Levas' crop_rec_4k experiments for 6D

[Levas]

Corrupted frames is not good   
Made a new crop_rec module where the max vertical resolution is lowered by 16 pixels. (So 16 lines less to read out by the sensor)

5496/3x1944 @ 24 fps
5496/3x1856 @ 25 fps

I expect this should prevent corrupted frames. Could be that it's still a little too much.
If you still experience corrupted frames, let me know, then I make a crop_rec build with the vertical resolution slightly more lower till no more frames are corrupted.

The crop_rec.mo file can be downloaded here from my google drive. you can replace this crop_rec.mo file with the crop_rec.mo file on you sd card (location on SD card:  ML/MODULES/ )
https://drive.google.com/file/d/14Rff7xmEOOVF9nxSFRYYVyKutGqhYS-D/view?usp=sharing

[zezic]

Corrupted frames is not good   
Made a new crop_rec module where the max vertical resolution is lowered by 16 pixels. (So 16 lines less to read out by the sensor)

5496/3x1944 @ 24 fps
5496/3x1856 @ 25 fps

I expect this should prevent corrupted frames.

Looks like it fixed the issue. I was able to record this ~1 minute example at 5496/3x1944 @ 24 FPS : https://www.youtube.com/watch?v=ePmpH5QNuGU
Thank you! I'm very happy, going to post more tests with steadicam soon.
By the way, do you have a repository at GitHub or other Git hosting where it's possible to see the edits you made to the source code?

UPD: Another quick test, now with steadicam: https://youtu.be/Lrc_npKrWaQ

[Levas]

Good to hear it's working without corrupted frames 
I don't have the source online.
But I could upload it to my google drive.
It's mostly the crop_rec module where changes are made.

The steadycam footage looks good on youtube in 4k 

[lightwriter]

Looks like it fixed the issue. I was able to record this ~1 minute example at 5496/3x1944 @ 24 FPS : https://www.youtube.com/watch?v=ePmpH5QNuGU
Thank you! I'm very happy, going to post more tests with steadicam soon.
By the way, do you have a repository at GitHub or other Git hosting where it's possible to see the edits you made to the source code?

UPD: Another quick test, now with steadicam: https://youtu.be/Lrc_npKrWaQ

Is the steadycam a real steadycam or a stabilized lens? Looks amazing, by the way!

P.S.: I can buy a 6D with 190.000 actuations for €180 euros. I know the MTBF is 150.000. What do you think about this? Should I go for it? Right now I have the EF 35mm f/2 (the non stabilized version).

[zezic]

Is the steadycam a real steadycam or a stabilized lens? Looks amazing, by the way!

Thanks! It's a real steadycam, the model is Glidecam HD-2000, I'm not very good at it yet, just started practicing. The lens is Helios 40-2-C 85mm (dumb metal+glass, no electronics). The lens combined with 6D and this steadycam make the whole setup a heavy thing, but I like it!

P.S.: I can buy a 6D with 190.000 actuations for €180 euros. I know the MTBF is 150.000. What do you think about this? Should I go for it? Right now I have the EF 35mm f/2 (the non stabilized version).

Probably, it can be a good setup for recording some static shots and maybe even some scenes with slow careful panning without stabilization (the shake is less noticeable on 35mm than on 85mm, but you have to try it yourself to check if you can hold it steady in your hands). What's about shutter live span – honestly, I don't even know how to estimate it, I bought mine 6D with ~23k shutter count so I'm not worrying about it at all. The 190k shutter count can turn out to be a lottery and maybe it's not that important for video, however, I'm not sure that 6D is the best option for RAW videos at its own.

[anto]

Corrupted frames is not good   
Made a new crop_rec module where the max vertical resolution is lowered by 16 pixels. (So 16 lines less to read out by the sensor)

5496/3x1944 @ 24 fps
5496/3x1856 @ 25 fps

like 5k anamorphic on the eos M? there is a real time preview?

[Levas]

No realtime correct framing liveview.
There is realtime liveview, but it is squeezed and zoomed in.
So to check framing, you need ML preview (the preview with slow refresh rate)

[Levas]

For those interested:
Posted some comparison shots 6D to R5 and posted them in the R5 topic.

https://www.magiclantern.fm/forum/index.php?topic=24827.msg239189#msg239189

[mkh50]

Hi,

is there a chance to make full frame RAW more than 1824x1026? Anything around 2.5 or 3K?

[Sténo]

I was wondering where is the ettr module in this version

where can we find it on the site to install it?

[orcslayer]

Any idea what is causing flickering effect in some of my videos? Same effect during camera preview - 1X1 pixel binning 3072x1152@25fps[magiclantern-Nightly.2020Jul16.6D116] 7 out of 37 videos have this flickering effect going on.

https://www.youtube.com/watch?v=zM5plJ9ZQGM

https://www.youtube.com/watch?v=c32OHEonoMY

https://www.youtube.com/watch?v=vsMBUHrZW5c

[lyrata]

Hey mates, what speed sd card do i need to utilize the latest crop rec and 4k?

[Levas]

The build is based on the same branch as the ML build on the download page on magiclantern.fm -> experiments -> 4K raw video recording; lossless compression

https://builds.magiclantern.fm/jenkins/job/crop_rec_4k/80/artifact/platform/6D.116/magiclantern-crop_rec_4k.2018Jul22.6D116.zip

That build contains a ettr module, see folder 'ML/modules/ettr.mo', you can copy the ettr.mo file to the build on your card in the same folder as the other modules, should work.

Haven't used the ettr module myself, so not sure how good it works, I'm using raw histogram and adjust exposure to that.

[orcslayer]

Any idea why the flickering keeps happening during video in 1:1 5x zoom crop mode?

[theBilalFakhouri]

is there a chance to make full frame RAW more than 1824x1026? Anything around 2.5 or 3K?

Nope. Hardware limit.

[theBilalFakhouri]

Hey mates, what speed sd card do i need to utilize the latest crop rec and 4k?

Hi, this might be an off-topic question?
You can ask in General Help Q&A section.

Anyway, UHS-I SD card with W90 (capable of 90 MB/s write speed) should work with 240 MHz OC. Feel free to start a topic and ask about a certain SD card before buying it.

We should create a list of verified SD cards too.

[theBilalFakhouri]

Any idea why the flickering keeps happening during video in 1:1 5x zoom crop mode?

How do you process MLV clips, what is your workflow?

[theBilalFakhouri]

I think you are mixing up W90 and V90.

Ah, yes, sorry. It's W90.
Will edit the posts with W90 instead of V90.

[gherebean]

Levas et al, thank you so much for the work you've done. I used to us ML back in the day when I first started filmmaking on my 550D. I've switched to RED over the years but I've had an old 6D lying around for years that I used to use for BTS until the card slot stopped working. Turns out, using a dud SD card and spraying WD40 on the leads and inserting/ejecting like 10-15x cleans those leads enough to work.

All that to say, I've loaded your version of ML. I've dug through and read this entire thread. I've enabled SD overclocking which doesn't work for me at the highest setting even though I have the same Sandisk Extreme Pro 170 MB/s card you're using. I'll be testing it for fun with my Contax Zeiss Cinemod primes over the next few days/weeks.

I'm wondering four main things:

1) Is there a way to shoot 4K in something like 10-bit h264 without the 3x1 pixel binning? (I have a feeling I know the answer is no)

2) Related to #1, How does the 3x1 pixel binning work and how do I configure it? Moreover, if I'm understanding how it works correctly, it's taking 3 adjacent pixels and somehow averaging the color. When brought into MLV App for post production, does this do very well on parts of the frame where simple color gradations occur and perform horribly in the high contrast areas? Maybe this is better explained in one of the other ML forums. If so could you point me there so you don't have to waste your time regurgitating something that I have to imagine has been asked many times before.

3) I'm having a little trouble understanding exactly how to use the crop modes and how to correctly configure my settings to work well. Is there also some diagram that shows the crop factor in the respective modes?

4) Perhaps this is just my lack of experience with MLV App, but I'm noticing that I'm not able to get the sort of highlight retention through it as I would through taking a canon raw file into Lightroom. Maybe I'm doing something wrong or just actually totally clipping? Admittedly it's been some time since I've used my Canon so I could just be misjudging things slightly. Moreover, is there some best practices for the 6D in MLV App for various modes?

Thank you so much in advance.

[gherebean]

btw, wanted to upload a couple minutes of test footage I shot yesterday throughout the day.
https://www.dropbox.com/s/n03v5mj8p4mceok/MLTEST1.mp4?dl=0

Focused primarily on getting exposure right to look at the colors/skintones/etc. My post processing for all of these shots was as follows: loaded in MLV app and disabled "Enable RAW correction" because I initially tried a few of those things and got some weird results so I figured I should start with the basics. I went through and maybe just increased/decreased exposure a tiny bit, made slight white balance adjustments, converted each shot into ARRI LogC, and for some shots I checked the highlight reconstruction box although I'm not seeing a whole lot of difference one way or the other. On the last shot of the Old Fashioned, I played with the Chromatic Aberration dials to reduce it a bit. Worked pretty well. I notice a lot of CA on the second close up of my dog at 1:03. I didn't adjust it at on that shot. I think I also started those two close ups on the zoom feature. Didn't realize it would shoot like that. Happy accident?

After exporting in ProRes 4444, I brought everything into Premiere, I simply added the built in ARRI LogC to REC 709 LUT. Earlier on a played with a couple shots which were exported on standard from MLV and found that I preferred the ARRI conversion especially on skintones and highlight retention. I added some warp stabilization to a few shots where it didn't look crazy. I shot this all on a 50mm Contax Zeiss f/1.4 at various apertures. The CA is pretty bad up until f/2.8 so I generally didn't drop below that but if I'm honest there probably are a few shots where I did.

Overall, I'm very happy with the color rendition. I tried to use the histogram, despite it being slow, and ETTR when applicable. It's much better color than I remember shooting in h264 on it back in the day. Being able to adjust white balance in post like I do with RED footage is really nice.

Now my gripes. The first shot, I'm seeing a lot of weird color noise in my dog's coat in the highlights. It also appears on the sunny parts of the fence boards around 1:36. I'm surprised at how well it handled the rug pattern around 0:12. The highlight retention is pretty good. Look at the sky around 1:22. Do you guys think the CA at 1:03 is from my lens or from the crop mode I was in?

Any input would be appreciated.

[names_are_hard]

I can't answer all of these but I can give you a start.

1) Is there a way to shoot 4K in something like 10-bit h264 without the 3x1 pixel binning? (I have a feeling I know the answer is no)

I don't think anyone knows if 10-bit h264 is even possible, or anything above 1080.  I suspect not, very sure it's not possible currently.

2) Related to #1, How does the 3x1 pixel binning work and how do I configure it?

Binning is done at hardware level by Canon.  It collects light from multiple sensels but sends back one measurement, allowing collecting light from a larger portion of the sensor for a lower bandwidth, at the cost of spatial resolution.  I don't know how the sensel data is combined, some kind of average I assume but I don't know which one would make the most sense.  Geometric mean possibly? 

4) Perhaps this is just my lack of experience with MLV App, but I'm noticing that I'm not able to get the sort of highlight retention through it as I would through taking a canon raw file into Lightroom. Maybe I'm doing something wrong or just actually totally clipping? Admittedly it's been some time since I've used my Canon so I could just be misjudging things slightly. Moreover, is there some best practices for the 6D in MLV App for various modes?

If it's clipped, it's clipped.  Different programs will use different algorithms to try and make the clipping look nicer, but the information is just as lost.  You can losslessly convert mlv to dng and compare recovery in two tools.  I haven't used MLV App, this question is probably better asked in an MLV App thread, maybe here: https://www.magiclantern.fm/forum/index.php?topic=20025.msg243653;topicseen#msg243653

You might get better feedback about the video if you ask here (and explain you want review / feedback): https://www.magiclantern.fm/forum/index.php?board=19.0

[gherebean]

I can't answer all of these but I can give you a start.

I don't think anyone knows if 10-bit h264 is even possible, or anything above 1080.  I suspect not, very sure it's not possible currently.
Binning is done at hardware level by Canon.  It collects light from multiple sensels but sends back one measurement, allowing collecting light from a larger portion of the sensor for a lower bandwidth, at the cost of spatial resolution.  I don't know how the sensel data is combined, some kind of average I assume but I don't know which one would make the most sense.  Geometric mean possibly? 
If it's clipped, it's clipped.  Different programs will use different algorithms to try and make the clipping look nicer, but the information is just as lost.  You can losslessly convert mlv to dng and compare recovery in two tools.  I haven't used MLV App, this question is probably better asked in an MLV App thread, maybe here: https://www.magiclantern.fm/forum/index.php?topic=20025.msg243653;topicseen#msg243653

You might get better feedback about the video if you ask here (and explain you want review / feedback): https://www.magiclantern.fm/forum/index.php?board=19.0

Thanks! I will be checking out those forums to see what more I can find. I appreciate you taking the time.

In the meantime, I did a little messing around with various settings outside in the parking lot of my studio. I shot the same shots in various formats including the 6D's own h264 format for comparison. Everything is shot on my Contax Zeiss 28mm at about f/6.3 with a VND filter at 1/50 shutter. All 24fps. All 100 ISO. Nothing was changed (not even focus) and everything was shot in the span of 10 minutes. RAW shots were processed to ProRes 4444 in MLV App with ARRI LogC and then the LogC to 709 LUT was added in Premiere.

4K 1:3 Crop video: https://www.dropbox.com/s/is50a52h4r13nk7/6D%2028mm%204k%201x3.mp4?dl=0
1080 various formats compared video: https://www.dropbox.com/s/idiht9wwwk6ur44/6D%2028mm%201080%20various.mp4?dl=0

Initial findings (I would love anyone's feedback to see if I'm missing something):

1. Raw mode without crop mode enabled is far superior to the Canon h264 (not surprising). Rolling shutter is about equal in both modes and very usable and maybe a little less noticeable in h264. Moire and aliasing is horrendous in both but worse in h264. The power lines and card edges render poorly. The patterns from the shop roll up doors is all over the place. Again, all worse in h264. I wish I would've used a flatter profile in h264
2. 1:1 crop mode is the cleanest but the crop factor makes even my 28mm telephoto. that said, there's practically no moire or aliasing. very clean lines even on power lines. At 100 ISO, the noise in the image is negligible but I would be curious to see how the noise patterns did in the dark at higher ISOs as compared to a mode where more of the sensor is utilized. Rolling shutter is insanely prevalent.
3. 1:3 crop is surprisingly good image wise. If you look at my reference videos, the power lines render really well even when magnified to 100% in the 1080 video I exported. You can see some softness in how the power pole is rendered as compared to the 1:1 cropbut this is surprising good!! The results are astonishing. That said, rolling shutter is increased intensely. Is this due to slower sensor readout or because we're actually recording less horizontal pixels?

I have the Sandisk Extreme Pro 170MB/s card and my SD overclocked at 192 but I can only get about 23 seconds of recording time on 1x3 crop mode. Does anyone have a solution to this?

[HugoMouteira]

Hello!
Frist of all, I would like to thank you guys involved for your contribution on ML on 6D.  I was looking to buy a 6D because I feel its a bargain for photography, one of the best cameras that exists for that job actually. But I would also love to record RAW video. Nonetheless, one think I noticed is that 6D lacks a 5k mode, unlike eos M or even 5dIII. Could you explain me why? It has even more 2MP than eos M, why it lacks that? And btw, Bilal posted hes working on a crop mood version for 5DIII, why didnt he mention 6D also, since these 2 cameras share the same processor, like eos M shares the same with 100D, 650D and 700D?
Im sorry for the annoyance, Im a noob.

Best regards from Portugal!

[names_are_hard]

Every camera needs unique work to get features working.  There are very few devs.  If no dev owns a 6D, it's not going to get worked on.

[HugoMouteira]

Its a shame. This camera is even better for video compared to 5dIII when it comes to ISO performance. This camera should get more atrention. Even for photography is better than 5DIII in image quality and autofocus, even though having less focusing points and competes vs 5DII when it comes to image quality. It has the legendary canon color science.  I hope some dev one day could be able to replicate the eos M crop mood modes with this camera, it deserves it.