Hello!

I just discovered that all my YCbCr Rec601 standard def 1:1/1:1 10b media that I've created over the years using MC up until v2018 are being decoded with wrong colors in newer Avid MC versions (v2024, v2023, maybe also v2022, v2021, v2020, v2019 - haven't tried those). This is with both regular MXF OP-Atom located in Avid MediaFiles and "AMA-linked" QuickTime files (exported Same As Source, back when it was still possible). At around 2019 I stopped using 1:1 for mastering, so I have no idea what happens with media files in this codec if they are created with MC v2019 and newer.

This was discovered by accident when I had to use a 1:1 clip created back in 2018 with MC v2018. The EBU 75% color bars at the beginning of the program seemed not quite right, and when I checked them with waveform monitor, it was clearly visible that the colors are  off - the luminance levels for almost all colors are wrong - each in a different way: greens were significantly darker etc. Also, the progam video picture itself doesn't seem correct. Checked this with several other random clips from different years (some created with very old version of MC, even pre-2010), and was is the same for all of them. And this doesn't seem to be a wrong overall video/full range levels issue, because the black and white levels seem to be fine. It is the color components that seem to be wrongly decoded and have different luminance and hue values.

For most of my uncompressed 1:1 masters I used to create ProRes transcodes (done internally in MC) as a lower-sized alternatives, so I compared those also, and color decodeing for ProRes files seem to be fine, regardless of MC version they are created or being opened in. So the problem seems to be only with DNA-era X:1 codecs - possibly created with MC v2018 and earlier. Not sure about 1:1 media created with MC v2019 and newer. I still have one older system running MC v2018, and there the same 1:1/1:1 10b files are being decoded correclty. All colors are exactly as they are supposed to be, both the color bars and the program itself.

Seems to me that starting with some MC version (I can imagine v2019, as that was a big revamp, if I remember correctly) something has been altered in decoding of X:1 codecs.

Could ths be a bug, or is it a deliberate modification of the way DNA-era X:1 codecs are being decoded in post-v2019 releases (for some reason, unknown to me)?

Best

Kriss

....try rewrapping a file in Shutter Encoder...that wont affect the file...just the wrapper...see if that helps....Shutter has the latest DNX codecs installed so it might clear this up......

....to test, flip a file in Shutter to MP4 and see if the levels are correct, both before and after rewrap...

....good luck....

Hi Robert,

Thanks for replying!

Just wondering what would rewrapping change anything if the problem seems to be with video essence? The problem is the same regardless of the wrapper in my experience - I tried both 1:1 and 1:1 10b QuickTime wrapped files (exported Same As Source from Avid), as well as the orignal MXF OP-Atom files (in Avid MediaFiles/MXF folder) that Avid created during video mixdowns or videotape transfers (depeding on project), and it is the same all the time.

Also, does Shutter really have Avid DNA 1:1 (2:1, 3:1 etc.) codecs support? Those are nothing to do with DNx codecs.

Best

Kriss

Hi Krishius,

I created and submitted a case for you so that the team can check on this one.

For reference the case number is 04939378

Thanks!

Hi Ajay,

Thanks a lot!

Another discovery: Davinci Resolve v19 supports decoding of Avid 1:1 and 1:1 10b codecs (only when in MXF OP-Atom wrapper, no Quicktime support for those codecs). And in Davinci also 1:1/1:1 10b clips are decoded with the same color offsets as in newer MC releases mentioned earlier.

Just wondering if Resolve built-in decoder for these codecs has anything to do with Media Composer.

Best

Kriss

...please try rewrapping the file.....or flipping it to a DNX codec.....then test it....use a copy to test.....but trying hard to use these old codecs when you can easily change it to something modern is a waste of your time....

Hi Ajay,

Just wondering if there's been any updates regarding case #04939378? I could not find a way how to follow-up the status of this case anywhere in my Avid account.

Thank you!

Best

Kriss

...dude...with all due respect...

....have you even tried to flipped these files to a modern codec?...or are you waiting for avid to fix this simple issue you could take care of?

...r u a pro?

A little update:

Just discovered that the problem only is present when the project is not set to SD PAL resolution. The files I have are SD PAL 25i, so they are in Rec601 color space. If the project that I have imported or AMA-linked the files in is also set to the matching resolution of SD PAL 25i (and Rec601), the decoding is accurate. However, when the the project is switched to any other resolution (e.g. HD 1080 50i), and the color space is Rec709, this is where the problems described initially appear. And this only seems to be happening to 1:1 (possibly also 2:1 etc.) codec clips. ProRes clips (generated in Media Composer SD PAL 25i Rec601 project, so they are Rec601, too) are being decoded accurately regardless of the resolution and color space (601 or 709) of the project they are viewed in.

I do not recall this being a problem back in the day when I had to use quite a lot of legacy SD 1:1 materials in my HD projects. The color of SD clips seemed always correct, no matter the resolution of the project.

Here are some samples of SD PAL 25i EBU 75% bars in 1:1 MXF OP-Atom format: https://www.swisstransfer.com/d/4dd80dae-3a0f-41fa-bb0c-ce18dbadd23e

Best

Kriss

Robert Martindale:

...dude...with all due respect...

....have you even tried to flipped these files to a modern codec?...or are you waiting for avid to fix this simple issue you could take care of?

...r u a pro?

Hi Robert,

I have a lot of those 1:1 materials, and I also already have everything transcoded to ProRes (as a lighter alternative to the uncompressed masters). So I already have the more modern alternative at my disposal. However, since the 1:1 is the original codec of those materials, I would like to keep them uncompressed. If a way to losslessly convert the old 1:1 codec to some more common/modern uncompressed or losslessly compressed codec will appear in Media Composer, I will do that right away. For now there don't seem to be any other uncompressed alternatives in Avid, and 1:1 is the only uncompressed codec that supports interlaced standard def.

Best

Kriss

Krishius:

Hi Ajay,

 

Just wondering if there's been any updates regarding case #04939378? I could not find a way how to follow-up the status of this case anywhere in my Avid account.

 

Thank you!

 

Best

Kriss

Hi Kriss,

There should be an email that was sent to you from our tech support team. Can you check your email. If you haven't received anything, please email me direct  at [email protected]

BTW I made a follow up to the team about your case.

Thanks,

Ajay

...its good to see you solved this problem long ago....when you playback the uncompressed files of standard def material do you see any difference with a modern codec like DNX or Prores at a high bitrate?

....standard def is, at best, 720x486 and the bitrates were much lower back then....are you saying a modern codec can't support that and there is a real world difference?

...perhaps you can post some screen grabs of both?

....are you sure of this...."1:1 is the only uncompressed codec that supports interlaced standard def"....1:1 is, by definition, "uncompressed".....the question is whether a real world difference exists in a 1:1 file of standard def material and that file flipped to a high bitrate modern codec?

...i can see the difference in good old DV25 material and 1:1 but you already have the 1:1 files.....why not make them modern?

Hi Robert,

If you don't touch the material colour-wise, then there is almost no visual difference between ProRes and uncompressed. You will only be able to see the difference in scopes. However, once you start to fiddle with colour and want to bring out details from extreme ends of the signal, there is quite big chance you will start to notice some minor compression artifacts for ProRes materials. Also, very noisy/grainy material in ProRes will most likely have some noticable compression artifacts (minor ones, but still - a thing to avoid in certain cases).

Anyhow, it would be great if Avid at least added a possibility to export to standard QuickTime Uncompressed 10-bit and 8-bit YCbCr 4:2:2 codecs (standard def, Rec601, with interlacing) that have been around for ages and still seems to be supported by basically everything out there, both decoding and encoding.

Best

Kriss

AVID did try to move away from Quicktime dependency...perhaps this particular codec combination did not get addressed?  

.....its good to see you can get this now very old, seldom used codec to work in Resolve for any color grading you need....