Magic Lantern Forum

Experimental builds (WIP) => Other experimental builds => Topic started by: dfort on October 04, 2017, 05:20:58 AM

Title: dfort's experiments for 7D Mark I
Post by: dfort on October 04, 2017, 05:20:58 AM
Quote from: Walter Schulz on September 17, 2017, 09:50:08 PM
Is there anything to test on 7D for mere users?

I've got a test any user can easily run. It is the same test as the one we just did for the 650D. Let's start with the current Nightly build and see if there's a problem with the audio meters.

Here's the test:

Quote from: dfort on October 02, 2017, 10:41:13 PM
The way to run the test is fairly simple. Turn on both the mlv_rec and mlv_snd modules and activate those modules. I prefer to reduce the frame to the smallest size because we're only checking for an audio issue. The first time you record an MLV clip the audio meters should work normally. What I'm interested in is what happens on the second through the forth clip. Make sure not to restart the camera between takes. Do the audio meters continue to operate normally?

The 7D is setup the same as the 650D and 700D. Both of those cameras had issues with the audio meters that somehow slipped past everyone so it went unreported. This was recently fixed.

If you're curious:

http://www.magiclantern.fm/forum/index.php?topic=20597.0
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on October 04, 2017, 04:16:53 PM
Quote from: dfort on October 04, 2017, 05:20:58 AM
Do the audio meters continue to operate normally?

Yes.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 04, 2017, 08:53:13 PM
Great!

Of course you know what's coming next--more tests. I uploaded three builds labeled 7D_test-A/B/C to my bitbucket download page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/). Start with build A, it should work just like the unified build you just tested. If not--stop, don't go any further, I messed up. If it does work, try the B and C builds.

What is going on is that we made a change the mlv_snd module in the crop_rec_4k branch that I'd like to have in the unified branch. It uses both StopASIFDMAADC and SoundDevShutDownIn to stop the audio recording. However, it turns out that some cameras need to use one or the other or both in tandem so I'm trying to figure out what is needed for each camera. On the EOSM mlv_snd needs to shutdown audio using only StopASIFDMAADC while the 100D is fine using both. I was surprised to find out that the 700D and 650D can't use StopASIFDMAADC or SoundDevShutDownIn.

So:

A has the new mlv_snd changes with StopASIFDMAADC commented out.
B uses both StopASIFDMAADC and SoundDevShutDownIn.
C uses neither StopASIFDMAADC or SoundDevShutDownIn.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on October 04, 2017, 09:22:08 PM
Audio meter working for all versions.
B gives "audio failed to stop, status 4" after stopping 2nd recording (and all following)
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 04, 2017, 09:34:58 PM
Interesting. So let's comment out StopASIFDMAADC for now since it isn't needed and works with the proposed changes to mlv_snd.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on October 17, 2017, 06:13:09 PM
It looks like the 5DMk3 has received the most attention in this forum and the feature set that it has received due to the enormous efforts of the developers is incredible.  Now, this camera has become a very powerful tool for professional video recording. 

Unfortunately, not everybody can afford a 5DMk3 and the question is, are there other alternatives that work well with Magic Lantern, can do most of what the 5DMk3 does and can be obtained at a fraction of its cost.  In my opinion, the answer is Yes and the alternative is the classical 7D (Mark I).  Here are a few basic features that support this statement:

1) The fast write speed on the CF-card.  On 1066x CF cards I get continuous writing speeds in the 90 to 91 MB/s range.  This means continuous 14-bit RAW video recording at 1728x972 resolution with sound until CF card is full.  This also means quite long RAW video recordings at 10 and 12 bit with sound at larger resolutions (up to 2496x1200) in the 5x zoom mode.  I got 2 min. record time in this mode at 10 bit and 24 fps.  Finally, it means also continuous 10-bit RAW recording with sound at 2144x1200 resolution (aspect ratio close to 16:9) at the 5x zoom mode which could be quite useful for interviews, high-quality recording of dancers on the scene, weddings and even short events. 

2) The EOS 7D is a very solidly built semiprofessional camera with good weather sealing that can stand quite tough working conditions.  Battery life is excellent even when shooting MLV video with Magic Lantern.  I never managed to get the camera temperature higher than 56 deg. C during RAW video recordings at 35 deg. C on hot summer days.

3) Image and video quality are superb.  I especially like the micro contrast, the colors and the dynamic range of the videos and time lapses coming out of this camera.

4) VAF filters for this model that remove moire and aliasing in scenes with lots of straight lines and repetitive patterns are readily available (expensive though, ~300$).
----------------------------------------------------

After playing for several hours with my 7D last weekend I was so happy with the camera performance and results that I obtained with a fairly old build that I decided to write this post in an attempt to revitalize the attention to this fantastic camera.   I used the December 01/2016 build from the download link in post 1314 from the following thread:

https://www.magiclantern.fm/forum/index.php?topic=5601.1300

My overall impression is that the feature set in the above build is excellent and appropriate for serious video work as described in #1 above.  I also checked the classical FRSP function in the silent mode which reproduces beautiful timelapses at the full resolution of the sensor.  It works very well, also in combination with Dual ISO. In my tests the camera was very stable and there were almost no crashes, except that it is hard to close the aperture of STM lenses in the FRSP mode. Here are a few things that need to be fixed and if the fixes are possible and successful, in my opinion, this will make the 7D second best after the 5DMk3 for Magic Lantern work:

A.  I miss the Movie crop mode which allows fast and easy switching from the normal to the 3x zoom mode with preserving framing during recording.  In other words you see on the screen what you get.  This is extremely useful when working with fast prime lenses.  I use the EOS 35/F2 IS which is my favorite for RAW videos and with the Movie crop mode it turns into a 105 mm/F2 lens on a push of a button.  In this way, I can quickly zoom to an important part of the scene, make a recording there at 105 mm with a beautifully blurred background and then switch back to the normal mode.  Very convenient!  I know, I can do this in the 5x zoom mode, however, in that case I see only a fraction of the frame.  This makes framing with camera in motion (steadycam work) very difficult.

B.  A well known phenomenon is the freezing of the screen when recording starts in the 5x zoom mode.  For static shots on a tripod this is not a problem but when camera moves (e. g. steadycam shots) framing is impossible.  There is a known temporary work around this, as described in the post 1314 above but switching back and forth between menus takes time and important moments are easily missed.

C.  I also miss the "ML digital ISO" function.  It was very useful when recording video in bright sunny days where light is too much and even F22 is not enough to get the exposure right.  ML digital ISO would allow me to gradually reduce the exposure by up to -2 stops which I used very often as a digital neutral density filter.  It worked quite well with time lapses at reduced frame rates as well as FRSP work where shortest exposure times are limited to about 1/25-th of a second.  Finally, it provides smoother transitions in the highlights which is useful when shooting contrasty landscape photos (sunsets, sunrises, etc.).

D. The 10 and 12-bit MLV-rec and RAW-rec is broken.  Every 2-nd frame is corrupt.  This forces the operator to use 14-bit RAW which makes the files too big and expensive cards get full very quickly.  I know, this bug is not easy to fix and we can live with it but a fix would be more than welcome.
----------------------------------------

These are the issues which make me think that, with the 7D, we are very close to a camera with superb functionality for serious video shooting.  If someone else has observed issues that I have missed, comments and further tests are very welcome. 
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 17, 2017, 11:20:28 PM
Quote from: IDA_ML on October 17, 2017, 06:13:09 PM
A.  I miss the Movie crop mode which allows fast and easy switching from the normal to the 3x zoom mode with preserving framing during recording.

I answered one of your posts on another topic about this but now I'm wondering--you mean the 7D had this option in the past and now it is gone? It isn't in the features matrix (http://builds.magiclantern.fm/features.html) under Movie CROP_MODE_HACK.
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on October 17, 2017, 11:48:41 PM
Quote from: IDA_ML on October 17, 2017, 06:13:09 PM
It looks like the 5DMk3 has received the most attention in this forum and the feature set that it has received due to the enormous efforts of the developers is incredible.  Now, this camera has become a very powerful tool for professional video recording. 

Unfortunately, not everybody can afford a 5DMk3 and the question is, are there other alternatives that work well with Magic Lantern, can do most of what the 5DMk3 does and can be obtained at a fraction of its cost.  In my opinion, the answer is Yes and the alternative is the classical 7D (Mark I).  Here are a few basic features that support this statement:

1) The fast write speed on the CF-card.  On 1066x CF cards I get continuous writing speeds in the 90 to 91 MB/s range.  This means continuous 14-bit RAW video recording at 1728x972 resolution with sound until CF card is full.  This also means quite long RAW video recordings at 10 and 12 bit with sound at larger resolutions (up to 2496x1200) in the 5x zoom mode.  I got 2 min. record time in this mode at 10 bit and 24 fps.  Finally, it means also continuous 10-bit RAW recording with sound at 2144x1200 resolution (aspect ratio close to 16:9) at the 5x zoom mode which could be quite useful for interviews, high-quality recording of dancers on the scene, weddings and even short events. 

2) The EOS 7D is a very solidly built semiprofessional camera with good weather sealing that can stand quite tough working conditions.  Battery life is excellent even when shooting MLV video with Magic Lantern.  I never managed to get the camera temperature higher than 56 deg. C during RAW video recordings at 35 deg. C on hot summer days.

3) Image and video quality are superb.  I especially like the micro contrast, the colors and the dynamic range of the videos and time lapses coming out of this camera.

4) VAF filters for this model that remove moire and aliasing in scenes with lots of straight lines and repetitive patterns are readily available (expensive though, ~300$).
----------------------------------------------------

After playing for several hours with my 7D last weekend I was so happy with the camera performance and results that I obtained with a fairly old build that I decided to write this post in an attempt to revitalize the attention to this fantastic camera.   I used the December 01/2016 build from the download link in post 1314 from the following thread:

https://www.magiclantern.fm/forum/index.php?topic=5601.1300

My overall impression is that the feature set in the above build is excellent and appropriate for serious video work as described in #1 above.  I also checked the classical FRSP function in the silent mode which reproduces beautiful timelapses at the full resolution of the sensor.  It works very well, also in combination with Dual ISO. In my tests the camera was very stable and there were almost no crashes, except that it is hard to close the aperture of STM lenses in the FRSP mode. Here are a few things that need to be fixed and if the fixes are possible and successful, in my opinion, this will make the 7D second best after the 5DMk3 for Magic Lantern work:

A.  I miss the Movie crop mode which allows fast and easy switching from the normal to the 3x zoom mode with preserving framing during recording.  In other words you see on the screen what you get.  This is extremely useful when working with fast prime lenses.  I use the EOS 35/F2 IS which is my favorite for RAW videos and with the Movie crop mode it turns into a 105 mm/F2 lens on a push of a button.  In this way, I can quickly zoom to an important part of the scene, make a recording there at 105 mm with a beautifully blurred background and then switch back to the normal mode.  Very convenient!  I know, I can do this in the 5x zoom mode, however, in that case I see only a fraction of the frame.  This makes framing with camera in motion (steadycam work) very difficult.

B.  A well known phenomenon is the freezing of the screen when recording starts in the 5x zoom mode.  For static shots on a tripod this is not a problem but when camera moves (e. g. steadycam shots) framing is impossible.  There is a known temporary work around this, as described in the post 1314 above but switching back and forth between menus takes time and important moments are easily missed.

C.  I also miss the "ML digital ISO" function.  It was very useful when recording video in bright sunny days where light is too much and even F22 is not enough to get the exposure right.  ML digital ISO would allow me to gradually reduce the exposure by up to -2 stops which I used very often as a digital neutral density filter.  It worked quite well with time lapses at reduced frame rates as well as FRSP work where shortest exposure times are limited to about 1/25-th of a second.  Finally, it provides smoother transitions in the highlights which is useful when shooting contrasty landscape photos (sunsets, sunrises, etc.).

D. The 10 and 12-bit MLV-rec and RAW-rec is broken.  Every 2-nd frame is corrupt.  This forces the operator to use 14-bit RAW which makes the files too big and expensive cards get full very quickly.  I know, this bug is not easy to fix and we can live with it but a fix would be more than welcome.
----------------------------------------

These are the issues which make me think that, with the 7D, we are very close to a camera with superb functionality for serious video shooting.  If someone else has observed issues that I have missed, comments and further tests are very welcome.

Good summary IDA_ML. I add (also wrote on another post but better here)

E. 7D needs proxy simultaneous h264 recording for preview and Easy Off Line/ online workflow (thanks to danne's switch app)

F. 7D with 10/12/14 bits lossless (we could reach résolution higher than 1080p and longer recording on cf cards)

G. At least 45fps raw

This camera is dream even 8 years after first release and a lot of people are still using it. I'd love to help to improve the functionnalities. I use it on a daily basis and I have the vaf-filter.

So the question is : how we can help ?like running tests (as we see with 6D, 600/650D, 100D...) to speed up 7D development which is far away from the rest ? Can we run 7D on quemu ?

Envoyé de mon iPhone en utilisant Tapatalk
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 18, 2017, 12:29:09 AM
10/12bit seems to be the biggest challenge on Digic 4 cameras, though it works on the 60D [EDIT - and 600D] so maybe?
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on October 18, 2017, 01:25:47 AM
Quote from: dfort on October 18, 2017, 12:29:09 AM
10/12bit seems to be the biggest challenge on Digic 4 cameras, though it works on the 60D [EDIT - and 600D] so maybe?
Yes it would be a big step in combination of H264 proxy // EDIT - Ok I understand now that there are 2 different builds (Crop_rec_4k and 10/12bits). We know now that 10/12 bits could be fixed.

Any idea if Crop_rec_4k can be ported to 7D ? Dfort, can digic 4 camera run Croprec_4k ?
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 18, 2017, 06:54:52 AM
Quote from: 12georgiadis on October 18, 2017, 01:25:47 AM
Any idea if Crop_rec_4k can be ported to 7D ? Dfort, can digic 4 camera run Croprec_4k ?

Maybe. There are a few hurdles to jump over. Someone needs to figure this out for the 7D--these are the other Digic 4 cameras that can run 10/12bit:

#ifdef CONFIG_60D
#define DEFAULT_RAW_BUFFER MEM(MEM(0x5028))
#endif

#ifdef CONFIG_600D
#define DEFAULT_RAW_BUFFER MEM(MEM(0x51FC))
#endif


In addition, Digic 4 cameras have lossless compression code in their firmware but so far none of these cameras have been ported over to ML lossless. Finally, the crop_rec module, more uncharted territory for Digic 4 cameras.

Seems like an interesting challenge.

[EDIT] Let's get started. Download the file named find-SRM_BUFFER_SIZE.2017Oct17.7D203.zip from my Bitbucket downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/) and shoot a simple silent DNG using that build. You should see something like this on your LiveView:

(https://farm5.staticflickr.com/4506/37028349634_4e65eea9c2.jpg) (https://flic.kr/p/Yq4Yih)

Please report back with the SRM_BUFFER_SIZE.
Title: Re: dfort's experiments for 7D Mark I
Post by: Danne on October 18, 2017, 08:21:37 AM
SRM_BUFFER_SIZE 0x1f80000
Taken in photo mode
Title: Re: dfort's experiments for 7D Mark I
Post by: Panz on October 18, 2017, 10:08:15 AM
It's the same in photo and video mode, 0x1f80000
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on October 18, 2017, 10:55:49 AM
Quote from: dfort on October 17, 2017, 11:20:28 PM
I answered one of your posts on another topic about this but now I'm wondering--you mean the 7D had this option in the past and now it is gone? It isn't in the features matrix (http://builds.magiclantern.fm/features.html) under Movie CROP_MODE_HACK.

Obviously, I did not make myself clear enough, Dfort and I appologize for that.  No, I don't mean that.  The Movie crop mode is abvailable on the 100D which I also own and I have been using it a lot lately on that camera.  However, I often use also the 7D because of the fast write speed (much longer record times), large buffer size (continuous recording with sound) and higher resolution in the 5x zoom mode.  And when I compare the 7D with the 100D, it is the Movie crop mode that I miss mostly.

I also greatly appreciate your statement: "Let's get started" and thank you for that.  If all this functionality is added to the 7D then this awesome camera will become a serious competitor to the 5DMk3.  Used 7D-s in mint condition and 20 to 30% of their shutter count are currently available for about 300 Euros, so now is the moment to give this camera what it deserves and make many budget oriented video shooters happy.

Here is a short film that I shot entirely on the 7D at the beautiful Kapinovo waterfall, in the foot of the Balkan mountain, in the central part of Bulgaria:

https://vimeo.com/134597071
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 18, 2017, 01:24:48 PM
Ok--we took the first step. Don't expect this right away or that all the features will work at first. If you want to follow the progress:

https://bitbucket.org/daniel_fort/magic-lantern/pull-requests/7/crop_rec_4k-for-7d203/diff

Please feel free to contribute.

@IDA_ML - You're getting great quality out of the 7D. As far as getting Movie crop mode working, that might not be possible. The only cameras that can do that are the EOSM/100D/650D/700D which are closely related. The feature is something that is apparently built into the 600D, I don't know because I haven't handled that camera. The 5D3 doesn't have it and that is the camera that has the most ML features working.

BTW - Thanks for the dip in a cool Bulgarian stream. We're having a heat wave here in the Los Angeles area.
Title: Re: dfort's experiments for 7D Mark I
Post by: Danne on October 18, 2017, 02:16:23 PM
The 5D3 doesn't have it and that is the camera that has the most ML features working.
1920 1:1?
http://www.magiclantern.fm/forum/index.php?topic=19300.0
Title: Re: dfort's experiments for 7D Mark I
Post by: nikfreak on October 18, 2017, 03:06:46 PM
FYI:

https://www.magiclantern.fm/forum/index.php?topic=18443.msg191319#msg191319

You will also find this useful for getting lossless compression eventually working:

http://www.magiclantern.fm/forum/index.php?topic=18443.msg183741#msg183741

Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on October 18, 2017, 03:38:22 PM
Quote from: dfort on October 18, 2017, 01:24:48 PM
@IDA_ML - You're getting great quality out of the 7D. As far as getting Movie crop mode working, that might not be possible. The only cameras that can do that are the EOSM/100D/650D/700D which are closely related. The feature is something that is apparently built into the 600D, I don't know because I haven't handled that camera. The 5D3 doesn't have it and that is the camera that has the most ML features working.

BTW - Thanks for the dip in a cool Bulgarian stream. We're having a heat wave here in the Los Angeles area.

Thanks for your kind words, Dfort.  But it's not me getting great quality out of the 7D.  It's you guys - the Magic Lantern developers that made RAW-quality video possible in this remarkable machine!  And guess what - 1728x972 MLV video scales so nicely up to FHD (1920x1080) in post that the quality (fine detail, smooth transitions, superb dynamic range and colors) is way better than the built-in H.264 video at that resolution. The same applies to the 100D too and, used, it is even cheaper than the 7D! What you guys are doing with ML is pure magic!

As far as the "Movie crop mode" is concerned, as Danne pointed out for the 5DMk3, I hope that this mode will be available under the "Crop mode" menu upon a successful Crop_rec_4k port to the 7D. If also 10/12 and 14-bit recording (non compressed) is available in that mode, then Sound should hopefully work too.  I know, I am dreaming too much but I like dreaming.  I'll keep my thumbs pressed that you guys are successful and will follow this development closely.  I feel bad that I cannot help in a more efficient way but I have absolutely no coding skills.  All I can do is testing features and share my experience with them.

By the way, "cool Bulgarian streams"!  Why don't you just plan your next family vacation in Bulgaria and see them life by yourself?  I am sure, you will not regret.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 18, 2017, 04:08:43 PM
Quote from: nikfreak on October 08, 2017, 01:32:25 PM
tried to match stubs for digic4 (7D) and it seems it can be done but TTL_SetFlags seems to be the only one making troubles. At least i can't find / match it.
Any hint on that one? I guess (not checked) it could be same for 60d, 5d2 etc.

Wow--so you have a 7D? I'm doing this blind. You got those stubs in one of your repositories? We can put more of our effort into searching for TTL_SetFlags.
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on October 18, 2017, 05:58:24 PM
Quote from: IDA_ML on October 18, 2017, 03:38:22 PM

also 10/12 and 14-bit recording (non compressed) is available in that mode, then Sound should hopefully work too. 

IDA_ML no need for 10/12bit to get sound. If crop_rec_4K is ported you get sound in recording mlv_lite + h264 proxy (that contains sound) and sync it with Danne's switch app (see thread). It is the best pro workflow available.



Envoyé de mon iPhone en utilisant Tapatalk
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on October 18, 2017, 07:30:37 PM
Quote from: 12georgiadis on October 18, 2017, 05:58:24 PM
IDA_ML no need for 10/12bit to get sound. If crop_rec_4K is ported you get sound in recording mlv_lite + h264 proxy (that contains sound) and sync it with Danne's switch app (see thread). It is the best pro workflow available.

I cannot quite follow why is this the best pro workflow available.  Please explain. 

My understanding is that if you record in 10/12 bit uncompressed with the MLV-rec module, you have synchronous sound right away.  It is built in the MLV file - no need for proxies and additional syncing software.  You can mount the MLVs with the MLVFS system into virtual cDNG folders on a click of a button, open them in Resolve and start working with video and sound right away, even directly from your camera card, connected to your USB3.0 port via a fast card reader.  This workflow is very fast and efficient and does not need round trips to other software.  Pros: 10/12 bit recording provides  consistent file sizes that do not depend on fine detail and exposure of the scene.  LL-compressed ones strongly depend on these latter parameters and you can experience an unexpected stop when working at higher resolutions, close to the write speed limit of your camera.  Drawback: 10/12 bit files are slightly larger than LL compressed ones but H.264 proxies also take up some space, therefore, differences should not be so significant, especially at 10 bit.   

By the way, I have not tested Danne's Switch yet.  It is for MAC and as far as I know, there is no Win version available yet.  I am on Win7 x64.
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on October 18, 2017, 08:37:02 PM
My workflow with 5dmkiii, crop_rec_4k on a 30' Doc with 12hrs footage :

Shoot comp raw, 8...11 bits or 10/12/14 (depends on the context) + h264 proxys with sound.
You have access to crop mode with accurate framing and 45fps raw also (not available in 10/12bits)

Post : sync and stamp tc raw and h264 proxy in switch. It took me less than 30' as it's not transcoding (for details see thread).
Edit with offline workflow (like 90% of cinema industry) using proxy file with sound. It means almost straight editing. Cndng from mlvfs doesn't work in fcpx and avid, not very good in premiere and slow in resolve (because vs stripes on, fpn... etc) compared to h264 edit on the NLE of your choice. Ideally these files could be auto transcoded by NLE in proxy APR for max stability on any config (even old laptops).
After the edit, export, edit and mix your sound in protools/ripper, consolidate and conform pictures in resolve with mlvfs + an aff or xml. Guess what ? It's rekinling and pointing on the raw perfectly.
IMO It's not necessary if you edit a 10' film with small files but recommanded for indy feature project to save time (or money).
Alternatively, we can transcode everything in APR444 with lut if you buy a 10TB (+10TB backup)... but, well, not the idea.
That's why it would be great to get crop_rec_4k for 7D!
Hope it makes sense :)
Oh, and You can emulate mac os on pc with a virtual box or hackintosh. or run switch on Linux.


Envoyé de mon iPhone en utilisant Tapatalk
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on October 18, 2017, 10:13:31 PM
Thank you very much, 12georgiadis.  I understand now.  The workflow that you describe is appropriate for long films and sounds too complicated to me.  However, I agree, crop_rec_4k for 7D + H.264 proxy with sound makes perfect sence, especially if one needs to record at larger resolutions with sound.
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on October 18, 2017, 10:19:43 PM
Yes, with crop_rec__4k, we can expect higher than 1080p + sound with h264 proxys


Envoyé de mon iPhone en utilisant Tapatalk
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 18, 2017, 11:09:03 PM
Don't expect too much too soon. I found the addresses needed to get a very basic crop_rec module working, just mv720 in 3x3 mode. However, there's going to be a lot more work (https://bitbucket.org/daniel_fort/magic-lantern/pull-requests/7/crop_rec_4k-for-7d203/diff) to do to get the 7D to compile in the crop_rec_4k branch. I'm in over my head and will need some help with this.
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on October 18, 2017, 11:29:24 PM
Quote from: dfort on October 18, 2017, 11:09:03 PM
Don't expect too much too soon. I found the addresses needed to get a very basic crop_rec module working, just mv720 in 3x3 mode. However, there's going to be a lot more work (https://bitbucket.org/daniel_fort/magic-lantern/pull-requests/7/crop_rec_4k-for-7d203/diff) to do to get the 7D to compile in the crop_rec_4k branch. I'm in over my head and will need some help with this.
Good to hear that Dfort. We'll help you the best as we can [emoji4]


Envoyé de mon iPhone en utilisant Tapatalk
Title: Re: dfort's experiments for 7D Mark I
Post by: taestea on October 21, 2017, 03:18:36 AM
Hello, new ML user and forum member here.

Been a 7D user for many years though.

I am using the mini usb > rca cable that comes with the 7D and using an RCA to 3.5mm female adapter to monitor the audio through my headphones but plugging into the A/V mini usb disables the LCD screen (the headphone monitoring does work)

I am sure you are all probably very aware of this scenario but I've been searching everywhere online (maybe not hard enough?) and can't seem to find a definitive answer but I was wondering if there was a hidden menu for Headphone/Audio Monitoring for the 7D on ML? (Maybe through enabling modules?)

On the ML user guide page (https://wiki.magiclantern.fm/userguide) under Headphone Monitoring it states that "This feature is not available on 600D/T3i" but nothing about the 7D so I am assuming it is an available option for 7D.

I've seen videos on youtube about headphone monitoring using ML (different canon models) but their ML audio menu had a lot more options than the 7D ML build that I am running (Latest Build (2017-10-03 22:30))

So I am wondering if headphone monitoring WHILE being able to view the LCD is possible on the 7D?

Thank you so much!

-T

Title: Re: dfort's experiments for 7D Mark I
Post by: tonij on October 22, 2017, 07:18:09 AM
It's cool to read there's others out there who also love the 7D. I agree it's quite the weapon even as it currently stands
For a long time I've been debating to get a VAF filter, if theres a 7D user out there that has one I'd be very interested to see DNG's with and without
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on October 23, 2017, 06:34:16 PM
Quote from: tonij on October 22, 2017, 07:18:09 AM
It's cool to read there's others out there who also love the 7D. I agree it's quite the weapon even as it currently stands
For a long time I've been debating to get a VAF filter, if theres a 7D user out there that has one I'd be very interested to see DNG's with and without
you have tests in EOSHD. But there is a new VAF filter version that is coming. More efficient with wide angle and better AF integration.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 23, 2017, 07:20:23 PM
New filter for an old camera? Interesting. Do you have a link for that?
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on October 23, 2017, 07:24:49 PM
Quote from: dfort on October 23, 2017, 07:20:23 PM
New filter for an old camera? Interesting. Do you have a link for that?
For almost all cameras (and we can ask for eos. Message from mosaic :
we've been developing the capability to intrinsically compensate for the optical thickness of our supplementary filter, and if we introduce a new filter for the 100D, I would very much like to include this compensation.  This would mean a couple of things:  First, there would be no back focus shift as we've had with all our previous DSLR filters, so all lenses should continue to focus normally, and for zoom lenses, to track focus normally during zooming.  It would also finally fix the difficulties our filters have when operating with extreme wide angle lenses.  I'm not sure how wide of lenses you've used before, but even our current DSLR filters have difficulties when the lens gets wider than about 10-11mm, with for example circular fisheye lenses such as the Samyang 8mm, presenting the most problems.  With our new compensation all of that should be fixed.  The filter would also install and remove easily, without requiring any shimming, as we currently do with our Blackmagic filters.




Envoyé de mon iPhone en utilisant Tapatalk
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 23, 2017, 08:31:22 PM
Was that message sent to you directly from Mosaic Engineering? I can't find it anywhere on their site or any site for that matter using Google. Sorry, I studied journalism and don't believe anything unless it can be verified.
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on October 23, 2017, 10:23:39 PM
Quote from: dfort on October 23, 2017, 08:31:22 PM
Was that message sent to you directly from Mosaic Engineering? I can't find it anywhere on their site or any site for that matter using Google. Sorry, I studied journalism and don't believe anything unless it can be verified.
Dfort, this is a private message that he sent to me yesterday. I know « one source = no source » but you can write him or I can forward you the e-mail ;-)


Envoyé de mon iPhone en utilisant Tapatalk
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on October 29, 2017, 01:13:48 AM
Is there anyway that the 7D, given it's good buffer speed, can sit an hour or two, recording 1080p h.264 video? or even 720, maybe?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on October 29, 2017, 01:28:51 AM
Buffer speed in H.264 was never a problem for all movie enabled cameras. 1080p25 in H.264 with IPB frames will just need about 6 MByte/s. Don't know what you mean by "even 720". Look into manual p. 157. Bandwidth requirements for both modes are listed as identical.

If you can live with about half a second lost every 29:59, cam will do just fine. Cam will stop recording, ML will force restart. Mirror and shutter will do the noisy thing during stop/restart.
Don't know if a battery grip will host energy for 2h though. External power source might come handy.

Setting: Movie tab -> Movie Tweaks -> Movie restart.

There is a bug with this option on 7D. You won't be able to stop recording that easy. See https://bitbucket.org/hudson/magic-lantern/issues/2065/movie-restart-7d-stuck-in-loop-cannot-stop

RAW/MLV does not have this limitation. No gap in recording. Bandwidth and memory space might cause issues.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on November 01, 2017, 06:16:51 AM
I'm still getting familiar with this camera. The LiveView screen looks unlike other cameras I've seen. Is this right or did I mess something up?

(https://farm5.staticflickr.com/4504/37368954254_4d4ba7b7a1.jpg) (https://flic.kr/p/YWaEbJ)
Title: Re: dfort's experiments for 7D Mark I
Post by: reddeercity on November 01, 2017, 07:06:25 AM
Looks like the good old Digic4 to me  :D
just like my 5D2 , the only difference is .raw loaded (for testing , no audio) with mlv audio meters are enabled
http://www.magiclantern.fm/forum/index.php?topic=19336.msg192525#msg192525
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on November 01, 2017, 04:21:15 PM
Interesting, good to know I didn't somehow mess things up. Yes, this is a Digic 4 camera but it has two processors, a slave and a master. I searched all over about what the 7D_MASTER platform is about but didn't find much information. So I decided to compile it and see what it does.

(https://farm5.staticflickr.com/4488/37362131024_74741b3bc6_z.jpg) (https://flic.kr/p/YVyFSQ)

This is running on a 7D.203, I would have expected it to do something more like dump the firmware but this is all I got out of it. The 7D_MASTER must be used for something and apparently it is being actively maintained. Just the other day g3gg0 made this commit for the 7D_MASTER (https://bitbucket.org/hudson/magic-lantern/commits/dcc15836c1f33d46274916b729a026454b39cf46?at=unified) to the unified branch. Can anyone enlighten us on this mysterious platform?
Title: Re: dfort's experiments for 7D Mark I
Post by: g3gg0 on November 01, 2017, 11:53:25 PM
Quote from: dfort on November 01, 2017, 04:21:15 PMCan anyone enlighten us on this mysterious platform?

various patches/changes have to be made on the master digic.
but autoexec.bin is executed on the slave only.

luckily FIR files get executed on both master and slave, so we can put ML code in a FIR if we need to patch stuff on master.
so the MASTER part is meant to implement the necessary stuff ML needs to be patched and then packed into the FIR along with "real" ML for the slave.

only when executed through a "firmware update", ML is then able to patch this stuff. think it was only cache hack related stuff.
IIRC only the high bitrate video hack used these features.

short: not really needed
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on November 02, 2017, 12:20:41 AM
Quote from: g3gg0 on November 01, 2017, 11:53:25 PM
short: not really needed

Great because I skipped over it. Got the 7D working with the current 2.0.6 firmware (https://bitbucket.org/hudson/magic-lantern/pull-requests/878/update-to-7d206/diff). I put a test build on my Bitbucket downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/) but warning, you need to make sure the camera boot flag is set before updating the firmware because there isn't a .FIR for 2.0.6 yet.

So is there anything new in the 2.0.6 firmware that will get the 7D into the crop_rec_4k branch? Not as far as I can see though it still might be possible. My biggest hurdle right now is trying to figure how to get CONFIG_EDMAC_RAW_SLURP working on this Digic 4 camera like was done for other Digic 4 cameras:

#ifdef CONFIG_60D
#define DEFAULT_RAW_BUFFER MEM(MEM(0x5028))
#endif

#ifdef CONFIG_600D
#define DEFAULT_RAW_BUFFER MEM(MEM(0x51FC))
#endif


Any clues how to figure out this turducken code for the 7D?
Title: Re: dfort's experiments for 7D Mark I
Post by: Audionut on November 02, 2017, 04:11:37 AM
Forgot to say thanks the other day for the sf_dump code you fixed, and the build you pushed for the 6D.

You've been pushing so much code lately.  Thanks dfort.  Exciting when the code base is constantly expanding.  :)
Title: Re: dfort's experiments for 7D Mark I
Post by: john5788 on November 02, 2017, 07:04:45 AM
Quote from: Audionut on November 02, 2017, 04:11:37 AM
Forgot to say thanks the other day for the sf_dump code you fixed, and the build you pushed for the 6D.

You've been pushing so much code lately.  Thanks dfort.  Exciting when the code base is constantly expanding.  :)

Just wanted to chime in and say I agree and thanks!

I've had a T5i and dfort got ML running on the latest 1.1.5 firmware. Now that I've switched to a 7D, I'm happy to see a push for ML on the latest firmware on this platform as well!
Title: Re: dfort's experiments for 7D Mark I
Post by: reddeercity on November 02, 2017, 07:17:52 AM
Quote from: dfort on November 02, 2017, 12:20:41 AM
My biggest hurdle right now is trying to figure how to get CONFIG_EDMAC_RAW_SLURP working on this Digic 4 camera like was done for other Digic 4 cameras:

#ifdef CONFIG_60D
#define DEFAULT_RAW_BUFFER MEM(MEM(0x5028))
#endif

#ifdef CONFIG_600D
#define DEFAULT_RAW_BUFFER MEM(MEM(0x51FC))
#endif


I been trying to figure this out too , it seems that "RAW_LV_BUFFER_ALLOC_SIZE"
#ifdef CONFIG_5D2
#define RAW_LV_BUFFER_ALLOC_SIZE (2040*1267)
#endif


works on digic 4 (mostly) as noted  here (http://www.magiclantern.fm/forum/index.php?topic=19336.msg192525#msg192525) , it funny that it works on 1100d (digic4) that's where I got the idea and code from the 10-12bit branch to try on my 5d2 , maybe give it a try on 7D.




Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on November 02, 2017, 03:49:17 PM
Quote from: Audionut on November 02, 2017, 04:11:37 AM
Forgot to say thanks the other day for the sf_dump code you fixed, and the build you pushed for the 6D.

You've been pushing so much code lately.  Thanks dfort.  Exciting when the code base is constantly expanding.  :)

Quote from: john5788 on November 02, 2017, 07:04:45 AM
Just wanted to chime in and say I agree and thanks!

I've had a T5i and dfort got ML running on the latest 1.1.5 firmware. Now that I've switched to a 7D, I'm happy to see a push for ML on the latest firmware on this platform as well!

Now where's that blushing emoji?

To be fair, I'm more successful on the easier, low priority stuff. Every time I try something extra challenging, like porting the 7D to the crop_rec_4k branch or get lossless compression working on a camera that I don't have in my hands (6D, 650D) I tend to hit ceilings and walls.

Anyway, enough about me, there's no "I" in Magic Lantern. Oh, wait....
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on November 02, 2017, 05:08:30 PM
Remarkable progress, Dfort!  I all the time keep my thumbs pressed that you get RAW_SLURP working on the amazing 7D!  The crop_rec_4k and lossless compression will make the 7D a hell of a camera with its fast write speed and long battery life.

It may be interesting for you to know that yesterday I used my 100D with the sound recording feature in the lossless 8...11 bit mode for the first time on a long event shooting, (a 4 hour long street protest meeting).  I shot almost 50 GB of RAW footage (about 25 min.) with sound in the 1736x976 resolution switching back and forth from Normal to Movie Crop Mode all the time.  And guess what.  I did not have a single camera crash !!!  Most of the time it stayed at 43 deg. C.  It would barely reach 50 deg. C even with clips that were 3 to 4 minutes long.  I just had to change the battery once when it ran out of juice.   I was amazed how stable and cool the camera was.  I wish, my 7D would work in the same way but at 2520x1200 resolution, lossless and with sound.  Can you imagine what video quality we would have with the 7D at this resolution?  I know, I am dreaming but nonetheless, take your time and keep up the good work!
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on November 06, 2017, 06:21:16 PM
dfort!

I appreciate all your hard work on this forum! Excellent job! I am following this thread closely and am ready to test with my 7D whenever its ready!

Thanks again!

Rob
Title: Re: dfort's experiments for 7D Mark I
Post by: laserbit on November 08, 2017, 09:31:31 AM
Hi Guys, I'm im trouble with my "new" 7D and Magic Lantern... I'm not able to let it work. I don't know if this is the right place to write for asking support or not since I am new and I don't know very well all forum sections...
If not, I kindly ask to the mods to move the post in the right place.

What I have:
-Eos 7D with FW 2.0.5,
-Adapter CF to MicroSD
-64GB MicroSD
-Magic Lantern Nightly Build "Nightly.2017Oct31.7D203"

What I did:
-Downgraded FW from 2.0.5 to 2.0.3
-Formatted MicroSD from my card
-Copied inside the Nightly files
-Run update to install ML on the card

What happened:
After the reboot the camera did not stat anymore

Then I made a backup of the MicroSD files and tried to format it again from the PC and copy the files again, but with this Card the camera does not want to boot anymore

If I try a normal CF (I have a pny 32gb 100mb/s) the camera boots properly
I tried also to put another microsd in the adapter (I have a 128GB Samsung MIcroSD), but also with this the camera does not want to boot. Probably the installation was not completed successfully, but the boot flag has been set on the camera.
What I have to do now? I have also seen the 2.0.6 FW, is there any chance to upgrade and let ML work with this FW?

Thank you
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 09, 2017, 12:21:51 PM
Most SD(micro-SD)-to-CF adapters are crap and will cause compatibility problems (=unusable) with ML. Avoid at all costs.
Title: Re: dfort's experiments for 7D Mark I
Post by: laserbit on November 09, 2017, 09:42:46 PM
Quote from: Walter Schulz on November 09, 2017, 12:21:51 PM
Most SD(micro-SD)-to-CF adapters are crap and will cause compatibility problems (=unusable) with ML. Avoid at all costs.

Removed the CF adapter and copied the files on the standard CF, it worked since first boot like a charm! Thanks, now ML is working properly. THe problem is that before testing ML, the adapter was working great, is it possible that the magic lantern .FIR install fried something in the CF adapter? it works great on a standard 5D, but it is not recognized by my camera with boot flag anymore.
Title: Re: dfort's experiments for 7D Mark I
Post by: Roberto Mena on November 09, 2017, 11:07:41 PM
I agree with Walter Schulz, I work at a university where we check out a/v equipment and people using micro SD cards with SD adapters have always had issues shooting video with our Canon T3i and T5i DSLR cameras with their regular, native, h.264 video codecs and ML is way more demanding. Just get the proper fast CF cards they work like a charm for my 7D shooting MLV RAW video.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 10, 2017, 09:05:08 AM
I was talking about SD-to-CF adapters, not about Micro-SD-to-SD adapters. Second one are just a bunch of connectors und wires without any electronics involved. I use them frequently in 650D and other devices and issues I had: Next to non-existent = About same amount I'm having with full-size SD.
Title: Re: dfort's experiments for 7D Mark I
Post by: laserbit on November 15, 2017, 10:21:40 PM
Uninstalled ML, the SD-to-CF Adapter works perfectly, once I reinstalled ML, it stopped working again. For some reasons it's Magic Lantern that does not support the adapter. Any idea to fix it? I would like to use some big SDs I have for registering RAW videos. Can I provide any kind of additio.al detail?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 15, 2017, 11:08:41 PM
Please lookup "crap" and "avoid at all costs" in a dictionary of your choice.
If you insist using those adapters with ML you have to contact adapter manufacturer and ask for remedy.
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on November 16, 2017, 11:28:38 PM
I have been using my 7D to record long takes for events that I film in MLV RAW with Sound at 1728x724. I can record for about 65 minutes for that resolution which is around 200-205 gb of space. I have discovered that it doesn't matter what resolution I record, it always stops when the card fills up to 200 gb of space. The gui freezes and I have to restart the camera and pull the battery to be able to record for the remaining time. Does anyone know why this happens? Is there a way to fix it to fill the entire card?

EDIT - When recording to a 256 gb compact flash card.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on November 17, 2017, 07:12:21 PM
Sounds like you reached the limit of a 256 GB CF card. Yeah, it is frustrating but because of formatting overhead and the different ways manufacturers measure capacity you will never be able to put 256 GB of data on a 256 GB card. You can Google to find out why.

Just out of curiosity, what is the free space of one of your cards that has filled up with 200 GB of MLV files?

[EDIT] Unless you are using 512 GB cards--you didn't specify the size of card you're using so I'm assuming 256 GB.
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on November 17, 2017, 08:53:24 PM
Hi Dfort,

Thanks for your response. Sorry, yes, it is a 256gb card that I am recording to. And once I restart the camera after it stops at 200gb I can record the remaining 56gb on the card. When I download the card the card is full at 256gb. :)

Its strange because I can film a take that is 30gb of space and then try to record another take and let it go as long as it can and the camera will stop recording once it hits 170gb. So it stops once it hits a total of 200gb. What I have been doing is just powering off the camera before I record a second long take so I can record at least 200gb. It just seems like there is something that is keeping it from recording more that 200gb. I was just hoping its an easy fix. :) Which I am sure is never really the case. 

Thanks again for your response! Do you have any ideas of what may be causing this?

Rob
Title: Re: dfort's experiments for 7D Mark I
Post by: a1ex on November 17, 2017, 09:11:38 PM
To narrow down, you can run a couple of tests: with a card formatted in the camera, record as much as you can, with:
- mlv_rec with sound (your current configuration that fails)
- mlv_rec without sound
- mlv_lite
- raw_rec (with some older version), only if the above did not work
- custom code that writes something to a file over and over (minimal example), only if the above did not work

Also, how does the crash look like? (e.g. can you film the camera screen during the crash?)
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on November 18, 2017, 12:10:11 AM
Quote from: a1ex on November 17, 2017, 09:11:38 PM
To narrow down, you can run a couple of tests: with a card formatted in the camera, record as much as you can, with:
- mlv_rec with sound (your current configuration that fails)
- mlv_rec without sound
- mlv_lite
- raw_rec (with some older version), only if the above did not work
- custom code that writes something to a file over and over (minimal example), only if the above did not work

Also, how does the crash look like? (e.g. can you film the camera screen during the crash?)

Alex!

Great idea on the tests to narrow it down. Here are the results for the tests and an interesting finding!

- mlv_rec with sound (your current configuration that fails) - Fails at 200 gb approximately 48 minutes (1728x972)
- mlv_rec without sound - Same as above... Fails at 200 gb approximately 48 minutes at (1728x972)
- mlv_lite - DOES NOT FAIL. I can record for the amount of space on the card at 256gb and about 60 minutes at (1728x972)

So I guess there has been something introduced in mlv_rec that causes it to stop at 200gb? Do you have any ideas of what it could be?

Another solution would be to add sound to mlv_lite. Is there a module that can be loaded to allow sound to work with mlv_lite?

Per your request, here is a video of how it stops when recording mlv_rec like your requested. A few seconds into the video you will see the timer stop and the word "Stopped" appear behind the recording icon. The image doesn't freeze but the magic lantern menu is frozen as you can see near the end of the video:
http://vimeo.com/243385811 (http://vimeo.com/243385811)

I am interested to hear what you think. :)

Thanks!

Rob
Title: Re: dfort's experiments for 7D Mark I
Post by: a1ex on November 18, 2017, 03:06:22 PM
Not sure what might be causing the issue, but mlv_rec's file I/O backend is pretty different from mlv_lite's; maybe g3gg0 has some ideas.

There is work in progress for making mlv_lite compatible with mlv_snd.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 18, 2017, 03:29:10 PM
Quote from: rob_6 on November 18, 2017, 12:10:11 AM
A few seconds into the video you will see the timer stop and the word "Stopped" appear behind the recording icon.

Are there 2 beeps when "Stopped" appears?
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on November 18, 2017, 05:46:44 PM
Quote from: Walter Schulz on November 18, 2017, 03:29:10 PM
Are there 2 beeps when "Stopped" appears?

Hi Walter,

I don't hear any beeps when I use mlv_rec. I don't even get any beeps when I start recording either. The Beep, test tones in the Audio tab of ML has it set to on but they don't beep when start, stop, or even freezing of recording happens.

Let me know if there is another way I can test that for you or if there is a way to enable those beeps.

Rob

Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on November 18, 2017, 06:38:38 PM
Quote from: a1ex on November 18, 2017, 03:06:22 PM
Not sure what might be causing the issue, but mlv_rec's file I/O backend is pretty different from mlv_lite's; maybe g3gg0 has some ideas.

There is work in progress for making mlv_lite compatible with mlv_snd.
Alex,

Yeah, that would be great if g3gg0 has any ideas. :)

Thanks again for helping me debug the issue. :)

Rob
Title: Re: dfort's experiments for 7D Mark I
Post by: Asanchezleache97 on November 22, 2017, 06:46:32 PM
Hi! I'm trying to uninstall and remove the ML bootflag to make my camera more responsive, since I use it for street.
I have the 2.03 firmware with the August Nightly installed, and when trying to follow the instructions (Select "Firmware Update") the camera simply tells me that there is no firmware in the CF card.
All Youtube tutorials I see use the same procedure on several cameras. What am I doing wrong?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 22, 2017, 09:17:01 PM
Format card using cardreader.
Format card in cam.
Copy extracted nightly build content to card.
Run firmware upgrade.
Report back.
Title: Re: dfort's experiments for 7D Mark I
Post by: Asanchezleache97 on November 23, 2017, 12:03:38 AM
Thanks a lot! No trace of ML anymore. Not that I don't think it's an amazing tool, just not for my work... I'll still miss the detailed DOF information and instant zoom in playback :)
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 23, 2017, 03:05:50 AM
I don't understand your reply.
Does your cam show some information with this card or not? If portable display test works your cam's bootflag is set and your diagnosis "No trace of ML anymore" is wrong.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on November 23, 2017, 10:03:03 PM
Quote from: tonij on October 22, 2017, 07:18:09 AM
It's cool to read there's others out there who also love the 7D. I agree it's quite the weapon even as it currently stands
For a long time I've been debating to get a VAF filter, if theres a 7D user out there that has one I'd be very interested to see DNG's with and without

I bought the VAF Filter for my 7d.
I'm slacking- haven't used it much, but I can pump out some examples for you if you'd like to see it!
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on November 24, 2017, 05:30:20 AM
Quote from: domo94 on November 23, 2017, 10:03:03 PM
I bought the VAF Filter for my 7d.
I'm slacking- haven't used it much, but I can pump out some examples for you if you'd like to see it!

Domo94,

I would love to see some tests you have ewiht your VAF filter. I would especially like to see some static shots with and without the filter so I can see the difference in RAW recording if you get the chance! Thanks!

Rob
Title: Re: dfort's experiments for 7D Mark I
Post by: Xylo on November 25, 2017, 12:19:57 PM
Hello to everyone
I was wondering if this new ML is also compatible with canon 7D firmware 2.0.6. If not which are the solutions to have ML installed?
Thanks,
Title: Re: dfort's experiments for 7D Mark I
Post by: tonij on December 07, 2017, 09:01:45 AM
Quote from: domo94 on November 23, 2017, 10:03:03 PM
I bought the VAF Filter for my 7d.
I'm slacking- haven't used it much, but I can pump out some examples for you if you'd like to see it!

Yes please!!
Title: Re: dfort's experiments for 7D Mark I
Post by: laserbit on December 08, 2017, 06:20:05 PM
Quote from: Xylo on November 25, 2017, 12:19:57 PM
Hello to everyone
I was wondering if this new ML is also compatible with canon 7D firmware 2.0.6. If not which are the solutions to have ML installed?
Thanks,
as far as I know only 2.0.3 is stable at the moment, so you have to downgrade. I downgraded from 2.0.5 to 2.0.3 with no issues, but I think that from 2.0.6 will be the same.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on December 09, 2017, 12:32:55 AM
Sort of a low priority project and highly experimental but I do have a build for 2.0.6 on my downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/). Just updated the pull request (https://bitbucket.org/hudson/magic-lantern/pull-requests/878/update-to-7d206/diff) for this so there aren't any conflicts if anyone wants to play around with it.
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on January 02, 2018, 07:57:12 PM
Hello everyone,

is it possible to implement H264 proxy mode in the 10/12 bits builds ? I'd like to make some test with 10 bits raw and very low bitrate (ex : 0.6x) H264 proxy. If this works, we can have a real time playback to watch the shot and better evaluate actor's performance on set for fiction film ! And for post-production, we can have a direct offline/online workflow, thanks to danne's switch workflow.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on January 02, 2018, 08:56:27 PM
Quote from: 12georgiadis on January 02, 2018, 07:57:12 PM
is it possible to implement H264 proxy mode in the 10/12 bits builds ?

I haven't tried it yet. 10/12 bit needs some more work first. The plan (my plan) is to also add lossless compression and the crop_rec module and all the other goodies from the crop_rec_4k branch but it is going to take a lot of work to get there. Any help would be greatly appreciated.
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on January 02, 2018, 09:50:05 PM
Quote from: dfort on January 02, 2018, 08:56:27 PM
I haven't tried it yet. 10/12 bit needs some more work first. The plan (my plan) is to also add lossless compression and the crop_rec module and all the other goodies from the crop_rec_4k branch but it is going to take a lot of work to get there. Any help would be greatly appreciated.
Ok, I understand Dfort. As far as I understand, there is still a bug with cropmode 10 bits, right ? Ok, I can't code, but can test builds of course.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on January 03, 2018, 12:12:36 PM
Amazing plan, Dfort, hopefully you will be able to realize it!  You can hardly imagine the excitement with which 7D owners follow your work.  I few days ago, I talked to a friend of mine who is event and nature videographer.  He said that his decision to keep his 7D or move to another camera depends entirely on whether or not you will be able to implement crop_rec_4k or at least fix the 10/12-bit recording feature in this camera.  If you succeed, he said, he will not only keep his 7D for several more years to come but he will also buy a second one for backup.

So, Happy New Year and keep up your excellent work!  On my part, I also promise to test every new expreimental build that you release and provide detailed feedback on the video features as soon as possible.  Hopefully, other knowledgeable and skilled people like A1ex, Nikfreak, Danne, Reddeercity and others will be able to help completing this important development too.
Title: Re: dfort's experiments for 7D Mark I
Post by: 12georgiadis on January 03, 2018, 12:40:19 PM
We had exactly the same talk with friends here in Belgium ! A lot of them have 7DmkI with PL mount+vaf filter. With crop_rec_4K (and also 10/12 bits), these cameras can have a second life for fiction projects ! And my friends don't want to move for another new canon other than 7DmkI that they really love.
Title: Re: dfort's experiments for 7D Mark I
Post by: jkanter3 on January 13, 2018, 04:37:12 AM
I apologize in advance if this should be obvious but I swear I've been searching all night. I'm trying to install ML on a 7D with 2.0.6 firmware. I see the nightly build but when I extract it, it has no .FIR file. I've downgraded back to firmware 2.0.3 and the compatible ML, which works great. Am I supposed to use the same .FIR file from the 2.0.3 version of ML? Do I grab a .FIR file somewhere else other than the 2.0.6 zip file? Thanks for any help.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on January 13, 2018, 08:43:32 AM
ML for 7D.206 is *not* a nightly build but built by user dfort for people wanting to play. FIR is not included yet and FIR for 2.0.3 is not compatible.

Use a second card to store Canon firmware files. Outdated slow one with low capacity will do fine. Format card in cardreader, format card in cam and copy extracted Canon FIR files to card.

Now take your card with ML installed. Rename ML directory to ML.203, Autoexec.bin to Autoexec.bin.203 and ML-SETUP.FIR to ML-SETUP.FIR.203.
Copy extracted ML files for 2.0.6 to card.

Update Canon firmware to 2.0.6. After installation you insert ML card and everything is fine.
As long as you do not format that ML card ...
Title: Re: dfort's experiments for 7D Mark I
Post by: jkanter3 on January 13, 2018, 01:26:04 PM
Thanks for the quick help. I will try all of that. And I understood that 2.0.6 was not a nightly build. My typing error. When I had firmware 2.0.6 installed, and extracted ML 2.0.6 to the card, nothing happened. That's why I guess I assumed it needed a FIR file. But with firmware 2.0.6 installed, and ML 2.0.6 on the card, what makes ML run? Does keeping the old ML 2.0.3 files there do something? That I hadn't done. Anyway, I'll try those steps as you've outlined. Thank you.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on January 14, 2018, 06:47:30 PM
There are 3 things necessary to make ML run:
- Cam with "boot flag" set. For 7D changing boot flag requires 2.0.3. Bootflag will not be affected by firmware up-/downgrades. One time procedure.
- A bootable card. Done by installing ML 2.0.3 in cam or using utilities like MacBoot, EOScard or shell scripts with a cardreader. Format card inserted into cardreader -> card non-bootable.
- ML files on card. That is: autoexec.bin and ML directory. They must match Canon firmware version!

A cam with bootflag set will look for a bootable card (thus causing some delay at startup and wakeup from sleep. Doesn't matter if card is bootable or not!)
If a bootable card is found cam will try to locate and load autoexec.bin. If found (and versions are matching) ML is loaded into memory.

Clear now?
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on January 15, 2018, 02:58:05 AM
Good news for 2.0.6 -- I've got a working ML-SETUP.FIR Been testing it out and it is working fine.

So now you don't need to downgrade to 2.0.3 in order to set the camera bootflag.

As usual, my test builds are posted on my Bitbucket downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/).
Title: Re: dfort's experiments for 7D Mark I
Post by: jkanter3 on January 15, 2018, 05:03:40 AM
Thank you both. I think I understand and will work on it. Much appreciated.
Title: Re: dfort's experiments for 7D Mark I
Post by: Sapporo on January 18, 2018, 02:54:57 PM
Quote from: dfort on January 15, 2018, 02:58:05 AM
Good news for 2.0.6 -- I've got a working ML-SETUP.FIR Been testing it out and it is working fine.

So now you don't need to downgrade to 2.0.3 in order to set the camera bootflag.

As usual, my test builds are posted on my Bitbucket downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/).
Does IO crypt also work with the new ML?
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on January 18, 2018, 05:41:45 PM
Quote from: Sapporo on January 18, 2018, 02:54:57 PM
Does IO crypt also work with the new ML?

I have never used that module but if I got the addresses right it should work.

https://bitbucket.org/hudson/magic-lantern/pull-requests/878/update-to-7d206/diff#chg-modules/io_crypt/io_crypt.c

[EDIT] The example shows the addresses for 2.0.3 so I updated the example. The values used in the module didn't change with the firmware update.
Title: Re: dfort's experiments for 7D Mark I
Post by: DeafEyeJedi on January 18, 2018, 09:21:49 PM
I can also confirm that updating to 2.0.6 on the 7D also works with ML from @dfort's download page. Nice work!

However, I decided to run the self test module (15 minute session) and it ended up getting stuck with the 'Busy' single on LCD viewfinder as well as tried turning it off while it continues to say 'Busy' on the LCD.

Kind of was expecting this anyway. Everything else seems to work like normal though I would prefer to go back to 2.0.3 just to have the 10/12-bit options and whatnot for the time being.  8)
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on January 18, 2018, 09:32:19 PM
Yeah, 2.0.6 is a rather low priority work in progress. If anyone wants to test or correct my mistakes please to do so!
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on January 21, 2018, 10:14:10 AM
Good to see 2.0.6 on 7d working. Latest and greatest news.

But what's the difference? Does 2.0.6 7d work a little smoother or more efficiently than 2.0.3? What are the benefits?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on January 21, 2018, 10:28:26 AM
https://www.usa.canon.com/internet/portal/us/home/support/product-advisories/detail/Firmware%20Notice%20EOS%207D%20Firmware%20Version%202.0.6
QuoteFirmware Notice: EOS 7D: Firmware Version 2.0.6

Thank you for using Canon products.

We would like to offer our sincere apologies to customers who have been inconvenienced by the phenomenon described below.
Phenomenon

When using the EF-S 18-135mm f/3.5-5.6 IS USM and EF 70-300mm f/4-5.6 IS II USM with certain cameras, lens aberration correction does not work.

If lens correction data is registered to the camera, "Correction data available" will appear on the camera's setting screen, but even if "Enable" is selected for lens aberration correction, lens correction will not be applied.

However, please note that Digital Photo Professional can be used to apply correction to RAW images.


http://canoncanada.custhelp.com/app/answers/detail/a_id/45748/~/firmware-notice%3A-eos-7d%3A-firmware-update-version-2.0.5
QuoteThank you for using Canon products.

Canon has released a firmware update for the EOS 7D digital SLR camera.

Product
EOS 7D Digital SLR

Firmware
Firmware Version 2.0.5 incorporates the following fix:

    Fixes a phenomenon in which the image files cannot be transferred using the FTP protocol via USB cable after the Canon EOS 7D camera has established a wireless connection to the Wireless File Transmitter WFT-E5A.
Title: Re: dfort's experiments for 7D Mark I
Post by: banertop on January 24, 2018, 02:35:46 PM
Hi guys,

Thinking about buying 7d for ml raw...

So, where are we in this moment?

1.Still, no 4k crop mode for 7d?
2.10 bit, 12 bit is possible, right?
3.What are the max resolutions in crop mode? Is 2.88k possible for 10 or 20 sec, or 2.5k is the max?

And, maybe, most important question......Would you go for 5d2 instead of 7d, and why, or why not?

Have some experience with 5d3 and ml raw,shoot in 4k crop mode for last 6 months, but zero experience with 7d and 5d2, so that is why i am asking....

Any suggestions could help

Thanks for your time>))
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on January 24, 2018, 04:09:44 PM
For current status of the 7D, please read here:

https://www.magiclantern.fm/forum/index.php?topic=5601.msg195624#msg195624

and here:

https://www.magiclantern.fm/forum/index.php?topic=5601.msg195585#msg195585

You can find additional information on the last few pages of that same thread.

For me, the 7D is second best for serious RAW video work with Magic Lantern after the 5DMkIII.  This is because of the fast CF-card writing speed.  I hope that the developers will continue their work in porting 4K_crop to this camera and if they succeed, this will be a real break through keeping in mind the serious interest in it among the 7D shooters.
Title: Re: dfort's experiments for 7D Mark I
Post by: banertop on January 24, 2018, 06:10:42 PM
IDA_ML, Thank you,

that is what i thought about 7d, so i am closer to buying 7d in general.....aliasing is an issue for 5d2 as well, so better write speed sounds like a better option.

But, i did not find answers for my questions on the links you send.

What i understood is, that if i use 7d, for higher resolutions, i have to use 10/12-bit RAW video build, instead of 3k corp mode, witch i used while i was shooting with 5d3.
But it is very confusing reading the treat.....one post is about one camera, next about the other camera.....i lose my self>))

So, if you are a 7d user, please tell me the higher continuous and higher non continuous resolutions with 7d in regular mode, in crop mode, and with fps override 50fps or 60fps

For example, I liked very much to shoot 2.88k with 5d3, so I was hoping to get at least 10 or 20 usable seconds in that resolution with 7d....is it possible?

Thanks a lot
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on January 24, 2018, 06:45:11 PM
Banertop,

There is no 4k_crop_recording on the 7D yet.  You can record only in 10, 12 and 14-bits uncompressed RAW video.  In the uncropped mode you can record continuously at 24 fps with sound, at all three bit rates and the resolution is 1728x972.  There is aliasing in this mode, so you may need a VAF filter to suppress it.

The maximum resolution in the 5x-magnification mode is 2496x1198 and you can record at 24 fps with sound at 10 bits quite long (I believe, more than a minute). Here video quality is fantastic and you don't need a VAF filter for this mode. 

You can record at 50/60 fps with the 7D and the horizontal resolution is 1728p while the vertical one was something like 670 or so but you need vertical stretching in post.  I tried that a long time ago but as far as I remember, there was aliasing in that mode too, so I was not very satisfied with the results.

I personally use the normal uncropped and the 5x-magnification modes with the 7D and am very satisfied with the results.
Title: Re: dfort's experiments for 7D Mark I
Post by: banertop on January 24, 2018, 07:14:45 PM
Thank you very much....very informative....you save me a lot of time>))

If i am not boring, I have just two more questions,

1.what are "all three bitrates" (no such a thing in the builds that I shoot with)?
2. Witch cards do you  recommend, since 7d is an old camera....is it work with newer cf cards, or with some specific card? I ask, because, in some treat guy said that he ended up frying couple of cf cards with 7d, and stop using 7d for raw....

Thanks again
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on January 24, 2018, 09:51:58 PM
1) 10, 12 and 14 bits

2) Any CF-card with at least 100MB/s write speed.  Transcend, Sandisk, Lexar x1066 work fine.
Title: Re: dfort's experiments for 7D Mark I
Post by: banertop on January 24, 2018, 10:29:20 PM
Thanks>))
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on January 26, 2018, 09:55:50 PM
Quote from: dfort on January 02, 2018, 08:56:27 PM
I haven't tried it yet. 10/12 bit needs some more work first. The plan (my plan) is to also add lossless compression and the crop_rec module and all the other goodies from the crop_rec_4k branch but it is going to take a lot of work to get there. Any help would be greatly appreciated.

How can a non-coder like me help
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on January 26, 2018, 10:40:23 PM
In a few months when I've earned some good profits from crypto trading, can I donate some bitcoin to the development team for the 7d?
We need raw video on the 7d to be incredible. Working better than 5dmk3
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on January 26, 2018, 11:30:00 PM
Well you could see if this is still valid: https://www.magiclantern.fm/donate.html

However, if you read the signature of g3gg0, one of the main developers, you'll see this:

QuoteONLY donate for things we have done, not for things you expect!
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on January 27, 2018, 01:36:51 AM
Shit, I just want to donate is my point.

You guys have done enough, but I'm broke in terms of donation money.

Soon, though. soon.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on January 27, 2018, 05:07:29 AM
Quote from: domo94 on January 26, 2018, 10:40:23 PM
...can I donate some bitcoin to the development team for the 7d?

Quote from: domo94 on January 27, 2018, 01:36:51 AM
...but I'm broke in terms of donation money.

???
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on January 27, 2018, 04:49:11 PM
Quote from: dfort on January 27, 2018, 05:07:29 AM
???

I meant for right now.

Like, I'm gonna donate, but in the next couple months when I stock up again. Just a temporary thing haha
Title: Re: dfort's experiments for 7D Mark I
Post by: scotophorus on February 05, 2018, 12:54:32 AM
@dfort First of all thank you for your kind contributions. I'm here to report that i've been getting tearing on a lot of frames on your build for the 2.0.6 firmware for the 7d when recording 14bit raw video with audio and using an external hdmi monitor. Even when using fps override. I didn't had this trouble when i was using your 10..12bit builds for the 7d 2.0.3.


(https://thumb.ibb.co/k8j3yx/2018_02_04_18_48_56.png) (https://ibb.co/k8j3yx)
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on February 06, 2018, 04:47:05 AM
Wow, someone is actually using the 2.0.6 version. That was mostly a challenge because I used a 5D3 to do the firmware dump.

https://www.magiclantern.fm/forum/index.php?topic=16534.msg192336#msg192336

This isn't a high priority project but it would be nice to get the latest Canon firmware working with Magic Lantern.

So this is only happening when using an external HDMI monitor? I don't have one to test with so I'll need some help resolving that issue. Not really sure where to start.
Title: Re: dfort's experiments for 7D Mark I
Post by: a1ex on February 06, 2018, 08:55:29 AM
http://www.magiclantern.fm/forum/index.php?topic=12760

Not a 2.0.6 issue; most likely, the memory bandwidth required by full-res HDMI (6 times as high, compared to built-in LCD) is simply too high for DIGIC 4 (it does not leave much room for raw recording).

The advice to switch to HDMI VGA should probably be in the menu; unsure whether this applies to any of the D5 models, or whether only D4 is affected.
Title: Re: dfort's experiments for 7D Mark I
Post by: reddeercity on February 06, 2018, 07:34:56 PM
HDMI work find at full res with 5d2 I regularly record the HDMI signal while recording raw video @1856x928 23.976p in 14bit . I'm know that other 7d user use to be able do the same , but
I remember talk about some frame tearing , but was fixed I thought .
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on February 06, 2018, 10:01:18 PM
Hello Dfort,

I have tried to test your latest 7D build:

raw_video_10bit_12bit_LVState-wip.2018Feb04.7D203

As much as I wanted to report some good news, I am afraid, I don't have any.  I tried to download the modules but after turning camera off and on to activate them, they don't load and I get an "Err" message behind each module.  Maybe, you could continue the work on the 7D starting from the Dec. 11-th, 2017 build.  It was the last one that was working pretty well. 
Title: Re: dfort's experiments for 7D Mark I
Post by: arrinkiiii on February 07, 2018, 05:51:14 AM

Hi guys, long time...

So, it's already possible to record 10bits and 12bits in HD?  :o  Here i can download this version for give it a trie? And the Lv is working good?

If i understand good, the problem now is to record in crop mode for the 4K?

Thanks 4all
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on February 08, 2018, 12:01:06 AM
Quote from: arrinkiiii on February 07, 2018, 05:51:14 AM
So, it's already possible to record 10bits and 12bits in HD?  :o 

Not quite. I've been experimenting but not being a seasoned coder I started off going in the wrong direction. There are test builds on my downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/) but they are far from stable. The best way to test is to set up a development environment and start tweaking what I started or try something new. It is actually quite easy to compile Magic Lantern. I posted tutorials for Mac (https://www.magiclantern.fm/forum/index.php?topic=16012.0) and Cygwin/Windows (https://www.magiclantern.fm/forum/index.php?topic=15894.0) that are well suited for beginners.

Quote from: IDA_ML on February 06, 2018, 10:01:18 PM
Maybe, you could continue the work on the 7D starting from the Dec. 11-th, 2017 build.  It was the last one that was working pretty well. 

The latest is a continuation of the Dec. 11-th build. I think that what you might be seeing is a strange phenomenon where sometimes all is well and other times it is broken. Try resetting ML to default, either through the menu or delete the SETTINGS folder, turn on only mlv_lite, restart--now if you shoot a video in either mv1080 or zoom modes in 10/12bit it will probably work. Reboot the camera and it will probably record corrupted files. Not sure what is going on with that. It happens to me on the latest as well as the older builds.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on February 08, 2018, 12:17:18 PM
Thank you Dfort!  I am glad to see that, despite all difficulties, you haven't given up on the marvelous 7D and this is the best news for me.

I will test again as soon as I find some time for that.
Title: Re: dfort's experiments for 7D Mark I
Post by: arrinkiiii on February 09, 2018, 12:15:16 PM

Thanks DFort, appreciated for the answer. So, its the best recording in 14bits for not having tearing or errors, cant have that in a paying gig :)

keep the good work, for sure that one day ours 7D's will record 10 and 12 bits without errors 8) 
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on February 09, 2018, 12:48:04 PM
Dfort,

A very brief feedback from first tests with your build:

raw_video_10bit_12bit_LVState-wip.2018Feb04.7D203

Following your instructions, I managed to upload and activate the mlv_lite, mlv_rec and mlv_play modules.  I tested RAW video recording with the MLV_REC module.  And it works - both in the normal uncropped mode at 1736x972 and the 5x-magnification mode at 2496x1198 resolutions. The Preview at 5x-crop is broken and screen freezes when recording starts.  If the RAW_twk module is also activated, the already recorded clips can be played back in the camera.  If an attempt is made to record in the presence of that RAW_twk module, the recorded 1736x976 video has again earthquake shaking.

My overall impression is that in the MLV_rec module, this build works in the same way as the "miracle" December 11-th build.  This is VERY GOOD NEWS since it proves that the 7D is capable of working without earthquake shaking !!!  It also appears to provide quite stable operation in both - the normal uncropped and the 5x-magnification modes at 10 and 12 bits which is EXCELLENT NEWS !!! If you could fix the freezing screen issue at 5x-magnification, this build will be perfectly useable !  Please keep up the good work, Dfort !
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on February 24, 2018, 04:22:26 AM
In other words, here's a video I made with footage shot in late 2016.
Details in description.

7D raw was a lot more stable back then. I don't have this kind of stability anymore.

https://vimeo.com/257201475

Not sure if it's my camera or not.
Enjoy !

Quote from: a1ex on February 06, 2018, 08:55:29 AM
http://www.magiclantern.fm/forum/index.php?topic=12760

Not a 2.0.6 issue; most likely...

How is the 2.0.6 build?
Title: Re: dfort's experiments for 7D Mark I
Post by: a1ex on February 24, 2018, 12:03:14 PM
Quote from: domo94 on February 24, 2018, 04:22:26 AM
7D raw was a lot more stable back then [late 2016]. I don't have this kind of stability anymore.

Old builds are still available; if you can identify a regression, please do a proper bug report (https://www.chiark.greenend.org.uk/~sgtatham/bugs.html).

In particular, since 2016, the main builds received almost exclusively bug fixes on the raw recording side (look at the changes (https://builds.magiclantern.fm/7D-203-all-builds-changes.html)). In particular, for video, the relevant changes since September 16 2016 (15eb9aa) are:
- mlv_rec: RAWC metadata, minor file I/O fix, minor speed tweak (round MLV header to 512 bytes), version metadata
- raw_rec -> mlv_lite
- file I/O backend: fixed an important bug that resulted in invalid video files (like 1 bad video out of a few hundreds)
- raw backend: use the same raw type for all cameras, autodetect resolution rather than hardcoding it, raw capture info (pixel binning factors), preview fixes.

Even when going back to January 2016, the only relevant changes (between f87b759 and 15eb9aa) are:
- mlv_rec: fixed pink frames, wrong black level, crop offsets, added RAWC metadata, refactored powersave, blocked the zoom key while recording
- mlv_snd: persistent settings, show audio meters when recording sound (and hide them when not recording)
- raw backend: black/white level fixes.

That was all on the raw recording side. Notable new features that landed in mainline in this period: Lua scripting, advanced intervalometer, 5D3.123, 700D.115, Makefile updates, various refactors, benchmarks/tests moved to modules, submenu icons, various minor updates.

Given the above, I'm very surprised about your comment ("I don't have this kind of stability anymore."), so please report the issue properly.

BTW, noticed you have been trying the 10/12-bit builds. I hope you are not talking about them - are you aware these builds are not present on the download page, not even as experimental builds, for the 7D? There is a reason for that ;)

Quote
How is the 2.0.6 build?

I have absolutely no idea - the current feedback doesn't allow me to draw a conclusion. Are you aware that we are relying on user feedback?
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on February 24, 2018, 03:41:10 PM
Quote from: domo94 on February 24, 2018, 04:22:26 AM
How is the 2.0.6 build?

Somewhat neglected at the moment due to lack of feedback from users.

Let me give you the back story to the 2.0.6 pull request (https://bitbucket.org/hudson/magic-lantern/pull-requests/878/update-to-7d206/diff). I heard lots of positive comments what a great camera the 7D is and I've got a few friends that own it so I picked up a real bargain on craigslist.org. I've got a 5D3 and comparing these cameras I can say that, well they weigh about the same but otherwise they are completely different beasts. In order to get to know this camera better I thought I'd try to do a firmware update on it. Just dumping the firmware proved to be quite a challenge (https://www.magiclantern.fm/forum/index.php?topic=19417.msg192618#msg192618). So does it work? Yes, there are issues that should be resolved but I have tested it out briefly and it does work. However, a higher priority on this camera should be to get it working with 10/12bit, lossless, crop_rec and all the other recent advances so the firmware update project has moved to the back burner for now.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on February 24, 2018, 06:48:50 PM
Dfort,

I have the 2.0.3 firmware version installed and do not see any reason why I should upgrade to 2.0.6.  I would prefer to get 10/12 bit video working on the 2.0.3 version before we move on with 2.0.6.  If you want some feedback on the latest build that you have been working on, please upload it, post a link and I will be happy to test it and report my findings as usual.   I provided a brief report on your February 04-th build (see my post #1486).  If there is some progress since then or if you want a particular function tested, please let us know.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on February 25, 2018, 12:46:02 AM
Quote from: IDA_ML on February 24, 2018, 06:48:50 PM
...do not see any reason why I should upgrade to 2.0.6....

I posted several reasons in the Porting a Canon firmware update topic (https://www.magiclantern.fm/forum/index.php?topic=19417.0). I've got another reason now because one of my new lenses is on the list of bug fixes for 2.0.6.

Sure, it would be great to get 10/12bit, lossless compression and the crop_rec module working all the way up to 4k. This should all be possible on the 7D. However, I'm stuck trying to figure out why things that are almost working stop working after the configuration file is saved.
Title: Re: dfort's experiments for 7D Mark I
Post by: atomtan on February 28, 2018, 02:59:02 AM
I have searched forum and don't see any others having pointed this issue, so not sure if just me.  A couple years ago when installing ML on my 7DM-I the external audio input worked fine.  Then sometime a while back there was an ML upgrade (sorry, didn't note the build or would have reverted) which seems to have killed the analog audio input levels.  Boot camera without ML, the audio input levels are fine, with ML and they are so very very low that at a minimum I have to bump the analog DB up to maximum within ML but that isn't as clean as getting normal levels to begin with.
What am I missing here why ML external mic audio levels are so through the floor on a 7D Mark-I that the level indicator barely registers? 
Without my Rode video-mic plugged in the built-in camera mic levels are fine and normal.
Yes, yes, I know to use an external audio recorder, which I always do.  BUT, I prefer to use the hotshoe videomic compared to internal camera mic for my reference audio AND as a backup to my audio recorder/shotgun mic, as you sometimes need that decent backup.

I've had to revert to not using Magic Lantern-loaded CF card when doing important interviews, JUST for that decent backup audio.
Any ideas why my ML levels through the floor?
Title: Re: dfort's experiments for 7D Mark I
Post by: a1ex on February 28, 2018, 03:59:33 PM
Older builds are available since late 2013. If the issue was not present in some older build, then builds older than that are probably also good.

However, I remember Trammell mentioning the default configuration with Canon audio (on 5D2, but likely applies to all models) was with pretty high analog gain.

edit: found the old post: http://www.dvinfo.net/forum/canon-eos-full-frame-hd/237675-canon-5d-mark-ii-audio-exposed-beachtek-dxa-5d-juicedlink-cx231.html#post1161097

I can also find the exact audio settings used by Canon (as the linked post refers to old 5D2 firmware, without manual audio controls), but not right now.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on March 03, 2018, 03:38:44 AM
Quote from: a1ex on February 24, 2018, 12:03:14 PM
BTW, noticed you have been trying the 10/12-bit builds. I hope you are not talking about them - are you aware these builds are not present on the download page, not even as experimental builds, for the 7D? There is a reason for that ;)

I can't give any specific instances, but it's definitely just consistency, speed, performance, etc. I can tell and feel the difference from before, but I'll get a bug report.

Yes, I have been on the 10/12. Could be a significant factor. I just head to the download page and all it does is give me nightly. I don't know what's stable without digging through forums.
Title: Re: dfort's experiments for 7D Mark I
Post by: gordongee on March 06, 2018, 04:04:01 AM
i unfortunately installed fw 2.0.6 on my eos 7d and now may be too late found out the magic lantern isnt possible to be installed... but i want to install on my eos 7d ..
as far as i know i cant downgrade or install it as far as i could figure out... if there is a possibility pls tell me

big hugs

Quote from: dfort on February 06, 2018, 04:47:05 AM
Wow, someone is actually using the 2.0.6 version. That was mostly a challenge because I used a 5D3 to do the firmware dump.

https://www.magiclantern.fm/forum/index.php?topic=16534.msg192336#msg192336

This isn't a high priority project but it would be nice to get the latest Canon firmware working with Magic Lantern.

So this is only happening when using an external HDMI monitor? I don't have one to test with so I'll need some help resolving that issue. Not really sure where to start.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on March 06, 2018, 06:15:52 AM
Quote from: gordongee on March 06, 2018, 04:04:01 AMas far as i know i cant downgrade or install it as far as i could figure out... if there is a possibility pls tell me

Just do it.
Title: Re: dfort's experiments for 7D Mark I
Post by: scotophorus on March 06, 2018, 07:34:36 AM
Quote from: gordongee on March 06, 2018, 04:04:01 AM
i unfortunately installed fw 2.0.6 on my eos 7d and now may be too late found out the magic lantern isnt possible to be installed... but i want to install on my eos 7d ..
as far as i know i cant downgrade or install it as far as i could figure out... if there is a possibility pls tell me

big hugs

you can downgrade from 2.0.6 to any firmware, you have to put the firmware file on the root of your cf card and execute firmware upgrade on the camera menu. Here is the link to the 2.0.3 firmware http://pel.hu/down/eos7d-v203-win.zip (http://pel.hu/down/eos7d-v203-win.zip)

Thank you for the info Alex, i've been using dfort's december builds with 10-12bit raw and havent noticed any tearing, probably because of the reduced bandwith as you said. I'm yet to test dfort's feb 28 build for 2.0.6 firmware of the 7d. I'll edit this post as soon as i test that one.

https://bitbucket.org/daniel_fort/magic-lantern/downloads/magiclantern-Nightly.2018Feb28.7D206.zip (https://bitbucket.org/daniel_fort/magic-lantern/downloads/magiclantern-Nightly.2018Feb28.7D206.zip)
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on March 07, 2018, 05:10:09 PM
Pelican's site sometimes goes down (https://www.magiclantern.fm/forum/index.php?topic=19417.msg193999#msg193999) and some files might not download so I've been posting the Canon firmware updaters on my downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/). Just noticed that I didn't have a "7D Canon Firmware for Testers" package so I just put one up. Now you can easily switch between 203 and 206 if you want.

Note that I usually just point to my downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/) instead of individual files. That's because I often delete old files and don't want my posts to contain broken links.

As far as the 7D.206 ML build -- This was a low priority project just to see if I could do it because getting a firmware dump from this camera (https://www.magiclantern.fm/forum/index.php?topic=19417.msg192618#msg192618) was very challenging. It seems to work quite well though it is having a problem passing some tests. If anyone wants to pitch in and figure out what needs fixing, please do!
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on March 07, 2018, 07:07:58 PM
Quote from: IDA_ML on February 09, 2018, 12:48:04 PM
Dfort,

A very brief feedback from first tests with your build:

raw_video_10bit_12bit_LVState-wip.2018Feb04.7D203

Following your instructions, I managed to upload and activate the mlv_lite, mlv_rec and mlv_play modules.  I tested RAW video recording with the MLV_REC module.  And it works - both in the normal uncropped mode at 1736x972 and the 5x-magnification mode at 2496x1198 resolutions. The Preview at 5x-crop is broken and screen freezes when recording starts.  If the RAW_twk module is also activated, the already recorded clips can be played back in the camera.  If an attempt is made to record in the presence of that RAW_twk module, the recorded 1736x976 video has again earthquake shaking.

My overall impression is that in the MLV_rec module, this build works in the same way as the "miracle" December 11-th build.  This is VERY GOOD NEWS since it proves that the 7D is capable of working without earthquake shaking !!!  It also appears to provide quite stable operation in both - the normal uncropped and the 5x-magnification modes at 10 and 12 bits which is EXCELLENT NEWS !!! If you could fix the freezing screen issue at 5x-magnification, this build will be perfectly useable !  Please keep up the good work, Dfort !

IDA_ML,

Would you mind sharing the instructions that you said dfort gave you when using the new firmware: raw_video_10bit_12bit_LVState-wip.2018Feb04.7D203

I tried based on what you mentioned in your post, but I must be missing something. I am getting earthquake shaking in about half of the clips I tested even though I didn't have the RAW_twk model activated. I am testing out 10 bit with the standard resolution of 1736x972 and the 5x-magnification mode at 2496x1198 resolution. Just wanted to make sure I had all the settings right that dfort gave your for this build.

Thanks!

Rob

Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on March 08, 2018, 12:40:49 AM
Quote from: rob_6 on March 07, 2018, 07:07:58 PM
Would you mind sharing the instructions that you said dfort gave you...

The "earthquake" seems to come up only when ML reads the configuration files. Delete the "SETTINGS" folder from the ML directory on your card and chances are that the clips will be fine. Restart the camera and you'll probably get the "earthquake" phenomena. Not sure what is causing that or how to fix it.

Also, make sure not to load the raw_twk module. In fact, it is best to load only mlv_lite and nothing else.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on March 08, 2018, 12:36:19 PM
Rob,

The key to get the 7D working without earthquakes is to disable the RAW_twk module!  Please, read carefully my post 1486 and report back what you found.  I have just followed Dfort's instructions in his post #1483.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on March 08, 2018, 04:42:22 PM
Right, Reply #1483 (https://www.magiclantern.fm/forum/index.php?topic=9848.msg196949#msg196949) -- It is more convenient to reset using the Config menu.

(https://farm5.staticflickr.com/4777/40691123501_9a5e9e2f93.jpg) (https://flic.kr/p/24ZJDmR)

Quote from: IDA_ML on February 09, 2018, 12:48:04 PM
If you could fix the freezing screen issue at 5x-magnification, this build will be perfectly useable !

Interesting thing is that reduced bit depth recording at 5x zoom mode was working before I started tinkering with this. This was discovered when I first started posting builds for users that couldn't compile but wanted to test the early 10bit/12bit development. Now that I look back on this (https://www.magiclantern.fm/forum/index.php?topic=5601.msg174306#msg174306) it didn't make any sense compiling for platforms that didn't support this feature but as it turned out it worked on the 5D2 and other Digic 4 cameras but only in zoom mode.

I think reddeercity came up with a way to get 5x zoom mode working again but as far as I can see he hasn't published his code changes (https://bitbucket.org/reddeercity/).

BTW--I'm sounding like a broken record but if you can follow the simple tutorials to compile ML from source it would really help! You don't need to know how to code, just pull in the changes and build.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on March 08, 2018, 05:20:22 PM
Dfort,

I am sorry, I have been very busy at work lately and this is not going to change for some time. Maybe, some other 7D owner could jump in and spend some time learning how to compile and help.  All I can do for now is that if some new build appears and you notify me with a PM, I will do my best to test it and report back as soon as I can.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on March 08, 2018, 05:31:12 PM
Quote from: IDA_ML on March 08, 2018, 05:20:22 PM
...spend some time learning how to compile...

The tutorials take maybe 15 minutes to complete.
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on March 08, 2018, 06:54:46 PM
Quote from: dfort on March 08, 2018, 05:31:12 PM
The tutorials take maybe 15 minutes to complete.
dfort,

I am willing to help out with compiling. Couple questions. You mentioned it takes 15 minutes to complete tutorials for compiling, which tutorials are your referring to? Is it compiling on a PC? I have a Mac. I saw your thread about compiling on a Mac (https://www.magiclantern.fm/forum/index.php?topic=16012.0) Is that a method that would work for the 7D?

Lastly, could let me know your ideas that I could test when compiling for the 7D? I basically want to start off in the right direction. :)

Thanks!

Rob
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on March 08, 2018, 10:36:08 PM
@rob_6 - Yes, that's the tutorial. Shortcut, copy and paste the "Quick installation script" from the bottom of the first post into your terminal window, run it (just hit enter/return) and it will check your system and install the appropriate tools. One of the first checks is if you have a /usr/local directory that you can write to. If not, the instructions are near the top of the post. It involves temporarily disabling System Integrity Protection (SIP) in order to create a /usr/local you can use. Run the install script again, answer a few prompts, enter your system password and you are good to go. Honestly I have set up several Macs and PCs myself in just a few minutes and have gotten several ML users started on compiling. It is really easy!

Try cloning my repository and compile some of the various branches. I highly recommend the Sourcetree app (https://www.sourcetreeapp.com/) especially when you are first starting out. I keep my repository in sync with the main "hudson (https://bitbucket.org/hudson/magic-lantern)" repository so you won't miss out on anything unless I'm on vacation. Of course it would be best if you also keep a local copy of the hudson repository.

Instead of me having to decide what to build and post for testing you can see the repository being updated in real time and decide if you want to try out the latest changes. It will be obvious from the commit notes and the changes to the code when something interesting for the 7D comes up. Don't be afraid of editing the code. If you really mess up simply delete the local repository and re-clone it from Bitbucket.

Once you tried changing a few lines of code consider creating a free account on Bitbucket and forking the main repository. Start making your own branches, merge branches, create pull requests. You don't need to know much coding to get started. Some of the most useful testing is to simply follow a1ex's instructions. Ok--sometimes it isn't easy to follow his instructions but he has more patience with users that show some initiative and willingness to learn.

At this time the most interesting action for the 7D is in a new branch in my repository named crop_rec_4k_Digic4. The goal is to get as many features in that branch working on Digic 4 cameras. So far none of the new features, lossless compression, reduced bit rate and the crop_rec module are working but it looks like the features that work in the unified branch are still pretty much intact so that's a good sign.
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on March 11, 2018, 09:39:28 PM
Quote from: dfort on March 08, 2018, 10:36:08 PM
@rob_6 - Yes, that's the tutorial.
@dfort,

Thanks or the detailed response. I really appreciate it! :)

Sincerely,

Rob
Title: Re: dfort's experiments for 7D Mark I
Post by: Sapporo on March 19, 2018, 07:33:05 PM
Quote from: dfort on January 18, 2018, 05:41:45 PM
I have never used that module but if I got the addresses right it should work.

https://bitbucket.org/hudson/magic-lantern/pull-requests/878/update-to-7d206/diff#chg-modules/io_crypt/io_crypt.c

[EDIT] The example shows the addresses for 2.0.3 so I updated the example. The values used in the module didn't change with the firmware update.
Updated my 7D to 2.0.6 today. Installed magiclantern-Nightly.2018Feb28.7D206.zip from your https://bitbucket.org/daniel_fort/magic-lantern/downloads/

Downloaded io_crypt https://builds.magiclantern.fm/jenkins/job/io_crypt/2/artifact/modules/io_crypt/io_crypt.mo

When I activate the module I get: "io_crypt: camera unsupported".
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on March 19, 2018, 08:00:30 PM
True:
else if(is_camera("7D", "2.0.3"))
    {
        trace_write(iocrypt_trace_ctx, "io_crypt: Detected 7D");
        iodev_table = 0x2D3B8;
        iodev_ctx = 0x85510;
        iodev_ctx_size = 0x18;


else
    {
        NotifyBox(2000, "io_crypt: Camera unsupported");
        return CBR_RET_ERROR;


dfort? g3gg0?
Title: Re: dfort's experiments for 7D Mark I
Post by: Sapporo on March 19, 2018, 08:14:38 PM
Quote from: Walter Schulz on March 19, 2018, 08:00:30 PM
True:
else if(is_camera("7D", "2.0.3"))
    {
        trace_write(iocrypt_trace_ctx, "io_crypt: Detected 7D");
        iodev_table = 0x2D3B8;
        iodev_ctx = 0x85510;
        iodev_ctx_size = 0x18;


else
    {
        NotifyBox(2000, "io_crypt: Camera unsupported");
        return CBR_RET_ERROR;


dfort? g3gg0?
Not really an issue for me anymore. I updated to use the WFT-E5 I got last week, but it seems WFT-E5 isn't really my thing  :(

I will sell it and I have already downgraded the firmware to 2.0.3 again.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on March 20, 2018, 06:07:02 AM
Sorry for the delayed response. @Sapporo got the io_crypt module from the downloads page so it is specifying 7D.203. The 7D.206 build requires an updated module.

I followed the instructions in the code to find the proper values for the new firmware and turns out that it didn't change. If that's true for all firmware updates it should be possible to change the way we specify cameras like this:

    else if(is_camera("7D", "*"))
    {
        trace_write(iocrypt_trace_ctx, "io_crypt: Detected 7D");
        iodev_table = 0x2D3B8;
        iodev_ctx = 0x85510;
        iodev_ctx_size = 0x18;
    }

Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on March 20, 2018, 08:20:13 AM
IMO it's a design flaw to expect such things *not* to change. Disclaimer: IANAP
Suggestion: Just add another else-if for 2.0.6 and both 2.0.3 and 2.0.6 are covered.
Title: Re: dfort's experiments for 7D Mark I
Post by: Frank7D on March 24, 2018, 06:15:35 AM
I have been away for awhile and am checking in briefly, but I saw a comment about the screen freezing in 5x crop mode, so I thought I'd mention the way I deal with this (when using an old 12-bit raw 7D experimental build): Just before I hit the "START STOP" button to begin a shot, I hit the "INFO" button four times to cycle through the various modes until I'm back at the original mode.
Title: Re: dfort's experiments for 7D Mark I
Post by: Rogan Thoerson on April 10, 2018, 11:07:51 AM
I tried on my 7D it does seem to work. I needed to update MLV producer to decode corectly the 10-bit 2.5k that i made.
I could record 37sec, so it looks like that i am a bit over the speed of the card.
So maybe try 2.2k or 2k might work, still it is really awesome !!!!

I just wonder why is it not yet in the nightly build or if we can add the fuctionality to the Nightly build ?
And is there a possibility one day to make a 8bit raw ? Because it would still be better than the native H264 we have on camera. Or i am wrong ?

I am never programmed such things as camera so maybe it will be a long learning curve to learn such things especially if no tutorial because at this time i only tried to install so i don't know if there is some source / compiler / development program / emulator to test the code in computer before risking the camera...
But if can be any help i would be glad to learn.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on April 22, 2018, 01:12:22 PM
A few days ago, I decided to check again the 5x magnification mode of the old and still my favorite 7D at its maximum 2520x1200 resolution at 10 bits uncompressed.  Here is a short video shot with the December 1-st, 2017 experimental build ("the haunting one") which allows Life View operation for proper subject tracking:

https://vimeo.com/265597058

As you see, this old camera can provide ML RAW video with superb quality in terms of fine detail, dynamic range, transition smoothness, etc.  Due to its extremely fast interface, (mine writes/reads at 90/110 MB/s) and unprecedented battery life, reliability and stability, it still could rival the 5D3 for serious work only at a fraction of its cost.  Used 7D-s in a mint condition can now be obtained at about 300 Euros in Europe.

Dfort put an enormous effort in trying to get the 7D working at 10 and 12 bits uncompressed and even achieved a remarkable success with resolving the "earthquake" issue.  Unfortunately, without much help from the other developers, he could not get very far.  I hope, the above video will encourage Dfort, Nikfreak, A1ex and other developers to continue their efforts towards porting 4K crop recording to the 7D and hope, there will be soon a new experimental build to test on this fantastic camera.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on April 22, 2018, 01:14:58 PM
Quote from: IDA_ML on April 22, 2018, 01:12:22 PM
(mine writes at 110 MB/s)

How? Details, please. Both my cams (one dead) max out very much below and never heard or read about users running such write rate before.
Benchmark snapshot available?
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on April 22, 2018, 10:50:57 PM
Hi IDA_ML,

Where did you get the December 1st, 2017 experimental build? Is it still online? I didn't see it on dfort's download page. Your footage shot with that build looks amazing!

Rob

Quote from: IDA_ML on April 22, 2018, 01:12:22 PM
A few days ago, I decided to check again the 5x magnification mode of the old and still my favorite 7D at its maximum 2520x1200 resolution at 10 bits uncompressed.  Here is a short video shot with the December 1-st, 2017 experimental build ("the haunting one") which allows Life View operation for proper subject tracking:

https://vimeo.com/265597058

As you see, this old camera can provide ML RAW video with superb quality in terms of fine detail, dynamic range, transition smoothness, etc.  Due to its extremely fast interface, (mine writes at 110 MB/s) and unprecedented battery life, reliability and stability, it still could rival the 5D3 for serious work only at a fraction of its cost.  Used 7D-s in a mint condition can now be obtained at about 300 Euros in Europe.

Dfort put an enormous effort in trying to get the 7D working at 10 and 12 bits uncompressed and even achieved a remarkable success with resolving the "earthquake" issue.  Unfortunately, without much help from the other developers, he could not get very far.  I hope, the above video will encourage Dfort, Nikfreak, A1ex and other developers to continue their efforts towards porting 4K crop recording to the 7D and hope, there will be soon a new experimental build to test on this fantastic camera.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on April 22, 2018, 11:46:32 PM
The build is still on my downloads page (https://www.magiclantern.fm/forum/index.php?topic=14739.msg200395#msg200395). Look for the one labeled: raw_video_10bit_12bit_LVState-wip.2018Feb04.7D203.zip

I don't like linking directly to the file because the link will break when the file is updated.
Title: Re: dfort's experiments for 7D Mark I
Post by: rob_6 on April 23, 2018, 01:27:28 AM
Quote from: dfort on April 22, 2018, 11:46:32 PM
The build is still on my downloads page (https://www.magiclantern.fm/forum/index.php?topic=14739.msg200395#msg200395). Look for the one labeled: raw_video_10bit_12bit_LVState-wip.2018Feb04.7D203.zip

I don't like linking directly to the file because the link will break when the file is updated.

Thanks dfort!
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on April 23, 2018, 09:40:51 AM
Quote from: Walter Schulz on April 22, 2018, 01:14:58 PM
How? Details, please. Both my cams (one dead) max out very much below and never heard or read about users running such write rate before.
Benchmark snapshot available?

I performed that test with the Lexar x1066 card quite a while ago but did not bother to make a record of the benchmark test.  I just had the number 110 MB/s in my head.  Now I repeated it again and have to admitt, I have made a small mistake in the above post, meaning the "read" and not the "write" speed.  I appologize for that and have edited my post.  Since I don't have the Lexar any more, I repeated the test with a somewhat slower card (SanDisk x1066) and still get 89/106 MB/s write/read speeds respectively:

(https://thumb.ibb.co/j1QSrx/7_D_San_Di_64_GB.jpg) (https://ibb.co/j1QSrx)


In my opinion, this is still a quite impressive write speed on the 7D and perfectly suitable for 4K crop recording.  Only 5D3 is faster but it has also a larger sensor.  So, I expect, both cameras will be on par.


Title: Re: dfort's experiments for 7D Mark I
Post by: windh on May 12, 2018, 02:44:22 AM
Hi all.

Recently shot some raw material at my local airport with some aircraft landing and taking off. After conforming to .DNG and importing into Resolve, I realize that the audio drops out in every clip. here's a screencap:
(https://thumb.ibb.co/dAeKMJ/Untitled_1.jpg) (https://ibb.co/dAeKMJ)



Nov 2015 nightly build.
Title: Re: dfort's experiments for 7D Mark I
Post by: dcoughlan on May 15, 2018, 06:29:39 AM
Quote from: IDA_ML on April 22, 2018, 01:12:22 PM
A few days ago, I decided to check again the 5x magnification mode of the old and still my favorite 7D at its maximum 2520x1200 resolution at 10 bits uncompressed.  Here is a short video shot with the December 1-st, 2017 experimental build ("the haunting one") which allows Life View operation for proper subject tracking:

Inspired by this footage I just tried recording 2520x1192 / 10bits / FPS override to 23.976 FPS / MLV_LITE, using the build that @dfort linked to below.

The LiveView was fine for framing but froze up when I hit record. To clarify are you able to view the LiveView during recording, or just framing?

FWIW I'm using a slower card (Lexar 1000x), but able to get between :25 and :40 seconds of recording with the above settings. Unfortunately the LiveView consistently locks up after I hit record regardless of the resolution (tested down to 1920x1200). It's been years since I've used ML but I have a vague memory of this happening before so maybe a known issue?

Getting 4K out of this camera would be awesome, so whatever I can do to contribute I'm happy to lend a hand. It's been about a decade since I've done any coding so I doubt I'll be any help on that front, but I've bookmarked dforts compiling tutorial and will get an environment set up on my machine in the next few days. Really impressed with how far it has come already. Hats off to dfort and the rest for their efforts.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on May 15, 2018, 11:54:36 AM
Quote from: dcoughlan on May 15, 2018, 06:29:39 AM
Inspired by this footage I just tried recording 2520x1192 / 10bits / FPS override to 23.976 FPS / MLV_LITE, using the build that @dfort linked to below.

The LiveView was fine for framing but froze up when I hit record. To clarify are you able to view the LiveView during recording, or just framing?

FWIW I'm using a slower card (Lexar 1000x), but able to get between :25 and :40 seconds of recording with the above settings. Unfortunately the LiveView consistently locks up after I hit record regardless of the resolution (tested down to 1920x1200). It's been years since I've used ML but I have a vague memory of this happening before so maybe a known issue?

Getting 4K out of this camera would be awesome, so whatever I can do to contribute I'm happy to lend a hand. It's been about a decade since I've done any coding so I doubt I'll be any help on that front, but I've bookmarked dforts compiling tutorial and will get an environment set up on my machine in the next few days. Really impressed with how far it has come already. Hats off to dfort and the rest for their efforts.

Dcoughlan,

There is an even older build that can unlock the frozen Live View screen during 10/12-bit recording at 2520x1200 resolution on the 7D.  I have been using it a lot lately because it is very stable and easy to work with and provides excellent results.  That build is called:

magiclantern-10bit_12bit_raw_twk_crop_rec.2016Dec01.7D203

You can find it when you scroll down the following repository:

https://bitbucket.org/reddeercity/magic-lantern_10-12bit/downloads/

To unlock the Life View please do the following:

1) Press 5x-magnification and set "RAW-video" to 2520x1200 resolution and 10-bit

2) Focus and adjust your exposure.  Use Global draw On.  Activate "Zebras" to avoid overexposure in the highlights

3) While at 5x-magnification, press the "trash" button, go to "RAW-video" and then press "Q".  This will bring the submenu of "RAW-video"

4) Go to "Preview" and press "Canon", then change to "Auto".  Pressing half shutter will show you a frozen Life View screen

5) Press the "trash" button again and then the "Q"-button.  Now switch the preview mode again to "Canon" and press half shutter again.  This unlocks the Life View screen and you can start recording while enjoying a 5x-magnified composition during recording.  If you turn also Global Draw OFF before you start recording, this will give you a substantially longer record time.  I get 45 sec. on a x1000 card at 2520x1200 resolution at 24 fps.

If you need sound with 10-bit RAW video, you may switch to "RAW video (MLV)".  In this case, your maximum resolution will be 2496x1200.  Also you can film at 1728x972 14-bit with sound in the Normal uncropped mode.  10 and 12-bit recording in that mode will provide earthquake shaking.

==============================================

Your kind offer to help with further development of ML for the 7D is God sent and I am sure, Dfort will greatly appreciate it since he badly needs someone with knowledge of compiling to continue his work.  With your help, I am sure, we will soon see 4K-crop recording on the fantastic 7D.






Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on May 15, 2018, 06:24:42 PM
@dcoughlan - If you want to play around with my experiments you are more than welcome. If you want to tweak my experimental builds that are still being hailed as "magical" you can open the autoexec.bin file with a text editor to see where it came from. For example the magiclantern-10bit_12bit_raw_twk_crop_rec.2016Dec01.7D203 on reddeercity's download directory looks like this in a text editor:

Magic Lantern Nightly.2016Dec01.7D203
Camera   : 7D
Firmware : 203
Changeset: b3dfbe7194f3 (x-perimental) tip
Built on : 2016-12-01 23:00:07 by rosiefort@RosieFoComputer


If you clone my repository -- https://bitbucket.org/daniel_fort/magic-lantern -- and use hg to find that changeset "hg update b3dfbe7194f3" you can branch off from there and see if you can get all the pieces working.

Good luck!
Title: Re: dfort's experiments for 7D Mark I
Post by: dcoughlan on May 16, 2018, 05:36:49 AM
Excellent, thank you both. As I mentioned I'm a little rusty so it'll take me some time to wrap my head around everything but I hope to be able to contribute to the efforts soon!
Title: Re: dfort's experiments for 7D Mark I
Post by: masc on June 13, 2018, 05:44:13 PM
Hej 7D owners, could you please help me?!
-> https://www.magiclantern.fm/forum/index.php?topic=20025.msg202673#msg202673
Title: Re: dfort's experiments for 7D Mark I
Post by: MateX on July 02, 2018, 02:50:48 AM
Hello im new i want to install magic lantern but i have got me 7D with the 2.0.6 frimware so can i downgrade to 2.0.3
Title: Re: dfort's experiments for 7D Mark I
Post by: Sapporo on July 02, 2018, 05:41:21 PM
Quote from: MateX on July 02, 2018, 02:50:48 AM
Hello im new i want to install magic lantern but i have got me 7D with the 2.0.6 frimware so can i downgrade to 2.0.3
For 2.0.6 https://bitbucket.org/daniel_fort/magic-lantern/downloads/magiclantern-Nightly.2018Feb28.7D206.zip

Or downgrade to 2.0.3. http://pel.hu/down/
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on July 02, 2018, 09:12:13 PM
Quote from: Sapporo on July 02, 2018, 05:41:21 PM
Or downgrade to 2.0.3. http://pel.hu/down/

If Pelican's site is down or hard to figure out I also have the 2.0.3 and 2.0.6 updates on my Bitbucket downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/). Look for the file named 7D Canon Firmware for Testers.zip.
Title: Re: dfort's experiments for 7D Mark I
Post by: mister.rings on July 11, 2018, 03:05:29 PM
Hello everyone.
sorry, I downgraded to 203 thinking that the magic lantern worked more, but not changing ... the menu does not change .... it's all the same ... but you can make video without interruption with the 7d and ML?
Thank you.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on July 11, 2018, 08:11:39 PM
Top of page -> User Guide -> FAQ/User Guide/Install Guide

H.264 recording lasts 29:59 and ML is able to restart recording but loosing some frames every 29:59.
MLV/raw recording is seamless.
Title: Re: dfort's experiments for 7D Mark I
Post by: dj_FoXX on September 13, 2018, 09:01:42 PM

Hello! sorry for my english, this is all Google translator) I have 7d, installed ML, on the first attempt, no problems, the firmware works. I use 7d only for photography, and it would be interesting to take part in testing and optimizing the firmware.

I found the lens adjustment function (+/-), did not understand how to use it.

I select DOTtune, I switch the lens to manual focus, I press play, the screen flashes several times, showing info, and the message "press INFO and tray again" appears.
OK. I go to LV, zoom 10x, focus in manual mode, than dot tune, press play - and the same "press INFO and tray again".

that I would like to add:
- autofocus illumination (as in nikon)
- confirmation of focusing on lenses without a chip.

and the last question, just installed ML, without changing the settings, does it affect any algorithms? (autofocus, exposure metering, photo processing, etc.)
Title: Re: dfort's experiments for 7D Mark I
Post by: BrokenSyntax on October 05, 2018, 12:44:56 AM
I have done some poking around on dfort's fork. Is the crop_rec_4k_5D2_7D branch the best one for progress? I'm keen to see if I can spend some hours seeing how to tackle the problems. So far I have things building which is relatively painless thanks to all the resources available.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on October 05, 2018, 10:06:43 AM
Quote from: BrokenSyntax on October 05, 2018, 12:44:56 AM
I have done some poking around on dfort's fork. Is the crop_rec_4k_5D2_7D branch the best one for progress? I'm keen to see if I can spend some hours seeing how to tackle the problems. So far I have things building which is relatively painless thanks to all the resources available.

BrokenSyntax,

If you have coding skills and are willing to help with 7D crop recording development, it might be a good idea to send a PM to Dfort and offer him your help.  I am sure, he will greatly appreciate that and will be able to give you some useful directions on how to proceed and what to do so that your help is efficient.  I personally am very excited about every step forward with the 7D because of its high potential to become the second best crop-recording camera after the 5D3.  And this is not only because of the very high write speed of 90 MB/s but also because of the unique film look that its sensor provides.  I have no coding background and skills whatsoever but would be glad to test and provide feedback on every new 7D build that you guys may be releasing.  I keep my thumbs pressed that you succeed.
Title: Re: dfort's experiments for 7D Mark I
Post by: BrokenSyntax on October 05, 2018, 03:35:43 PM
Quote from: IDA_ML on October 05, 2018, 10:06:43 AM
If you have coding skills and are willing to help with 7D crop recording development, it might be a good idea to send a PM to Dfort and offer him your help.  I am sure, he will greatly appreciate that and will be able to give you some useful directions on how to proceed and what to do so that your help is efficient.

I have just now done that, thanks.

I think to get lossless working correctly I need to have a deep dive into the silent module and figure out a way to get the lossless compression working there as that seems a more sensible place to start for me being relatively new to the codebase. At a glance it definitely looks like if I make progress with the silent module (by extension lossless logic) that the lossless 14-bit compression should work on the 7D with a few minimal tweaks. Time will tell, I'll have a stab at it over the weekend as I am very eager to breathe new life into the 7D I picked up yesterday.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 05, 2018, 09:26:48 PM
Quote from: BrokenSyntax on October 05, 2018, 03:35:43 PM
I think to get lossless working correctly...

Right -- these are the changes I was playing around with:

https://bitbucket.org/daniel_fort/magic-lantern/pull-requests/25/crop-rec-4k-5d2-7d/diff#chg-modules/silent/lossless.c

Not quite working yet.
Title: Re: dfort's experiments for 7D Mark I
Post by: r_non on October 31, 2018, 10:00:24 PM
Hi all, can somebody show me how to contribute to Magic Lantern development in general? I've got no coding knowledge (unless you count Matlab as one) but would like to support as much as possible.

For a start, I'm currently following this guide at https://www.magiclantern.fm/forum/index.php?topic=6783.0

Am i in the right direction? Or should be using cygwin/mingw approach https://www.magiclantern.fm/forum/index.php?topic=15894.0

Yea it would be cool to see 10 bit raw video on the 7d but is anyone still working on it? Which project is currently being focused on, maybe i can start there?

Many, many thanks!
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on November 01, 2018, 02:55:26 PM
The Cygwin tutorial  (https://www.magiclantern.fm/forum/index.php?topic=15894.0)should be the easiest to get you up and running on any version of Windows. You can't use QEMU with it but the 7D and QEMU aren't playing well together so that's not a deal breaker.

Setting up a Linux environment will give you a more feature rich development environment. If you are running Windows 10 you don't need to run a virtual Linux box because it has an embedded Linux subsystem (https://www.magiclantern.fm/forum/index.php?topic=20214.0).

A few of us are working on the 7D but there are very few of us and development isn't keeping up with other cameras from that era like the 5D2. Having a few more users willing to jump in and try to get some new features working on this camera would be greatly appreciated.
Title: Re: dfort's experiments for 7D Mark I
Post by: beanz on November 17, 2018, 01:18:46 PM

(https://thumb.ibb.co/cGyfZf/IMG-20181117-135124.jpg) (https://ibb.co/cGyfZf)

Has anyone come across this before? (See image). After recently upgrading to 2.0.6 and doing some successful tests I put my camera away. Today when I turned it on it was displaying multiple module fails. I turned them all off, restarted, turned them on again, restarted and everything is back to normal. Thanks
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 17, 2018, 01:32:00 PM
mlv_snd.mo does require mlv_rec.mo to be loaded. If not present this screen will be shown.
Unclear why mlv_rec.mo failed to load.

BTW: Why loading mlv_lite and mlv_rec together?
Title: Re: dfort's experiments for 7D Mark I
Post by: beanz on November 17, 2018, 01:40:44 PM
Quote from: Walter Schulz on November 17, 2018, 01:32:00 PM
mlv_snd.mo does require mlv_rec.mo to be loaded. If not present this screen will be shown.

BTW: Why loading mlv_lite and mlv_rec together?
Hi Walter, not sure why that was on as well, I never use it and must have loaded it by accident. It's now off. I've loaded mlv_snd and she's back to normal. Cheers man
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on November 23, 2018, 11:01:19 PM
You are on 2.0.6?

That is not mainstream yet. If you are using it please let us know how well it is working out for you.


Sent from my iPhone using Tapatalk
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on November 24, 2018, 12:15:18 PM
So I'm not too sure, simply because I've been so mindless (so to speak) in using ML, but I've never paid attention to the card format because on my PC, it formats automatically to exFAT, but when I do the camera, it reverts to FAT32. Is that normal? Are we able to use the 7D with an exFAT formatted card?

I had a gig today that was absolutely horrendous for me on a technical standpoint - I had interviews, I did them raw, but apparently ML kept recording til I hit stop, but my card wrote only the 4GB limit. So now I'm stuck with incomplete interviews. Is there any possible way to get the 7D working with an exFAT format?

I don't recall ever being capped out with my RAW recordings on ML. I just filmed a bunch of footage back in June in Costa Rica and I don't remember having this issue & I'm sure as all hell I recorded way over a 1:18 time frame. (That's what it capped out today).

128GB LEXAR 800x CF CARD.

I'm going to upgrade to 2.0.6 firmware to be more attentive and critical of the performance of ML in order to help development. Please bug me if I'm slow. @lifestyledq is my handle on everything you'll find me on.
I will also upload some footage of the 7d, 14 bit RAW, with the Anti-aliasing filter bought from MOSAIC Engineering because I had told a thread in this forum that I would do so, but never actually did so.

Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 24, 2018, 12:22:47 PM
Quote from: domo94 on November 24, 2018, 12:15:18 PMbut when I do the camera, it reverts to FAT32. Is that normal? Are we able to use the 7D with an exFAT formatted card?
https://www.magiclantern.fm/forum/index.php?topic=23064.msg208567#msg208567

Quote from: domo94 on November 24, 2018, 12:15:18 PM
but apparently ML kept recording til I hit stop, but my card wrote only the 4GB limit. So now I'm stuck with incomplete interviews.

Please check for files <Filename>.M00, M01, ... on card.
MLV recording on 7D uses file chunks up to file system limit.

If they are not there, please respond with build number, modules used.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on November 24, 2018, 01:05:43 PM
Quote from: Walter Schulz on November 24, 2018, 12:22:47 PM
https://www.magiclantern.fm/forum/index.php?topic=23064.msg208567#msg208567

Please check for files <Filename>.M00, M01, ... on card.
MLV recording on 7D uses file chunks up to file system limit.

If they are not there, please respond with build number, modules used.

Ah, yes, that's what it was! Thank you!
Maybe I glitched out today. Not sure.
I knew I could record more, I was so surprised to see my videos cut to the max today.

As for the reliability... I don't know anymore, to be honest. I recorded a solid 40 seconds of footage, and the next few minutes, I had about 5-7 seconds of footage before 'Frame skipped, stopping recording"

I just did the 2.0.6 upgrade. Running some basic tests and seeing how well it performs. Maybe it's a placebo, but I feel confident about this. dfort has done great work.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on November 24, 2018, 06:08:56 PM
 8)

Make sure you also comment on the pull request (https://bitbucket.org/hudson/magic-lantern/pull-requests/878/update-to-7d206/diff).
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on November 24, 2018, 07:35:37 PM
Quote from: dfort on November 24, 2018, 06:08:56 PM
8)

Make sure you also comment on the pull request (https://bitbucket.org/hudson/magic-lantern/pull-requests/878/update-to-7d206/diff).

I have a lot of extra time this weekend to learn more about this part to assist you.
So far 2.0.6 is pretty stable. Just getting a glitchy overlay when I switch between menus, live feed, playback, etc.
When I'm focusing, the overlay kind of glitch out and have a bunch of gray pixels everywhere.

I was curious, is there a possible way to keep ratio display with ML overlay (ex. I film 2.20:1 / 2.35:1 often & I need the framing area to frame properly) while disabling Focus Peaking when recording? I notice when I process my raw footage, I think I've created hot pixels on my camera, which can be fixed no problem, as well as getting stuck focused peaked pixels.

I'd like to lock my focus in with my focus peaking, but not have it appear when filming so that way I retrieve clean footage.
How much of a work around is that?
The option to disable global draw is there, but I need my framing if I'm gonna film 2.20:1 ratio. Maybe add a feature to focus peaking to turn off whilst recording?

How much of a task is that?

EDIT
*I have realized I have forgotten how to use Magic Lantern and completely overlook the simple things. Haha. Nevermind!*
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on November 25, 2018, 01:06:19 AM

(https://i.ibb.co/x339Jgz/image.png) (https://ibb.co/x339Jgz)

I shot this clip in 14 bit raw, high ISO at 6400 I believe, WITH NOISE REDUCTION in post, exported at ProRes 422.

Sorry I'm posting so much, but is my camera sensor bad? Is it supposed to look like this? You can see those lines in the video that look like they are BAKED on the sensor.

I've done long exposure low light photography and I get hot pixels and just a bad photo overall. Or is the 7D Low light performance just not that good?

This shot is before post noise reduction, same clip.

(https://i.ibb.co/S6HFSrp/image.png) (https://ibb.co/S6HFSrp)
Title: Re: dfort's experiments for 7D Mark I
Post by: nickxc on December 08, 2018, 12:19:32 AM
Quote from: domo94 on November 25, 2018, 01:06:19 AM

(https://i.ibb.co/x339Jgz/image.png) (https://ibb.co/x339Jgz)

I shot this clip in 14 bit raw, high ISO at 6400 I believe, WITH NOISE REDUCTION in post, exported at ProRes 422.

Sorry I'm posting so much, but is my camera sensor bad? Is it supposed to look like this? You can see those lines in the video that look like they are BAKED on the sensor.

I've done long exposure low light photography and I get hot pixels and just a bad photo overall. Or is the 7D Low light performance just not that good?

This shot is before post noise reduction, same clip.

(https://i.ibb.co/S6HFSrp/image.png) (https://ibb.co/S6HFSrp)


That noise reduction is actually really impressive. The Canon 7D has pretty bad low light performance. Those lines you are seeing is fix patterned noise and is totally normal.

personally I try to never shoot above 200iso. At 400iso I start getting hot pixels and the image noise gets really bad from then on.

your camerasensor is fine :)
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on December 13, 2018, 07:28:46 AM
Quote from: nickxc on December 08, 2018, 12:19:32 AM
That noise reduction is actually really impressive. The Canon 7D has pretty bad low light performance. Those lines you are seeing is fix patterned noise and is totally normal.

personally I try to never shoot above 200iso. At 400iso I start getting hot pixels and the image noise gets really bad from then on.

your camerasensor is fine :)

Thanks, I'm feel better now with 3rd party validation, reference. Thank you! It's just pretty heavy denoising with neat video. Well worth the money, to be honest. I paid for like the multiple keys for Premiere/ After Effects.

Also, this is for @dfort - I am on firmware 2.0.6 now and I am running pretty stable, I have all the essential features I would like out of Magic Lantern to keep my camera worthy and usable for everything I do.

Couple issues I have - when recording, some of the focus marks and overlays(?) tend to freeze on the screen & stay there which is very ugly and distracting while I film, but it doesn't burn into the actual video, which is good.

Second, I was preparing to do full 14-bit raw, up to 12 minute long takes at 1600x900. The camera got that far. No problem. It was a huge file, which was expected. Whilst building the index files, it was too much for the camera I noticed. But even building OTHER index files for smaller shots, my camera glitched out and gave me an ugly purple/pink distortion.
(https://i.ibb.co/stX7xQT/20181203-224102.jpg) (https://ibb.co/stX7xQT)


This lasted until I reboot the camera, of course. But for future instances, I don't want to have to rely on rebooting the camera in case I do.. say 5-6-7 minute or more longer takes and I hit the playback button. This would mean I would NEED to have MLV play module OFF which is an extra hassle/step. I don't mind it, but I don't prefer it, either.

I also ran some tests when I upgraded to 2.0.6 and some passed, some didn't?, the camera just went into a nonstop test mode- it was for benchmarking, I believe. It was some time ago.
You tell me what I need to do for you so you can work your magic.
Thanks!
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on December 13, 2018, 08:38:09 AM
Wow, you're giving 2.0.6 quite a workout.

Quote from: domo94 on December 13, 2018, 07:28:46 AM
when recording, some of the focus marks and overlays(?) tend to freeze on the screen & stay there which is very ugly and distracting while I film, but it doesn't burn into the actual video, which is good.

Have you tried turning off Global Draw? That tends to clear up problems like this.

Quote from: domo94 on December 13, 2018, 07:28:46 AM
Second, I was preparing to do full 14-bit raw, up to 12 minute long takes at 1600x900. The camera got that far. No problem. It was a huge file, which was expected. Whilst building the index files, it was too much for the camera I noticed....

You are running mlv_play? I haven't tested that but when you come up with issues like this try the 2.0.3 version and see if this is a new bug.

Quote from: domo94 on December 13, 2018, 07:28:46 AM
I also ran some tests when I upgraded to 2.0.6 and some passed, some didn't?

Of course it would help to be more specific but the previous comment on running the tests on 2.0.3 also apply. I believe that there are some tests that this camera can't complete. Also note that this is the one camera that doesn't play well in QEMU so there are no tests on the build server.

(https://farm5.staticflickr.com/4873/44479706410_1b48571fbf_n.jpg) (https://flic.kr/p/2aLw84Q)
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on December 14, 2018, 01:16:38 AM
I need this to make my camera magic.
I'm stubborn, super experimental, and it's my only camera that makes me produce wonders for my work / business.

I do photo/video and for photography, Canon takes the cake, everytime.

I might switch to 5d mk iii in the near future for a full frame/low light upgrade, but it'll be some time because I've invested so much into this camera already. But using proper lighting and Neat Video saves my image almost every time, so I make do!

PLUS THESE COLORS ARE TO DIE FOR. Film lives on forever!

I use global draw for focus assist because I just can't seem the nail the focus anymore manually without it. I'm not sure if it's my lens or something, but even the auto focus doesn't nail it properly sometimes, either.

I know it won't happen with Global Draw, I believe. I need to run some more tests.

Right now I bought a dummy battery and am building a battery bank rig for my camera so I can have a very long battery life.
I will update you all on how it goes, but so far it's working solid. Building it is the biggest problem.

I am running MLV_play but I think it was just because it was about 6-9 index files it had to build at the time and it was just WAY too much for the processor to handle.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on December 15, 2018, 07:00:18 PM
Update on the 7d - 2.0.6.

Still pretty stable. It seems every now and then I need to let the camera warm up with a failed clip if I'm shooting 16:9 at 1600x900 for consistent 23.976fps RAW 14 bit.
I'll get the first 7 seconds to fail then the rest of the filming works just fine.

I would still like the option to turn OFF focus pixels ONLY when filming and not JUST global draw because I normally shoot 2.20:1/2.35:1 ratio at max resolution available for 23.976ps. I still want to see the crop marks, audio levels, and framing & I would like to know what's in focus for me PRIOR to shooting and then NOT have the focus pixels show up when filming because when I process the footage. I don't want the pixels showing up and I don't want to have to worry about that.

I just bought a dummy battery for the 7D and connected it to a portable 16750 mAh battery bank at 4.5A output for consistent power. It works, just not as good as having a battery INSIDE the camera. I took a photo whilst filming & in liveview for video mode and the camera actually turns off, glitches out, the mirror lock up gets stuck as if it's trying to flip up, but it's not, so it's like a stuck motor.

Very scary moment because it feels like the camera is about to just flat out break, but it didn't.

It's similar to an elevator door trying to close but there's something in the path of the doors but it won't close and it'll make multiple attempts to do so.

I might go back to 2.0.3 for other tests and to see how well it works with a battery bank. We'll see!
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on December 17, 2018, 05:40:14 AM
@dfort - Hey, I have an update for today.

I powered my camera after the long night of shooting with the external battery, I reconnected a regular battery because I was shooting on a glidecam. I had my camera crash many a times & almost give me a heart attack because I was shooting an entire car show today.

I had left the Canon displays on before I inserted the card the night before, then inserted the card and switch to Magic Lantern overlay & this happened. I also couldn't take a photo at one point & review. 3 different incidents.

Dropping the crash logs here -
Crash 1
Quote
ASSERT: 0
at Memory\Memory.c:566, task PropMgr
lv:1 mode:3

PropMgr stack: 112090 [112118-111118]
0xUNKNOWN  @ ff084c98:112110
0xFF20C168 @ ff20c224:1120e8
0xFF07D088 @ ff20c1e8:1120c8
0x1F800478 @ 1f8007d0:112090

Magic Lantern version : Nightly.2018Jul04.7D206
Mercurial changeset   : 053a73548de5 (update_to_7D.206)
Built on 2018-07-04 21:36:26 UTC by rosiefort@RosieFoComputer.
Free Memory  : 340K + 1484K

Crash 2
QuoteASSERT: 0
at Memory\Memory.c:566, task DbgMgr
lv:1 mode:3

DbgMgr stack: 10bd68 [10c4f0-10b4f0]
0xUNKNOWN  @ ff084c98:10c4e8
0xFF20C168 @ ff20c224:10c4c0
0xFF07D088 @ ff20c1e8:10c4a0
0xFF067F80 @ ff011cac:10c480
0xFF0689E0 @ ff067ffc:10c460
0xFF20C3E8 @ ff0689f4:10c458
0xFF07D088 @ ff20c43c:10c438
0xFF067F80 @ ff011cac:10c418
0xFF0689E0 @ ff067ffc:10c3f8
0xFF20C3E8 @ ff0689f4:10c3f0
0xFF07D088 @ ff20c43c:10c3d0
0xFF067F80 @ ff011cac:10c3b0
0xFF0689E0 @ ff067ffc:10c390
0xFF20C3E8 @ ff0689f4:10c388
0xFF07D08Magic Lantern version : Nightly.2018Jul04.7D206
Mercurial changeset   : 053a73548de5 (update_to_7D.206)
Built on 2018-07-04 21:36:26 UTC by rosiefort@RosieFoComputer.
Free Memory  : 340K + 1434K

Crash 3
QuoteASSERT: 0
at Memory\Memory.c:566, task clock_task
lv:1 mode:3

clock_task stack: 176c28 [176d58-174d58]
0x1F82EF28 @ 1f824e64:176d40
0x1F81CE98 @ 1f82ef70:176d30
0x1F852658 @ 1f81cff8:176ca8
0xFF0689E0 @ ff067ffc:176c88
0xFF20C3E8 @ ff0689f4:176c80
0xFF07D088 @ ff20c43c:176c60
0x1F800478 @ 1f8007d0:176c28

Magic Lantern version : Nightly.2018Jul04.7D206
Mercurial changeset   : 053a73548de5 (update_to_7D.206)
Built on 2018-07-04 21:36:26 UTC by rosiefort@RosieFoComputer.
Free Memory  : 340K + 1428K

What do you gather from this? Let me know if you need to try to recreate anything.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on December 18, 2018, 09:05:34 AM
So it was working fine when you were messing around with the camera and when you had something important to shoot it borked? Yeah, that sounds about right. Seems like it is having issues getting through the startup tasks.

Try resetting the settings. Either through the ML menu or delete the ML/SETTINGS directory from the card. You'll have to redo your settings but that usually clears up problems like this. Since the issue came up when switching from external to internal batteries, try a battery pull. Doing a battery pull is needed after a hard crash where the camera won't turn back on but sometimes there are more subtle issues that can sometimes be resolved with a battery  pull.

Then again it could be none of the above. Check to see if the camera works fine without loading ML. Maybe reset the Canon settings if the problems persist.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on December 18, 2018, 08:20:41 PM
Hey dfort - yes I did all the above.

I got it working again.
I had to do MANY battery pulls.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on December 18, 2018, 10:24:05 PM
Quote from: domo94 on December 18, 2018, 08:20:41 PM
I had to do MANY battery pulls.

That's also true when experimenting around with 2.0.3.

So is 2.0.6 stable now?
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on December 19, 2018, 12:58:07 AM
The break was on 2.0.6 but honestly, it's 8/10 stable.

Would've given 9 but that crash kinda happened, so I'm not going to be lenient with it especially if it means I'm telling people to move over from 203 to 206

I'm using it for the daily now and I just went back to 2.0.3 for 10-12 bit raw tests but I don't like needing to 5x zoom and all to record raw, let alone for slo-mo and all. So I stick to low res 14 bit because it's high quality ANYWAYS.

I give 2.0.6 a green light but only for 14 bit RAW and every other reason to use ML. But reliability is more important to me than anything!
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on December 19, 2018, 02:01:23 AM
Quote from: domo94 on December 19, 2018, 12:58:07 AM
...I just went back to 2.0.3 for 10-12 bit raw tests...

That's another experiment that has issues.

What I'm interested in is you could compare apples to apples -- the nightly 7D.203 vs. the nightly 7D.206. Any differences you found?

Note that my 2.0.6 port has a few pieces that I wasn't able to update, mainly due to not having access to the master processor.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on December 19, 2018, 02:39:31 AM
For the most part, which I'd have to do more extensive testing & casual shoots to see what the camera performs like - was the responsiveness.

The 2.0.3 build seemed very smooth & responsive, almost little to no lag WHERE EVER I was flipping through the menus.

2.0.6 seems to slow down at times, lag a bit, freeze for a moment or so then catch up, not to mention the earlier issue I mentioned about global draw specs getting kind of stuck on the screen when filming and changing up a bit every now and then. Let me recreate that issue tonight & upload what I mean. I never had that before in 2.0.3 that I ever remember.

I think with more love, 2.0.6 would be perfect and just as stable.

I just invited a friend who was paid 6 figures by the school because of his expertise in coding to join the Magic Lantern development train and help us develop this further. I wanted him to spend his time with me somehow since he endured a terrible accident a year ago and kind of does nothing now. I need to catch him up to speed on what to do for this.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on December 19, 2018, 06:23:03 AM
New tests & experimentation lead me to new issues on the 2.0.6.
NOT a game changer, just things to note.

I left a little video there to explain what's going on...

But basically 720p 60fps with FPS OVERRIDE kinda just messes up the image & everything when shooting RAW 14 bit.

https://drive.google.com/open?id=1-ccS7vwlAXpAzo6Lu5dZKGE_gQef6hzq

All settings and information is here.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on January 25, 2019, 08:49:38 AM
May I ask 7D users trying to reproduce a bug? AF lens required and about 3 minutes. Tripod optional but helpful.

====
Start cam in photo mode
Make sure to have ONE SHOT set in Canon menu.
Make sure to have AF selected on lens.

Enter ML menu -> Shoot -> Shoot preferences -> Use Autofocus ON

Then Shoot -> Motion Detect -> Q: Settings
Trigger by Expo. Change
Trigger level 8
Detect Size medium
Delay OFF

Leave menu with Q and press Set -> Motion Detect ON = EXP, level=8
Point cam to something it can focus.
Leave menu by half shutter press. -> Cam goes in Liveview mode
Wait about 5 seconds
Now cover lens with something opaque (lens cap or cardboard) and give cam some time to find out it cannot focus.
Now remove cap/cardboard.

What my cam (7D.206 with EF-S 10-18 STM) does: After focus is found ... shoot like crazy! Have to turn it off!
Please test and report. I'm unable to test 2.0.3 right now.

Other users with Digic IV cams are invited to test it, too.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on January 26, 2019, 05:13:54 AM
Interesting test. I'll check it out with 2.0.3 and 2.0.6 when I get 3 minutes -- uh, maybe it will take 6 minutes?
Title: Re: dfort's experiments for 7D Mark I
Post by: critix on January 26, 2019, 07:02:54 PM
OK, I did the test on 500D  ... Unfortunately, after the focus is found, the camera shoots like a crazy ...
Just like in your description. So this camera does not work well either.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on January 26, 2019, 07:10:37 PM
Thanks for testing!
Which lens was mounted?

I suppose all D4 cams act the same way.
D5 cams are not working well either but they do not go into crazy mode ...

Next bug report approaching ...
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on January 26, 2019, 07:22:25 PM
Did the test on the 7D in both 2.0.3 and 2.0.6 and could not reproduce the issue. This isn't a feature that I'm familiar with but it seems to me that it doesn't work properly. Sometimes a change in exposure would trigger it, sometimes not. I couldn't reproduce the shoots like crazy issue.

Maybe the lens has something to do with it? I used an old Canon EF 28-105mm f/3.5-4.5 USM lens and a new Canon EF-S 18-135mm f/3.5-5.6 IS lens (that's the one the 2.0.6 firmware "fixes") with the same results.

Maybe try resetting to Canon factory default settings and try it again? Sometimes that makes a difference.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on January 26, 2019, 07:25:52 PM
Which 18-135? Nano-USM or STM?
After lens covered you have to give the cam time enough to give up on focus. If it won't start focus (which may happen) when darkened bug will not occur.

Just repeated the test with EF 100/2.8 Macro USM (without L) and 7D.
After 2 or 3 retries it went crazy. It's not the lens, no.

And 650D is affected by the bug, too. After loosing AF once it simply won't trigger focus motor anymore.
Title: Re: dfort's experiments for 7D Mark I
Post by: critix on January 26, 2019, 07:41:20 PM
Quote from: Walter Schulz on January 26, 2019, 07:10:37 PM
Thanks for testing!
Which lens was mounted?
I use Canon 18-55mm STM.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on January 26, 2019, 07:54:24 PM
Quote from: Walter Schulz on January 26, 2019, 07:25:52 PM
Which 18-135? Nano-USM or STM?

The Nano-USM. Though it looks like the lens doesn't make a difference.

Quote from: Walter Schulz on January 26, 2019, 07:25:52 PM
After 2 or 3 retries it went crazy. It's not the lens, no.

And 650D is affected by the bug, too. After loosing AF once it simply won't trigger focus motor anymore.

Ok--good to know. So it probably affects other cameras and it isn't always triggered so we have to keep trying to reproduce the bug.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on January 26, 2019, 09:07:18 PM
Bug report:
https://bitbucket.org/hudson/magic-lantern/issues/2923/feature-use-autofocus-breaks
Title: Re: dfort's experiments for 7D Mark I
Post by: RhinoNelson on January 27, 2019, 06:36:51 PM
I tested with my 7D and a 55-250 IS STM and the shutter went crazy just as you experienced. Using 2.03 July 3, 2018.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on January 27, 2019, 06:42:54 PM
Thanks for jumping in and verifying this bug!
Title: Re: dfort's experiments for 7D Mark I
Post by: Temos01 on February 01, 2019, 08:01:04 PM
Having some issues shooting stills using 2.0.6 and dforts build.  Camera freezes on the preview after shooting an image.  Have to open the batt door and pop battery out to get the camera to shut down.

any suggestions?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on February 03, 2019, 05:56:14 PM
If freeze is reproducable:
Document steps to reproduce the issue and create an issue report.
https://bitbucket.org/daniel_fort/magic-lantern/issues/new
Cream on top: Test this procedure in 2.0.3
If it happens in 2.0.3, too:
Create a bug report here https://bitbucket.org/hudson/magic-lantern/issues?status=new&status=open instead. After checking if it is a new and not reported earlier.
Title: Re: dfort's experiments for 7D Mark I
Post by: skitron on May 14, 2019, 05:08:38 AM
Quote from: dfort on May 13, 2019, 08:21:50 AM
You are using one of my experimental builds for the 7D. Isn't that proof that there is still some development going on for this camera? Also note that when any branch gets updated, the 7D will probably also be updated.

Thanks so much for sharing your 7D builds! In the FWIW dept. here are some testing findings (...2019Mar22.7D203 build, global draw off):

I can't get a clean "stop record" (in crop rec mode, any res). The record icon stays on the screen and I have to exit live view, reboot camera. But the files are good. 2496x1198 @24fps 10 bit no problem.

In non-crop mode press record locks camera. Have to R&R battery to clear.

Cannot get mlv_play to load at boot time. (granted it is so slow playing back files this size, but it has use in verifying framing in crop rec mode)

In debug I noticed that two memory patches appear when in live view. From what I read on other camera model threads this appears to be normal?

For now I'll spend more time with the Mar02, will report anything I find, but it seemed pretty stable the last couple of days.

best regards
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on May 17, 2019, 12:32:51 AM
This topic went through a cleanup and some off topic items were split off so I wanted to make sure 7D users get the message:

I made a new build that supports 10bit/12bit raw video and posted it on my downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/). This works with both mlv_rec and mlv_lite and seems to be more stable than previous builds. In order to use mlv_play with reduced bit files you'll also need to load the raw_twk module.

Note that development is continuing on the 2.0.3 firmware. I also have an experimental 2.0.6 branch going on but I couldn't figure out how to get the Master processor to give up its secrets so it is only a partial firmware update. Any help with dumping the firmware for the Master would be greatly appreciated. Note that this would also help with other dual processor cameras like the 5DS, 5DS R, 5D4 and of course the 7D2.
Title: Re: dfort's experiments for 7D Mark I
Post by: skitron on May 20, 2019, 02:59:00 AM
Quote from: dfort on May 17, 2019, 12:32:51 AM
I made a new build that supports 10bit/12bit raw video and posted it on my downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/). This works with both mlv_rec and mlv_lite and seems to be more stable than previous builds. In order to use mlv_play with reduced bit files you'll also need to load the raw_twk module.

Note that development is continuing on the 2.0.3 firmware. I also have an experimental 2.0.6 branch going on but I couldn't figure out how to get the Master processor to give up its secrets so it is only a partial firmware update. Any help with dumping the firmware for the Master would be greatly appreciated. Note that this would also help with other dual processor cameras like the 5DS, 5DS R, 5D4 and of course the 7D2.

Thanks! I've loaded it and will let you know if I find issues, but initial testing was very smooth and uneventful.

best regards
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on May 20, 2019, 05:58:04 PM
I got a report that the 5x zoom mode is off center. Not surprising, it is probably affecting other branches too. There was a similar issue a while back on the 60D. Read through the Raw Video framing is off when using 60D's 'Movie Crop Mode'. Help! (https://www.magiclantern.fm/forum/index.php?topic=16912.0) topic to see how that was resolved.

There's also some reports of corrupted frames. I have not been able to reproduce the issue but I haven't tested extensively. Danne and Levas have been working on some solutions. It would be great to get the Digic 4 cameras onto their bleeding edge branches to see if these fixes are also working on these older cameras.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on May 20, 2019, 11:22:03 PM
Very first clip I used with the new build.

99% great, basically.

ONE bad frame, BARELY, out of what I did so far.


(https://i.ibb.co/f8WKpWt/image.png) (https://ibb.co/f8WKpWt)
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on May 21, 2019, 07:06:24 AM
Did the bad frame happen with mlv_rec or mlv_lite?

There are some experiments getting Digic 4 cameras working on the newer crop_rec branches but it only works with mlv_rec. Though don't expect the crop_rec module working on the 7D camera anytime soon.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on May 21, 2019, 07:35:31 PM
I don't personally care for crop_rec.

All I used was mlv_rec.

This happened with 'card backup' setting enabled.

I disabled it after.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on May 22, 2019, 07:02:53 AM
Ok--so the bad frame happened with mlv_rec. Good to know.

Have you tried mlv_lite on my experimental build? Advantage is that it should be using less processing power so possibly fewer corrupt frames. Disadvantage is that it doesn't record audio. (It is an early implementation of mlv_lite on this branch.)
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on May 22, 2019, 07:02:50 PM
Normally, no. But I don't find myself needing audio, most of the time. So I suppose I could use it for the better!
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on May 24, 2019, 06:33:04 AM
Ran MLV_Lite tonight while playing some volleyball.

Not very good results. Corrupted frames on the FIRST take upon module start / camera power on.
More corrupted frames than my first mlv_rec tests.

It barely ran for 10 seconds per series of shots, maybe a little longer on some others.
"EDMAC" error, problem producing frames, something like it.

MLV App is having a hard time reading the RAW data. Black levels are off by a bit, higher than the standard value I tend retrieve from normal clips. Value averages at 128, 129 (for optimal color balance / black level balance). Some peak at 145.

These tests were done in 10 bit, 60 fps. Varying resolutions / ratios. Mainly 2.20:1 or 2.35:1. 1600x680 resolutions or higher. I was trying to maximize quality whilst keeping optimate write speed.

The 7D can do like a 61 mb/s write speed, but even at 49/55 mb/s write speed, the camera still skips frames and stops. Until I turned on 'ALLOW frame skipping' and I never had an issue, and probably dropped very very very few frames.

I suppose the write speed jumps around a lot (because it does) to where one frame that peaks above 65 mb/s or so~ tends to cause a frame skip, therefore stopping my whole entire filming process.

As for MLV_rec... with newer tests, and FRAME SKIP to STOP recording activated, I get roughly around 1:25, TOPS in recording time. That's 10bit, 60 fps.

It varies, but that is my maximum output.
Title: Re: dfort's experiments for 7D Mark I
Post by: kitor on May 24, 2019, 09:05:29 AM
I don't see any info what card are you using, only this:
QuoteCard Warm Up: 128 MB (I have 128GB CF Card)

Have you tried with different cards?
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on May 24, 2019, 05:01:08 PM
Quote from: domo94 on May 24, 2019, 06:33:04 AM
These tests were done in 10 bit, 60 fps.

Hum--that could be the problem. All I tried was 24fps. Are you shooting mv720 or using fps override or ??

It would help to include information about the card you're using in your tests. I'm using SanDisk Extreme Pro and Komputerbay 1066x cards.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on May 24, 2019, 05:16:23 PM
Lexar 128GB 800x.
I have no other cards

I remember when I first started with the 7d back in 2016, I would be able to record max res 24fps raw 14 bit continuously without fail, almost always. It was the Sept 2016 build if I was correct, but I was silent on the forums so I never reported anything.

No FPS override.

720p Canon video mode, which allows 60fps in raw.
Title: Re: dfort's experiments for 7D Mark I
Post by: billyllid on June 03, 2019, 06:36:15 AM
hello i need help please i have 7d mark 1 with 2.0.6 firmwar and i want to instal ML can you give link to download a ML for my canon 7D 2.0.6 version ? thank you
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on June 05, 2019, 10:28:47 AM
2 Options: Downgrade to 2.0.3. Link to firmware is included in install instructions.
Or use dfort's build for 2.0.6. https://bitbucket.org/daniel_fort/magic-lantern/downloads/
Title: Re: dfort's experiments for 7D Mark I
Post by: billyllid on June 08, 2019, 05:17:06 AM
thank you !! it's safe to instal it with 2.0.6 ?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on June 08, 2019, 05:50:41 AM
Safe Magic Lantern?
Doesn't apply: https://wiki.magiclantern.fm/faq#is_it_safe
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on June 19, 2019, 11:05:25 PM
Hey dfort, can you post the LAST Mar build for 10-12 bit raw for the 7D?

The latest one is too finnicky and I can't get it to record past a minute, or 2. I got lucky and got 7 minutes yesterday but the last one seemed to be more consistent.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on June 24, 2019, 04:23:48 PM
Quote from: domo94 on June 19, 2019, 11:05:25 PM
Hey dfort, can you post the LAST Mar build for 10-12 bit raw for the 7D?

That would be a step backwards (https://bitbucket.org/daniel_fort/magic-lantern/commits/c599b6651c3c4e83c34ffcb51b0275e23f69f3d9). The one posted on my downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/) should be working much better than the previous versions. My suggestion if you want to help move this forward is to set up a development environment and build the various changesets. It isn't that difficult.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on June 27, 2019, 12:49:47 AM
I'm just the worst with code.
I have bare minimum knowledge.

I know you've posted the resources before, could you point me in the right direction on what I can to do contribute?
The basics I need to learn, any programs to download?

I'm great with numbers / patterns.
Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on June 27, 2019, 01:43:32 AM
You can download the Sourcetree app (https://www.sourcetreeapp.com/) to clone and view the history for the main ML repository or any of the developers' repositories. Mine is located here:

https://bitbucket.org/daniel_fort/magic-lantern

To set up a development system I created these tutorials:

Macintosh (https://www.magiclantern.fm/forum/index.php?topic=16012.0)
Windows/Cygwin (https://www.magiclantern.fm/forum/index.php?topic=15894.0)

Then there is this discussion for using the Linux subsystem in Windows 10 -- which can also help get you started if you are using Linux without Windows.

Compiling Magic Lantern on Windows 10 (using its Linux subsystem) (https://www.magiclantern.fm/forum/index.php?topic=20214.0)

Once you are set up it is simply matter of picking a revision to of any of the various branches and compile for your platform. When you look at the changes between some of these revisions you'll realize just how nonsensical some of these claims are about old builds being so much better than the latest versions.
Title: Re: dfort's experiments for 7D Mark I
Post by: Mr. NVT on September 14, 2019, 11:39:25 PM
Apologies for being new here, new to ML and new to the d7 - the answer might be right in front of me, but I'm apparently blind.

I'm **REALLY** struggling with getting a clean HDMI from the camera.  I'm on v 2.0.6 - and can't find anywhere in the menu that works. 

I'm streaming on YouTube with Streamlabs OBS, capturing live view through the HDMI.  The issue I keep getting is that darn white focus box, or the face detection auto focus box.  It's always one or the other. 

Our show streams for 40 min at a time, and we're trying to upgrade from crappy logitec 920 webcams to our first DSLR through the elgato camlink.   

Thanks dfort for the build!
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on September 15, 2019, 08:12:19 AM
You need to downgrade to 2.0.3 and run experimental "lua_fix" build (Top of page -> Downloads -> Downloads (drop down menu) -> Experiments.

30 minutes issue: "Prefs" screen -> Powersave in LiveView -> 30-minutes timer Disabled
This setting is not available in nightly builds.

Clean HDMI: "Display" screen -> Clear overlays HalfShutter/WhenIdle/Always/Recording
Adjust as needed.
Title: Re: dfort's experiments for 7D Mark I
Post by: jack16277 on October 02, 2019, 07:01:35 AM
Hello, I've been looking around this post to find something about the 2019Mar20 update.

I've tried it and seems stable (the screen started flickering, but it faded away and went back to normal), but the only thing that doesn't work is the autofocus in live mode (sigma lens, quite new model)

Maybe it's a known issue/bug, but couldn't find something about this specifically in the thread  ::)

Also, is it worth to update from the stable version?

Sorry, I know those are pretty stupid questions, I'm new to this and I'm experimenting, hope you understand  ;)

Title: Re: dfort's experiments for 7D Mark I
Post by: dfort on October 05, 2019, 06:45:21 AM
Quote from: jack16277 on October 02, 2019, 07:01:35 AM
Hello, I've been looking around this post to find something about the 2019Mar20 update.

What are you talking about--the magiclantern-lua_fix.2019Mar20.7D203 branch that is posted on the experiments downloads page (https://builds.magiclantern.fm/experiments.html)? If autofocus is working with your lenses using the unified branch (a.k.a. Nightly build) but doesn't work on the lua_fix branch, file a bug report. Hum--if autofocus works with Canon lenses and not Sigma lenses maybe that's a bug too? I have no idea who would be able to look into that to fix it.

Quote from: Mr. NVT on September 14, 2019, 11:39:25 PM
I'm **REALLY** struggling with getting a clean HDMI from the camera.  I'm on v 2.0.6 - and can't find anywhere in the menu that works. 

Does it work with 2.0.3? Sorry, I don't use HDMI on this camera and haven't looked into the 2.0.6 builds lately.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on October 16, 2019, 08:15:42 AM
Hello guys,

Since I'm a 7d Mk i user, I decided to post this here.

For those consistently using the 7d mk i, I have the VAF filter up for sale, it's in great condition, still clear, no scratches, still have the box, cloth, and tweezers.

My 7d is reaching it's life span and my CF card broke. I don't have anymore intention or drive to work on the 7d and I need to get a camera that's consistent and reliable. I also need one that produces an image that's acceptable in low light.
I've not found any time or need to keep using my camera so I'd rather sell the VAF filter to somebody who is still interested in this beautiful camera/ software.

I love the VAF filter a ton, but it's just time to say goodbye.

The website sells it for $295.
I can let it go for $200 as it's been used and definitely wiped down and cleaned.
Shipping included.

Email me at [email protected] about it, thanks!
Title: Re: dfort's experiments for 7D Mark I
Post by: PaulHarwood856 on October 16, 2019, 12:07:11 PM
Hello domo94,

    I just sent you an email regarding the VAF-7D.

- Paul
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on October 16, 2019, 05:15:32 PM
Would anybody be willing to buy the canon 7D as well?

It works, great for testing purposes for when you wanna be worry free on a build.

It's got a few several hot pixels that show up, especially on raw.

Still works great, it's been taken care of.

Fell once and cracked the top left of the screen, but it's unobtrusive, with a corner scuff on the body.
Can include (non branded) batteries, charger, strap, and cap. No card(it broke on me).
The core basics, not an issue.

I wish I could justify using this but I have no other work camera now, and I need something that people won't say- oh that's old.

Might switch to Fujifilm.
I need faster workflow.

Again - [email protected]

(I'll be using a t2i still for fun!)
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on January 30, 2020, 07:17:19 AM
I've been so discouraged when my 128 GB CF failed on me. My main lenses are both broken and unusable.
I was getting too many bad pixels, my low light was nothing compared to Sony. Not to mention I lost all my batteries except one.
I retired the 7D, or so I thought.

I've been letting a friend use my t2i as of recently, I've been left with no choice but to use my 7D.
I've had terrible months since September. I can't upgrade my gear. I decided to sell the anti-aliasing filter for cash flow. I sold my rig to my friend with the Sony. I sort of just dismissed all my gear.

A friend invited me to a shoot a shoot and realized I had enough to salvage parts make a mini rig. I decided to buy a couple pieces of equipment (a 1/4 screw and 64GB Sandisk CF for $40!!), and I've been using my 7D.

I told myself I wouldn't be filming in RAW anymore because it was just not producing the images the way I wanted/needed them to.
I have no clue how I was doing anything wrong when I was properly exposed on the camera, but my imagery always came out dark with the magic lantern.
The anti-aliasing filter was just too much to manage and some of my lenses didn't sit well with it.

I saw a video a while ago of super sharp t2i footage and the person basically just bumped up the bitrate recording setting. I decided to do the same with the 7D instead so I can get more space out of my card and more reliable footage.
Filming RAW was extremely experimental. I did get some good footage, but overall, RAW really did slow me down a lot- might as well just get a black magic cinema cam for that big boy stuff.

I decided to use my 7D again. See what I could rebuy for cheap/used to have resource.

I'm gonna be not using RAW anymore and just film normal with a higher CBR bitrate for filming. Hopefully 2.0 will work well.

Today I did 1.8 and it was smooth, let me tell yah.
Title: Re: dfort's experiments for 7D Mark I
Post by: SourMashII on February 12, 2020, 12:04:03 AM
Maybe this should have been its own thread? but I'm a 7d.. so here we go. I've downloaded the Nightly build, and my CF card won't recognize it in-camera. Mac shows me it's on the card, but the camera doesn't. I did end up on Canon 2.06 out of the whole thing. so not all is lost.. Help?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on February 12, 2020, 12:46:57 PM
Sorry, your description lacks detail.

If you want to use an official nightly build you have to downgrade to 2.0.3 first (as described in intallation instructions). If you want to use firmware 2.0.6 you have to use dfort's custom build https://bitbucket.org/daniel_fort/magic-lantern/downloads/
If there are errors you may have to add some details:
Full error description, maybe screenshots, name of build used... Be precise!
Title: Re: dfort's experiments for 7D Mark I
Post by: Biojack on April 20, 2020, 12:52:03 AM
Hi all, haven't been on the ML forums in years but I've got two 7D Mark I and I'm thinking Magic Latern might be useful for a new project. I've just spent many hours trying to figure out what is possible with ML these days and there is so much info on here... I'm not sure if I'm wasting my time in 2020 or of I'm on the right track with trying to give a new life to the gear we already own...


Question 1: I can't seem to find a clear answer for is what memory card we need.  We have old Sandisk extreme 32GB at 60 MB/s. I understand these would limit what we can do with magic lantern.  If I want to upgrade, what would you recommend? Do I need the 160mb/s Sandisk or would the 120 MB/s do just fine? I would rather get the cheapest cards that would work and be reliable. Also, Would you recommend 128GB over 64GB?

Question 2: I'm assuming it's somewhere in the forum.... but if someone has time and wants to help I'm wondering if ML would also allow us to use our 7D cameras and get a clean HDMI feed? And if so, what can be expected and for how long it might work? My spouse would like to setup a home studio to add a live classes to her online course business and wondering if we should get new cameras or if we can manage with what we already have. I came across a few posts that makes me believe it works but wondering if it's a good solution for our need. Anyone has experience with this? We would need to also purchase an HDMI capture card and after this post I'm going to have a look at the new Elgato products.

Question 3: What firmware version would you recommend for A) Quality recording of dialogues/interviews B) Quality recording for city/wildlife C) Webcam use with HDMI feed. I'm assuming different firmwares might be needed?

In the mean time I'll keep reading the forums and watching YouTube examples.... I was just about to see if I could sell the 7Ds but perhaps there is still life in them... haven't filmed videos since 2014 but just upgraded my PC and reactivated my Adobe Creative Cloud... so much catching up to do.

Thanks for your time
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on April 20, 2020, 11:23:28 AM
@Biojack,

My advice to you is to keep your 7D-s no matter what.  This camera produces the most beautiful and highest quality continuous 1728x972 resolution RAW video with sound that I have ever seen - deep natural colors, very high dynamic range, excellent fine detail, amazing post processing headroom with both - MLVApp and DaVinci Resolve.  Its crop mode also works fine if you are willing to accept a significant crop factor.  The card write speed is about 80 MB/s.  Please read this entire thread very carefully and make notes of posts that you find important.  Then download the latest build called:

magiclantern-raw_video_10bit_12bit_LVState.2019May15.7D203.zip

from here:

https://bitbucket.org/daniel_fort/magic-lantern/downloads/

and start experimenting.  Dfort did a great job with this build! If you don't need 4k you will very soon find out that this is your camera, especially for studio work.  I kept mine and have never regreted.

Good luck!
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on April 20, 2020, 01:02:36 PM
Our 7D fanboy answered first.

7D is hampered in several ways: Mostly because of dual processor design (and devs haven't found out how to control the second unit) there are some features not available:
No HDR and Dual-ISO in video mode, for example. See https://builds.magiclantern.fm/features.html Don't be mislead by Dual-ISO info there. It doesn't tell Dual-ISO in photo and video mode.
And because of old age (Digic 4) there are additional limitations: Considering todays standards HDMI options are outdated. You get 1080i in 8 bit YCbCr 4:2:2 But that doesn't mean you get a view in 1920x1080. Nope, on my 1920x1200 monitor I get black borders left and right because it is actually 4:3 ratio (remember old times with 1024x768 screens? Same ratio.)
My SL3/250D/Kiss X/200D Mark II does true 1080 via HDMI. And 4k. And 10 bit. Canon learned a lesson or two between Digic 4 and 8.
And no 7D experimental build with 4k options. See experimental build page.

You haven't mentioned what (in numbers) "Quality recording" means for you = Resolution, frame rate, bit depth. And if you want to use crop mode or not, if you want to invest in a VAF for uncropped mode, ...
And if you have an external mic ... I hope you do, because an internal mic is not that great and that is not because it is mono only ... and in case of HDMI usage the only way to get sound online is an external mic connected to PC because 7D will do audio via HDMI for playback only. Not during recording!
dfort's build does not include latest lua_fixes which means there is a 30 minute limit on HDMI. Don't know if Elgato's workaround works with 7D. Don't even know if 7D is in their compatibility list. Look it up first.

Cards: Sandisk 120 MByte/s has a maximum write rate of about 70 MByte/s. Source: https://www.cameramemoryspeed.com/reviews/cf-cards/sandisk-extreme-32gb-cf-card/
7D can do better. If you need it depends on your needs, obviously. See https://rawcalculator.netlify.app/calculator_desktop
Same applies to capacity. 64 GB at 7D's maximum data rate will give you around 13 minutes raw recording (give or take).

The 7D is a workhorse. Build quality (apart water resistance) is top, battery life is quite good (still recommend an external power supply for streaming) and CF-card interface is pretty fast (80+ MByte/s).

But it is not a great cam for web streaming! Not at all.
Just connect your 7D to your monitor/AVR via HDMI and test for yourself.
Title: Re: dfort's experiments for 7D Mark I
Post by: ilia3101 on April 20, 2020, 02:23:38 PM
Quote from: IDA_ML on April 20, 2020, 11:23:28 AM
Please read this entire thread very carefully and make notes of posts that you find important.

Don't read the entire thread ever! (if it's this old) You will fill your head with a lot of outdated facts.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on April 21, 2020, 11:23:56 AM
Quote from: Walter Schulz on April 20, 2020, 01:02:36 PM
64 GB at 7D's maximum data rate will give you around 13 minutes raw recording (give take).

According to the calculator that you linked, you get 15 minutes raw recording at 14 bits and 22 minutes at 10 bits on a 64 GB card.  Quality loss at 10 bits vs. 14 bits is negligible and very hard to percept even by experienced pixel peepers.
Title: Re: dfort's experiments for 7D Mark I
Post by: Biojack on April 22, 2020, 05:52:34 AM
Thanks everyone! I'm amazed by how fast those answer came considering how old the camera is.  I guess I'll need to have a better look and decide if it's worth keeping the two 7Ds. From what I get here I'm thinking a more modern tool might be better for our current project. That said I'd love to experiment a bit, though now that I think of it the price of a new memory card is perhaps more than I'm willing to spend.

I loved using the 7Ds for wildlife photography. Even though it was a heavy toy I it up many mountains and on many multi day hikes in the Pacific NW. I even took one and 5 lenses around Thailand, Laos, Cambodia, New Zealand and Australia on a 6 months trip in 2012. Bullet proof cameras! I probably should have sold at least one of them in 2012... 
Title: Re: dfort's experiments for 7D Mark I
Post by: Wlad81 on May 07, 2020, 11:04:35 PM
Is this camera is capable now of recording 1920x1080 12 bit 25fps video?

Can it write 80 MB/s or 90 MB/s on 1066x CF?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on May 07, 2020, 11:10:35 PM
https://rawcalculator.netlify.app/calculator_desktop
Title: Re: dfort's experiments for 7D Mark I
Post by: Wlad81 on May 07, 2020, 11:22:51 PM
So the data on this page is incorrect?

https://kameratrollet.se/canon-write-speeds/
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on May 07, 2020, 11:41:53 PM
My own measurements with Komputerbay 1066x 128 GB:
write 85 - 91.x MB/s (in photo mode)
In video mode around 55 MByte/s.

These are benchmark numbers and real data rates may be not well matched.
Title: Re: dfort's experiments for 7D Mark I
Post by: yourboylloyd on May 07, 2020, 11:42:31 PM
Quote from: Wlad81 on May 07, 2020, 11:22:51 PM
So the data on this page is incorrect?

https://kameratrollet.se/canon-write-speeds/

Use the calculator that Walter sent you. It will answer all of your questions
Title: Re: dfort's experiments for 7D Mark I
Post by: Wlad81 on May 07, 2020, 11:52:21 PM
So, I suppose, this camera is better than 50D?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on May 08, 2020, 12:20:17 AM
Quote from: Wlad81 on May 07, 2020, 11:52:21 PM
So, I suppose, this camera is better than 50D?

It has a higher data rate, that's all.
Title: Re: dfort's experiments for 7D Mark I
Post by: Wlad81 on May 08, 2020, 12:24:54 AM
Higher data rate = higher resolution (continuous).

Quote from: Walter Schulz on May 07, 2020, 11:41:53 PM
My own measurements with Komputerbay 1066x 128 GB:
write 85 - 91.x MB/s (in photo mode)
In video mode around 55 MByte/s.
Why so?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on May 08, 2020, 12:26:52 AM
Nothing to write home about.
10 percent higher data rate translates to next to nothing higher vertical/horizontal resolution.

To get double resolution you need quad times data rates.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on May 08, 2020, 12:38:35 AM
Quote from: Wlad81 on May 08, 2020, 12:24:54 AM
Why so?

Photo mode benchmark rates = Numbers close to theoretical date rates, overhead by computing mostly ignored
Video mode benchmarks rates = Numbers close to numbers in real life, overhead by computing taken into account.
Title: Re: dfort's experiments for 7D Mark I
Post by: Wlad81 on May 08, 2020, 12:40:22 AM
Are you talking about h.264 or raw video?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on May 08, 2020, 12:41:29 AM
Are you trying to troll?
7D's typical data rate in H.264 in 1080/IPB is about 5.5 MByte/s.
Title: Re: dfort's experiments for 7D Mark I
Post by: Wlad81 on May 08, 2020, 12:12:24 PM
OK, I see. Thank you so much.
Title: Re: dfort's experiments for 7D Mark I
Post by: Blandest on August 06, 2020, 07:34:27 AM
I set up ML on my 7D and shot some RAW video with it.  Thanks everyone for all the work, great forum posts and guides.  It was all pretty easy to do.

However, the footage that I got out of the camera was unusable to me.  It was the aliasing.  I can handle having to light my shots.  But the aliasing was intense, and even using a soft lens or diffusion filter didn't help much.  I was just testing shooting some nature (trees and leaves).  Is all the good footage I saw done with the 7D likely also with the use of an OLPF filter? Is there a cheaper alternative than obtaining that filter to get rid of the extensive aliasing?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on August 06, 2020, 09:16:17 AM
I suppose you were in un-cropped mode (1736 horizontal resolution)? Have you tried crop mode?
Title: Re: dfort's experiments for 7D Mark I
Post by: Blandest on August 06, 2020, 06:47:14 PM
Quote from: Walter Schulz on August 06, 2020, 09:16:17 AM
I suppose you were in un-cropped mode (1736 horizontal resolution)? Have you tried crop mode?

Ah, valid point, thanks.  Yes, I was using the un-cropped mode as that works better with the lens focal lengths I own.  But maybe all those videos I saw online that looked great and didn't have aliasing were using the cropped mode.  I didn't consider that. 

So basically un-cropped mode requires the OLPF filter?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on August 06, 2020, 08:16:50 PM
Only every 3rd line used and build-in AA filter designed for photo, not video: Sure, aliasing galore! VAF recommended.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on August 09, 2020, 07:40:08 AM
Last 7D 2.0.3 build with working MLV 10 bit raw, slow mo, and sound together.

https://vimeo.com/445929142

This will be the last time I'm using the 7D, it seems.

Hot pixels everywhere, MLV app kept crashing on me, overall outclassed by cameras today to do simple & effective jobs.

I used a Canon 80d here to fill the rest of the footage. About 80% of footage was 7D at 10 bit raw.

I love this beast.
Resolution was at 1440 x 448 ~ something like that. Still sharper than Canon 80d at 1080p.

Not too much noticeable aliasing. I bumped up to 508p as well on some shots.
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on August 09, 2020, 03:48:17 PM
Quote from: domo94 on August 09, 2020, 07:40:08 AM
This will be the last time I'm using the 7D, it seems.
Hot pixels everywhere, MLV app kept crashing on me, overall outclassed by cameras today to do simple & effective jobs.

Could you please give a more detailed explanation on why you are giving up on the 7D?  Are hot pixels getting more and more with time?  Didn't you try removing them with MLVApp?  And why does MLVApp crash on you?  I often leave it rendering overnight, I also run multiple copies on it simultaneously when I have a lot of footage to process and convert and never experienced any crashes.  Could it be that there is a hardware/software problem with your PC?  And finally, if you want to invest a lot of money into a new system, what is it and what real improvements do you expect from it compared to 7D RAW filming?
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on August 09, 2020, 09:06:30 PM
I just need consistency and an efficient workflow.

I find myself sitting at the PC correcting more than I am producing.

Yes more hot pixels throughout time and MLV app even though I mapped the pixels, still show up for some reason. They become very prominent the heavier I color  grade.

I left it rendering and it crashes within 1-2 clips sometimes immediately. There was one pass I did where it rendered all the videos but they were literally no where to be found.

I just upgraded to Ryzen 9 hardware and am overclocked with 3200MHz 32GB DDR4 Ram.

Honestly it's a tough choice for me, but I just realized I'm not producing enough money for crazy gear.

I can't be buying multiple memory cards for small projects.

I love the RAW workflow and it's flexibility. The sharpness, dynamic range, and fidelity is unmatched.

But I need an image I can drag and drop into the timeline.

I need solid consistent 60FPS Because sports. And I need the autofocus, which is still a give or take because I'm very good at focusing manually.

I can't use crop mode because I need a wide field of view to get into the action for my work now.
This client is very low maintenance but I also don't want to fail him so the job I have now entails me to have a solid camera.

I need better low light performance as well. 7D caps at 1600 and the 6400 ISO sets a new floor for noise and is virtually less then 1600, I can't also sit there denoising all my work. Every time.

Unless I can figure out a way to get my 7D to work well with 10bit 60fps raw and no crashing , I need to stick to a camera that provides a good image and just works.

Another huge advantage with the 80d is that I can stop and start recording instantly as well and I will always get the same image.

If I need long term recording, my 7D won't produce the same image shooting normal MOV vs RAW which will eat up all my cards.
Title: Re: dfort's experiments for 7D Mark I
Post by: domo94 on August 09, 2020, 09:09:44 PM
Though I'm still considering making my own content with a RAW workflow to match that of a RED or BlackMagic.

14 BIT raw with talking head interview type work or so.

But it's just not something I can consider for my professional workflow anymore, unfortunately.

Hard to break up with this camera and firmware hahahaha
Title: Re: dfort's experiments for 7D Mark I
Post by: IDA_ML on August 10, 2020, 02:41:23 PM
In my opinion, only 5D3 can match the colors, dynamic range and sharpness of 7D RAW footage, especially in the crop mode.  You should at least keep your 7D for ultimate RAW video quality.
Title: Re: dfort's experiments for 7D Mark I
Post by: dlprod on September 01, 2020, 01:20:40 PM
Hi Everyone

Does anyone know where Dfort's builds have gone?  Bitbucket appears to be unavailable.

https://bitbucket.org/daniel_fort/magic-lantern/downloads/
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on September 01, 2020, 01:31:44 PM
Use Web Archive Wayback Machine and enter string ...
Title: Re: dfort's experiments for 7D Mark I
Post by: dlprod on September 05, 2020, 01:18:10 PM
Nothing there, but thx.
Title: Re: dfort's experiments for 7D Mark I
Post by: srsa on September 05, 2020, 01:25:16 PM
Quote from: dlprod on September 05, 2020, 01:18:10 PM
Nothing there, but thx.
https://web.archive.org/web/20200622002229/https://bitbucket.org/daniel_fort/magic-lantern/downloads/

Tried 2 files, they are still available after a redirection.
Title: Re: dfort's experiments for 7D Mark I
Post by: dlprod on September 05, 2020, 01:44:02 PM
OMG thank you sir!! :)
Title: Re: dfort's experiments for 7D Mark I
Post by: Mautdrunn on September 17, 2020, 06:03:46 PM
Hi, it´s been quite long I using very basic hacks of ML.

I first would like to know if it´s possible to achieve higher resolutions than 1920x1080 on a 7D, if it´s possible how do I set this up on ML?
Is there a specific CF card to improve that resolution? I´m using a Lexar Professional 400x - 32Gb.

I´m doing some anamorphic movies and that would benefit me a lot when stretch out a higher resolution.

I don´t have any need to shoot for long times. I mostly filming surfing... so my clips range from 8s-25s most of the time.

What else could I do to improve files quality to post process? I tried to go through the topics by you guys are next level! Bitrate hels, RAW movie? How to set this up?


This is kind of movies I´m doing:
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on September 27, 2020, 08:46:00 AM
You can record higher resolution with raw/MLV only. Continuous recording requires a fast (and quite expensive) 1066x CF-card like Sandisk Extreme Pro, Lexar Professional, Komputerbay. No idea if slower cards will do for 25s.
Bitrate manipulation does not apply to RAW/MLV.

Caveat: There are at least two Lexar varieties. One offers premium performance ...
Benchmark your cards after purchase!
Title: Re: dfort's experiments for 7D Mark I
Post by: ProFoto on September 30, 2020, 03:17:51 PM
Quote from: dfort on December 09, 2017, 12:32:55 AM
Sort of a low priority project and highly experimental but I do have a build for 2.0.6 on my downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/). Just updated the pull request (https://bitbucket.org/hudson/magic-lantern/pull-requests/878/update-to-7d206/diff) for this so there aren't any conflicts if anyone wants to play around with it.

I am not able to access the file to download ML...Since I am using 2.0.6 and I cant find a retro to 2.0.3 does that mean I cant use ML??  I cant get the page for this build.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on October 02, 2020, 01:23:55 PM
See reply 1637. There you will find firmware files for 2.0.3 and a custom build for 2.0.6.
Title: Re: dfort's experiments for 7D Mark I
Post by: Lensmeister on October 03, 2020, 12:40:13 PM
Morning,

Many moons ago I d/l ML version magiclantern.7D.203.Alpha2 .

I have a 7D with Canon firmware version 2.0.3 on it.

I was watching this video
https://www.youtube.com/watch?v=vpSdABzwQII
on How to record 2k RAW with the Canon 7d and Magic Lantern but the screens don't look the same and I cannot find out how to set the 7D up to record 2K Raw. When I do try to film with ML running it's a bog standard .mov file that comes out.

Please can you advise?

Thanks in advance.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on October 03, 2020, 12:45:22 PM
Alpha completely outdated. Obsolete. Junk.
Top of page -> Downloads -> Download nightly builds -> 7D
Title: Re: dfort's experiments for 7D Mark I
Post by: Lensmeister on October 03, 2020, 12:54:56 PM
Quote from: Walter Schulz on October 03, 2020, 12:45:22 PM
Alpha completely outdated. Obsolete. Junk.
Top of page -> Downloads -> Download nightly builds -> 7D

Thanks
Title: Re: dfort's experiments for 7D Mark I
Post by: smorgan3 on October 19, 2020, 09:53:13 PM
Can anyone point me to any recent guides or how-to videos on how to record raw? Everything I've found so far is at least a few years old. I've installed the latest (2018) nightly build onto my 7D and played around with the settings. Tried recording raw but it crashed, showed error 70, and I thought I bricked it until I switched the battery and it came back. Looking for more recent info so I know what I'm doing. Thanks!
Title: Re: dfort's experiments for 7D Mark I
Post by: Brawl on October 22, 2020, 11:07:35 AM
Hi, talking about raw video.... any hope for a new update with continuous shooting on crop mode?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on October 22, 2020, 11:10:11 AM
I'm able to record in crop mode for unlimited time (apart battery & card capacity) continuously. Please specify your request.
Title: Re: dfort's experiments for 7D Mark I
Post by: Brawl on October 24, 2020, 03:27:14 AM
Quote from: Walter Schulz on October 22, 2020, 11:10:11 AM
I'm able to record in crop mode for unlimited time (apart battery & card capacity) continuously. Please specify your request.
hi Walter I'm sorry I miss the forum since a bit, as well I miss playing with my camera (I'm more focused on audio todays as sound designer for films). I remember I couldn't record for long takes in crop mode there were stops after a while. But I had no problems in non-crop, continuous shooting was working great (I have a fast CF from sandisk), the only downside was moire/aliasing.
Which resolution can you record in crop mode continuous shooting? Which settings/raw module do you use? Thx
Title: Re: dfort's experiments for 7D Mark I
Post by: SavedGround on November 06, 2020, 02:56:10 AM
Hello I can't find where to enable a better resolution out of the HDMI. Can someone please point me in the right direction?

Any help would be appreciated.

Thank-you,
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 06, 2020, 07:08:45 AM
Well, there is no such option within ML.
Title: Re: dfort's experiments for 7D Mark I
Post by: SavedGround on November 13, 2020, 05:43:01 AM
Thank-you for the quick reply. I'm only getting 720x486 and thought I've read others getting 1080i or using an Atomos Ninja
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 13, 2020, 07:58:25 PM
Check Display tab -> Advanced settings -> Force HDMI-VGA
Should be OFF.
Title: Re: dfort's experiments for 7D Mark I
Post by: SavedGround on November 14, 2020, 04:46:58 AM
Unfortunately OFF is selected.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 14, 2020, 06:18:00 AM
Let me ask about your HDMI device. I suppose you are using an HDMI capture device. Have you tried to connect directly to a monitor/TV instead?
Title: Re: dfort's experiments for 7D Mark I
Post by: SavedGround on November 14, 2020, 02:33:04 PM
I am only using a monitor. I tried a professional video monitor as well as a consumer television.
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on November 14, 2020, 09:00:07 PM
Sorry, unable to do proper testing here. My monitor (my only HDMI input device) will just fetch anything and stretch it to Full HD without telling source resolution.
Title: Re: dfort's experiments for 7D Mark I
Post by: SavedGround on November 15, 2020, 02:26:14 AM
That's okay. Appreciate your help.
Title: Re: dfort's experiments for 7D Mark I
Post by: Peter Gaal on December 17, 2020, 01:09:28 PM
Hi! Currently availableIs is 10-12bit raw video for 7d? I found this link:
https://bitbucket.org/daniel_fort/magic-lantern/downloads/
But this page currently unavailable.
Many thanks!
Peter
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on December 18, 2020, 07:53:34 PM
https://web.archive.org/web/20200622002229/https://bitbucket.org/daniel_fort/magic-lantern/downloads/magiclantern-raw_video_10bit_12bit_LVState.2019May15.7D203.zip
Title: Re: dfort's experiments for 7D Mark I
Post by: Peter Gaal on December 20, 2020, 09:33:03 AM
Thank you!!
Title: Re: dfort's experiments for 7D Mark I
Post by: Elvis_Cheung_0215 on December 23, 2020, 12:09:08 PM
Hello everyone I got a problem with my 7d is that the raw zebra doesn't show when I am using video mode in my 7d set below 200 iso or over 1600 iso, however, if I use photo mode then it will be completely fine. hope someone knows how to fix it :)
P.S. I am an ML noob don't laugh at me  :'(
Title: Re: dfort's experiments for 7D Mark I
Post by: GODenish on January 12, 2021, 06:51:46 PM
hello. I have a Canon 7d. firmware 2.06. What ML build can I install on it?
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on January 13, 2021, 07:35:52 PM
With 2.0.6:
Custom nightly build by dfort https://web.archive.org/web/20200622002229/https://bitbucket.org/daniel_fort/magic-lantern/downloads/magiclantern-Nightly.2018Jul04.7D206.zip

After downgrade to 2.0.3
Nightly build from -> Top of page -> Downloads -> Download nightly builds -> 7D
or
lua_fix build -> Top of page -> Downloads -> Downloads (drop down menu) -> Experiments -> Latest Lua updates + fixes
Title: Re: dfort's experiments for 7D Mark I
Post by: Skylar on February 06, 2021, 04:42:42 AM
So, I'm curious as to if there is any recent builds for the 7D for either 2.0.3 or preferably 2.0.6.  If so how stable are they and could they be reliable enough for using everyday?
Title: Re: dfort's experiments for 7D Mark I
Post by: Mathieuremy on February 17, 2021, 04:17:21 PM
Quote from: dfort on December 09, 2017, 12:32:55 AM
Sort of a low priority project and highly experimental but I do have a build for 2.0.6 on my downloads page (https://bitbucket.org/daniel_fort/magic-lantern/downloads/). Just updated the pull request (https://bitbucket.org/hudson/magic-lantern/pull-requests/878/update-to-7d206/diff) for this so there aren't any conflicts if anyone wants to play around with it.


The links does not work
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on February 19, 2021, 05:54:30 PM
https://web.archive.org/web/20200622002229/https://bitbucket.org/daniel_fort/magic-lantern/downloads/magiclantern-Nightly.2018Jul04.7D206.zip
Title: Re: dfort's experiments for 7D Mark I
Post by: mblackshore on April 28, 2021, 04:11:16 AM
Hello is this tread dead?
Which is the latest stable working version for Canon 7D?
Which version of canon firmware should I use with that?

Latest Build (2018-07-02 22:37) accourding to this: https://builds.magiclantern.fm/7D-203.html
Is it stable?
Title: Re: dfort's experiments for 7D Mark I
Post by: domasa on May 17, 2021, 11:10:09 PM
Thx for mlv-lite. Please can you add lossless compression support?
Title: Re: dfort's experiments for 7D Mark I
Post by: someonedifferent on June 28, 2021, 11:42:25 PM
Dear amazing people of ML,

I am about to install the Nightly Build for my 7D on 2.0.6. It is my first try for ML, so I hope everything works and my camera won't go up in flames (lol).

As an enthusiast techie and photographer, are there any tests I can do to help you? I am a physicist, so I'm capable and very happy to provide thorough research and experiment documentation (since that is so important). I would be delighted to help with anything regarding this build.

Also, what is the criterion for calling a build stable/a main build? What needs to happen to make this a stable build? (Please kindly tell me if that is an obnoxious question, I am a newbie when it comes to ML.)

All the best!
Title: Re: dfort's experiments for 7D Mark I
Post by: Walter Schulz on June 29, 2021, 10:47:49 AM
Quote from: someonedifferent on June 28, 2021, 11:42:25 PM
What needs to happen to make this a stable build?

ATM nobody knows. Quite some time ago ML project went into hibernation and a full relaunch didn't take place yet.
Title: Re: dfort's experiments for 7D Mark I
Post by: Spups on December 20, 2021, 01:48:31 PM

Using Dfort's Last 7D build.
10 bit MLV to ARRI LOG C prores LT in MLV APP
From there I edit in premiere, grade with Filmconvert Nitrate (FUJI 8563 stock) and export as 4k to spoof youtube into adding decent bitrate
Title: Re: dfort's experiments for 7D Mark I
Post by: tek03 on December 24, 2021, 01:51:12 PM
Quote from: Spups on December 20, 2021, 01:48:31 PM

Using Dfort's Last 7D build.
10 bit MLV to ARRI LOG C prores LT in MLV APP
From there I edit in premiere, grade with Filmconvert Nitrate (FUJI 8563 stock) and export as 4k to spoof youtube into adding decent bitrate

I downloaded dfort's build from the links above, but I still can't manage to make 10 bit mlv work. I can't find the setting, I'm going nuts! I went trough all the settings


EDIT- I DOWNLOADED THE WRONG BUILD *FACEPALM*, NOW 10BIT WORKS, IT IS: "magiclantern-raw_video_10bit_12bit_LVState.2019May15.7D203"
Title: Re: dfort's experiments for 7D Mark I
Post by: Spups on January 20, 2022, 07:10:07 PM
LOW LIGHT examples from 10 bit mode. All footage ISO 800+. some shots lack contrast as lens was wide open
canon 7d magiclantern-raw_video_10bit_12bit_LVState.2019May15.7D203
Inside scenes were all ISO 1600. lighting was poor, fluorescent tubes with LED's overhead, I thought the footage held up and still looks pleasant depite the different sources.