Canon 650D / T4i

[Tongotongo]

Again about the sound settings (the ML one) is there any development in that subject on the 650D?
I really would like some VU meters on the live view or at least audio out monitoring through the usb interface of the camera.
Is this possible for the 650D?

Hola , creo que lo mejor es grabar con celular u otro dispositivo separado suena mucho mejor y le da un calidad al video muy aceptable, es mi opinión.

Hello, I think it's best record with phone or other separate device sounds much better and gives a very acceptable video quality at, is my opinion.

[dngrhm]

Again about the sound settings (the ML one) is there any development in that subject on the 650D?
I really would like some VU meters on the live view or at least audio out monitoring through the usb interface of the camera.
Is this possible for the 650D?

• 1% mentioned a long time ago in a thread far far away some success on the EOS-M (the 650D's evil twin) with audio.  There is some hope and it is somewhere on the list of stuff to do.
• As with all development on ML there is no estimate of when it will be available or worked on.
• It's a very useful feature, it is not the ONLY useful feature
• The developers are aware of a desire for this feature
• There is a so-so work around to use the canon menu to set the manual gain and check the level
• You are not alone in requesting this feature in this thread or the one before it.

[killr0y]

It's been a while since I've gotten onto the ML forums, and I've waded through a ton of pages and haven't seen the answer to this so please don't bite my head off if someone has already asked this question:

I know our shitty 650D's don't have the bandwidth for 1080p raw video, but what about the original raw mode where you could get 2 seconds of 1080p raw in the buffer and let it do the write-behind to the SD card?  I think that would be a feature worth having.  Of course, next I'm going to hear that the buffer is so tiny that would only work with about 1 second @ 24fps

[Alquimia16]

Hey man, continous 720p RAW is enough for the 650D to not be " shitty" at all!

What possible serious use could 1-2sec of 1080p raw have? Maybe im lost but  720p is WAY more than expected for a entry lvl dsrl ;=)

[davidtlong]

Hi all,

two questions

1.  It seems that to manually control ISO you need to be on "M" as any other setting is auto ISO and I can not seem to turn that off.  Is this correct?
2.  Has anyone tried integrating audio into footage?  I assume that one would have to record audio on an independent device and then integrate in software.  Is this correct?  Also if it is it would be interesting if that along with the raw recording, a separate audio recording could be made.

Thanks in advance,

dave

[brewerrj]

• 1% mentioned a long time ago in a thread far far away some success on the EOS-M (the 650D's evil twin) with audio.  There is some hope and it is somewhere on the list of stuff to do.
• As with all development on ML there is no estimate of when it will be available or worked on.
• It's a very useful feature, it is not the ONLY useful feature
• The developers are aware of a desire for this feature
• There is a so-so work around to use the canon menu to set the manual gain and check the level
• You are not alone in requesting this feature in this thread or the one before it.

Just looking for an update from the developers if there is any progress on audio.

[mdfaisal]

Sure, and you're lucky : I've timed most of the process !

Precisely, I've shot 11min12sec of RAW at 1476x626 (16 251 frames@24fps), for about 25GB on the memory card.

- The raw2cinema DNG (using RAW Magic on osx) took me 12min21sec (I ran the process directly from the rushes on the memory card and sets the output on the hardrive). Output 28GB of DNG (not 25 due to the 14bits > 16bits conforming I guess).
- PinkDotRemover tooks 34min44 (And gives me another 28GB of dots-free DNG). I used search expression (which I don't remember sorry) to selects all DNG without initial underscore (_) and delete them in order to keep only the dots free DNG.
- Finally, After Effects exports all rushes in uncompressed 8bit color-graded videos within 2h10 (Output : 39.7GB, always in 1476x626).

So globally, processing a RAW file from .RAW to 8bit video takes 15min for 1 minutes of shooting.

I haven't timed the final editing process (done with Premiere pro), but it was fast as all rushes were already processed in uncompressed 8 bits videos.

Same here... but mine 1 min video raw took about +/- 20 minute. It makes me think twice to record along project with raw.

Pink dot remover takes aloooot of time...
I hope someday developer team make a new version that we dont have to use pink dot remover to clean the dead pixels.  Is it possible 650d to have a clean raw video ?

[Tongotongo]

Same here... but mine 1 min video raw took about +/- 20 minute. It makes me think twice to record along project with raw.

Pink dot remover takes aloooot of time...
I hope someday developer team make a new version that we dont have to use pink dot remover to clean the dead pixels.  Is it possible 650d to have a clean raw video ?

Yes ,https://www.dropbox.com/s/tu8tgpbady8oyqp/ML.rar , no hay que usar pinkdotremove.

Yes, https://www.dropbox.com/s/tu8tgpbady8oyqp/ML.rar, do not use pinkdotremove

[Rewind]

Yes ,https://www.dropbox.com/s/tu8tgpbady8oyqp/ML.rar , no hay que usar pinkdotremove.

Yes, https://www.dropbox.com/s/tu8tgpbady8oyqp/ML.rar, do not use pinkdotremove

This method uses RAW_TYPE 78, which marks dead pixels, but produces a full frame grid of green dots, is absolutely not what you want, if you concerned in quality. Later your raw processor interpolates them, but the artifacts are terrible in comparsion with the Pink Dot Remover tool.
http://www.magiclantern.fm/forum/index.php?topic=6658.msg69459#msg69459

Just forget it. Use the latest PDR for now.

Pink dot remover takes aloooot of time...

Latest PDR works directly with RAW files, this is way faster then process every dng individually, also it includes drag&drop support now, so removing pink dots is just a matter of couple of clicks and a few seconds of time.

[Tongotongo]

This method uses RAW_TYPE 78, which marks dead pixels, but produces a full frame grid of green dots, is absolutely not what you want, if you concerned in quality. Later your raw processor interpolates them, but the artifacts are terrible in comparsion with the Pink Dot Remover tool.
http://www.magiclantern.fm/forum/index.php?topic=6658.msg69459#msg69459

Just forget it. Use the latest PDR for now.
Latest PDR works directly with RAW files, this is way faster then process every dng individually, also it includes drag&drop support now, so removing pink dots is just a matter of couple of clicks and a few seconds of time.

Disculpa , pero la calidad que logro con este ML + solo Rawanizer(todos los demás andan diferente)+sin puntos+sin regilla , no la logro con ningún otro . Con pinkdotremove funciona también bien pero no en todas las resoluciones y cuando son muchas las conversiones es un tiempo terrible.


Sorry, but the quality that achievement with this ML + Rawanizer only (all other walk different) + sin + sin regilla points, not with any other achievement. With pinkdotremove works too well but not at all resolutions as are many conversions is a terrible time.

[davidtlong]

Hi all,

two questions

1.  It seems that to manually control ISO you need to be on "M" as any other setting is auto ISO and I can not seem to turn that off.  Is this correct?
2.  Has anyone tried integrating audio into footage?  I assume that one would have to record audio on an independent device and then integrate in software.  Is this correct?  Also if it is it would be interesting if that along with the raw recording, a separate audio recording could be made.

Thanks in advance,

dave

Hi all,

I assume the answers to my questions are yes to both.  I can not change ISO in any other  settings for video raw other than "M".  However when put the camera into "M" the image cycles from light to dark.  Do you know if this is normal?

dave

[davidtlong]

all,

In addition to the iso comments (can I control) and problems with manual (may have solved) I just noticed that in TV mode when shooting the TV does not stay at 60, which we need.  So is my camera working right or do we need to to M mode at all times?

dave 

[davidtlong]

all,

In addition to the iso comments (can I control) and problems with manual (may have solved) I just noticed that in TV mode when shooting the TV does not stay at 60, which we need.  So is my camera working right or do we need to to M mode at all times?

dave

I think I answered my question.  M at all times I guess.

dave

[newsense]

I think I answered my question.  M at all times I guess.

dave

ML doesn't disable or change the functionality of the canon settings, it simply adds extra features. Your issues seem to be you don't know the camera well enough, i'd suggest reading the manual. I use M all the time but I know three are modes that let you control just one setting like Tv mode I believe is shutter priority, you set shutter speed camera sets iso and aperture. I really suggest looking in the manual a this really isn't ML related.

[midnite]

i shot another raw video. dealing with raw files is a pain i'm sayin you. this 73second non-compressed avi file's size was 13.7gb. uploading this to youtube took more than 24hours (shitty 1mb uploadspeed)

i still use ML build of 20.07.13 and haven't upgraded the camera's firmware. and probably i won't bother to do it unless some big change come up.

what i did:

1) i used raw2dng first to get dng files.
2) used pink remover tool to clear em.
3) opened dng's with PS6 and did some adjustment
4) extracted jpg files (which is downed to 8bit)
5) renamed the jpgs' with "bulk rename utility" (google it)
6) files are joined and mp3 added to auido with virtualdub. now i got a 1600x600 non-comp avi. sized almost 5gb.
7) avi file impoted  into after effects.
file interpolated to 1920x1080 (1920x720 actually)
9) extracted again non-comp avi. 13.7gb - ~195mb/s bitrate

even this 70second small footage took enormous time and computing power. so i think most of the time i ll stick with 1080p h264.

this is almost the same scene with h264 canon mode:

footage itself is a bit shaky thx to 50mm non-is lens. and youtube downgraded the quality of 15gb video to a 46mb one. (that's right)  enjoi.

[30dbs]

what i did:
1) i used raw2dng first to get dng files.
2) used pink remover tool to clear em.
3) opened dng's with PS6 and did some adjustment
4) extracted jpg files (which is downed to 8bit)
5) renamed the jpgs' with "bulk rename utility" (google it)
6) files are joined and mp3 added to auido with virtualdub. now i got a 1600x600 non-comp avi. sized almost 5gb.
7) avi file impoted  into after effects.
file interpolated to 1920x1080 (1920x720 actually)
9) extracted again non-comp avi. 13.7gb - ~195mb/s bitrate

That seems a lot of work, here's what I do.

1) PDR raw file
2) raw2dng
3) import dng into after effects (making colour adjustments here)
4) render out uncompressed into clips.
5) join clips in video editor and render out in h264.

You should never at any point in the process save to lossy compression (jpg) and I am not sure of the benefits of interpolating to 1080p, most players upscale nowadays anyway.

I'm also not sure of the advantages of uploading raw video to youtube, I'm pretty sure that if you (final)render in h264 , in one of the correct sizes/bitrates for youtube it doesn't re-encode anyway.

But you are right its much more work that h264, and unless you have a specific reason to shoot in raw, do not bother. if you really really need the extra dynamic range etc then its worth the extra workflow. otherwise h264 is 100% fine.

here's an example of my raw processed that way.

[spider]

You should never at any point in the process save to lossy compression (jpg)

But there is no visible quality loss.
Also I would never upload uncompressed Video too.

[midnite]

But there is no visible quality loss.
Also I would never upload uncompressed Video too.

well, i thought if  i upload the bigger and the better video, youtube would give me the better streaming video. that's why i upped non-compressed. i guess it is wrong.

[Fuma]

The more time needed for optimising the RAW files is definitely worth in my opinion.
But as long you are satisfied with Canon h264 everything is ok.

Still hoping the turnaround with the pdr will be obsolete in a near future.

[30dbs]

But there is no visible quality loss.

maybe not visible this compression round, but next compression, and subsequent compression it will be more and more visible. lossy compression should only be applied at final rendering stage.

[spider]

maybe not visible this compression round, but next compression, and subsequent compression it will be more and more visible. lossy compression should only be applied at final rendering stage.

It is still not visible. You have to do it 50 times and more to see a difference.

[spider]

I think RAW is very useful and easy to use.
I worked about 45 minutes on it except processing time.

[30dbs]

It is still not visible. You have to do it 50 times and more to see a difference.

Visibility it would show up waaay before recompressing 50 times. plus If you are not worried about the compression, then why on earth would you bother shooting in raw in the first place? a big benefit of raw is that there are no compression artifacts. (especially for me as i mainly use raw for greenscreen) let me make it clear that I understand that lack of compression is not the only advantage to raw but my original point is why waste such an advantage by using jpg as a middle man?

If you are shooting in raw it is because you want the footage in the very best condition possible. jpg WILL deteriorate the quality.

[davidtlong]

ML doesn't disable or change the functionality of the canon settings, it simply adds extra features. Your issues seem to be you don't know the camera well enough, i'd suggest reading the manual. I use M all the time but I know three are modes that let you control just one setting like Tv mode I believe is shutter priority, you set shutter speed camera sets iso and aperture. I really suggest looking in the manual a this really isn't ML related.

My canon

I will be polite    in my answer.  My canon cameras go back to the 20D, in terms of digital.  I know my camera, including being an early beta tester for Camera Raw. (added the camera raw part as a mod) 

My point is that in TV mode the speed should be fixed as it is with the straight canon software.  When shooting ML raw, the TV values on my camera change along with the iso and will not remain fixed.  Only in the M settings does TV values stay fixed.  So I was wondering if this is what I should be experiencing when shooting raw.

dave

[a1ex]

Ever tried to use your camera without ML?