Features on the TODO list - things that ARE possible

Started by a1ex, June 07, 2012, 12:51:45 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

a1ex

* Lua scripting ( http://groups.google.com/group/ml-devel/browse_thread/thread/6f708e124d700712 ).
* PicoC scripting ( http://www.magiclantern.fm/forum/index.php?topic=3769.0 )
* Android USB controller ( http://groups.google.com/group/ml-devel/browse_thread/thread/4b0d01d64a459485 )
* Follow focus / remote control with a simple TV remote ( https://groups.google.com/group/ml-devel/browse_thread/thread/6ec6723d7da4119e )
* Custom file prefix (for example, 5D2_1234.CR2, 02JA1234.CR2 etc). Maybe also absolute file numbering from shutter count.  Done: https://bitbucket.org/hudson/magic-lantern/commits/49682329d18213d88f74a9fe516c202091f14d22
* Exposure simulation for extremely dark scenes [DONE]
* Templates for HDR and focus stacking scripts
* Very short uncompressed 422 video clips (1-2 seconds, maybe more on 5D3) Done: http://www.magiclantern.fm/forum/index.php?board=49.0

You can find some more features still to be implemented in this google doc.

moosen


2xM

LUA scripting with an nice API would be awesome!

screamer

the absolute number based on shutter count will be a simple but useful new feature, i like all this new things :))
always trying to use the 100% of magic lantern..
Gear:
Canon 60D, all the samyang lenses, Canon 50 mm 1.4, Canon 60mm macro, Canon 70-300 usm, Sigma 4.5mm fisheye, Sigma 17-70 2.8, Canon 40mm f2.8 pancake, all the Lensbabies and a lot of other pieces, Flash metz 58 af2

Matthew Kieren


(Sorry a1ex! Question moved to separate thread.)
Canon 60D, 18-55mm IS, 55-250mm IS, Tascam DR-40
SMC Pentax-M 50mm f/1.4, Yongnuo YN560-II

a1ex


screative

By 422 you mean Apple Pro Rez?...please make it longer than 2 seconds recording time on this cause thats pretty useless if so....unless maybe for time lapse

1%


xcreativ

* Very short uncompressed 422 video clips YUV 4:2:2 recorded 1-2 seconds due to the very high bit rate, ie limited by the speed cards? Ie If you can achieve compression clips will be longer? Or is it due to the computational speed of the processor, memory, and so on? I understand that the function itself is only "in the future", but it's such great news!
Cameras:Canon 6D, Canon EOS 550D.
Lens: Sigma AF 70-200mm f/2.8 EX DG OS HSM, Sigma AF 17-50mm f/2.8 EX DC OS HSM, CANON EF 50 f/1.8 II, CANON EF-S 55-250 mm f/4-5.6 IS.
Sound: ZOOM H1, Prof Shotgun Uni-Directional Condenser Mic EM-320E.

a1ex

Well, the 5D3 might handle this quite well with fast cards, I have high hopes for it. Older cameras are probably too slow though.

hjfilmspeed

uncompressed 422 clips on the 5d3! 30 seconds of that would be awsome never mind 30mins. even if the 1080 quality was as sharp and easy to grade as its s2 jpeg that would be a hudge improvement! Im so pumped for the 5d3 ml! this is great! waaahhhhh hoooooo!

RenatoPhoto

Yes!! 30 seconds of 422 would be absolutely fantastic = more donations for sure!
http://www.pululahuahostal.com  |  EF 300 f/4, EF 100-400 L, EF 180 L, EF-S 10-22, Samyang 14mm, Sigma 28mm EX DG, Sigma 8mm 1:3.5 EX DG, EF 50mm 1:1.8 II, EF 1.4X II, Kenko C-AF 2X

Sergey Pylypenko

Quote from: hjfilmspeed on October 01, 2012, 09:17:12 PM
uncompressed 422 clips on the 5d3! 30 seconds of that would be awsome never mind 30mins. even if the 1080 quality was as sharp and easy to grade as its s2 jpeg that would be a hudge improvement! Im so pumped for the 5d3 ml! this is great! waaahhhhh hoooooo!
Hey guys, I was just today thinking about going on ML and ask someone if 422 is possible, and here we are.  More then I could expect...
.. it is soooo great!!!

Lcrusher

I have a question regarding the bit depth of video.
As I understand canon sensors have 14-bit processing, but the video is 8-bit. Is the bit depth reduced before or after the image is scaled to  1080? Will the uncompressed files be 8-bit or higher?

a1ex

1) From ARM you can only see it as 8 bit, with all curves applied. The pipeline seems like this:

- 14bit data
- linear transformation (digital ISO gain, black level)
- HTP (nonlinear) - if enabled
- 14 to 12 bit reduction (suggested by 14_12 strings and by lack of shadow detail in aggressive Flaat styles, compared with digital ISO gain)
- nonlinear curves (like picture style)
- result: 8-bit 422 (codenamed Craw or HD buffer, almost 1080p, see Debug menu for exact size).
- Craw is downsampled for LV display
- Craw is fed directly to H.264 encoder, where it becomes 420 and slightly upsampled to 1080p.

2) 8-bit, of course.

Guilherme

I don't know if this is possible or not, so here it goes: lots of us use external mics (not connected to the camera) so lots of times the  camera automatically stops recording and you guys have make it way more easy for us now. But the thing is: the sync is all messed up when this happens, do you have any ideas to automatically start recording but also be able to have that perfect sync in post?

Greetings from Portugal! :)

jplxpto

Quote from: Guilherme on November 08, 2012, 08:09:52 PM
I don't know if this is possible or not, so here it goes: lots of us use external mics (not connected to the camera) so lots of times the  camera automatically stops recording and you guys have make it way more easy for us now. But the thing is: the sync is all messed up when this happens, do you have any ideas to automatically start recording but also be able to have that perfect sync in post?

Greetings from Portugal! :)

http://www.magiclantern.fm/forum/index.php?topic=3475.msg19880#new

Seja bem vindo ao nosso forum. :)

Cumprimentos de Portugal.

Malcolm Debono

Quote from: Guilherme on November 08, 2012, 08:09:52 PM
do you have any ideas to automatically start recording but also be able to have that perfect sync in post?

Yep: Pluraleyes  :D
Wedding & event cinematographer
C100 & 6D shooter
New here?  Check out the FAQs here!

deleted.account.01

Quote from: Malcolm Debono on December 11, 2012, 11:15:19 PM
Yep: Pluraleyes  :D


;) i love to see how software make synhro from 5  audio sources recorded lets say on musci festival where all 5 recording devices are in diferente places :DD ..  software never replace human ear.

nbenadom

I know the 5D Mark III has it standard, but can a ML script be created to do the same for the T3i so that the images can be combined as a composite in-camera?
Regards,

Noel

Francis

Quote from: nbenadom on January 14, 2013, 05:16:29 PM
I know the 5D Mark III has it standard, but can a ML script be created to do the same for the T3i so that the images can be combined as a composite in-camera?

ML is not intended to replace common post-processing functions. You can use the extended HDR bracketing option to create a script which can be run on your computer to combine the stack with enfuse. If you are talking about the multi-exposure function, then just take multiple pictures and stack them in layers in whatever image editing program you prefer.

1%

That function sucks, it only uses Jpegs. The alignment leaves much to be desired too. Rather get the in camera working with raws on 6D/5d3 first. ML's HDR and multi expo is much better ATM.

groundlessfears

Using android for Magiclantern is the best option, as Android is one of the best and the most popular open source platform,. This is the future.

Great
Canon EOS 550D 1.0.9
Canon EOS 70D.111A
Canon EOS 5D Mark III 1.2.3

pixelgarden

Hi ML guys. I don't know if it's too much to ask for, But i think you guys can do a lot better in elevating the precieved latitude we are getting with 2.3 build.

Isn't it possible that we can pull down exposure in bright areas (highlights) while leaving the mids and shadows untouched? Some thing like a negative Digital ML iso, but only for sections of the frame that shows RGB unit value in neighbourhood of 255 (say 200 to 255). Also coz ML can pinpoint every single pixel with a equivalent RGB value, I think all that is needed is an algorithm to pull those higher values in lower range or to simulate a negative ML digital iso in bright sections.. Please pardon me if i am talking too ameturish, but it felt that if canon mewP can sense & achieve better shadows with ALO, why can't that algorithm be rev-engg. and those highlights be preserved in a much visible range. For me HTP is pretty useless as i don't see much benefit.

Huge regard for the entire ML team.

cthornhill

A1ex, Given what you have accomplished in raw, the prospect of 4:2:2 Craw is very interesting...not as much DR as RAW, but it would be really cool to get that data stream direct to a card on older cameras like the 60D...it might actually fit at full HD or close to it on the slow interface to SD. Not sure if that could be done, but I for one would be SUPER happy with 24fps 4:2:2 Craw compared to H264...now that we have a good model of post processing with 14bit raw (raw2dng, etc.) if we could just do that same thing with 8bit Craw at least we would have super nice HD video (way more filmic than most lower cost broadcast options I bet since we have a super35 sensor). Using a cool LUT like VisionColor or others at this deep color depth (and with an anti moire filter) this would make the 60D a very interesting production tool (and the 7D too if it can be done there on the funky chip set)...very interesting potential...if it could work. Might be much more practical than  my other question about All-I on the 60D...It might be a nice compromise between full raw and high bit rate H264...and would cost less due to the lower bit rate...at least a bit less anyway. - CT

PS - Did a little math in my head while driving, an it made me want to bolster the note...No matter what happens on the 60D, the prospects for the 600D, 650D and (I hope) 70D might be more exciting...sure 14bit raw is super nice, but if 8bit 4:2:2 Craw is actually about 40% (a bit more) smaller,  not only would faster SD interfaces on those cameras allow for higher resolutions/frame rates (yeah I know you may not be able to down sample resolution on Craw given how you get it from the pipeline), but the cost could be radically lower if it really works at below 45MB/s (lower cost cards)... in any event, the potential of having a sub $1K US camera or about $1K US camera (70D) that has an APC-C sensor, and super high quality 4:2:2 video with Canon color science is very exciting...in theory it might look as good or better than the C300...not 14bit, but still a darn nice image.

Not sure if this could work, but it could be almost as radical as raw considering what people have to pay to get that sort of image quality right now from other vendors (or even Canon). I mean the image internal H264 record quality of the C100 is way better than the regular video of even the 5D3 (much sharper esp.) - and the HDMI out from the C100 is 8 bit 4:2:2 and on a scope/screen identical to that from the C300...hence people jumping to it (C100) in droves for events and documentary work, and adding a recorder for higher end broadcast use. If you could get similar results from the 650/700/70D that would be about as massive as raw for many production tasks.

pps - Wow..I am slow...consider the implications of Craw on the 5D3 too...at the speeds it now has for 14bit raw, I bet it could do Craw and a audio with no sweat...making it at least as interesting as the C300 in image quality at 4:2:2 (not to mention 14bit raw)...this implies that a $3200US DSLR could surely record the same image quality as a $16000US camera (and we know that actually 14 bit raw is already better...but Craw could be a lower cost option if it takes less storage (or just allow longer takes with easier access to onboard audio...again adding to the disruptive and beautiful options ML is already providing.