SONY FS700 + ODYSSEY 7Q SLog2 color issues

The problem of magenta cast is not solved, after the 5.1.100 Odyssey 7Q firmware upgrade (with FS700).
The following images uses ambient light, ISO 2000, Color temperature 3200, Slog-2, 4k->HD ProRes
Left frame of each image below is the AVCHD version and right frame is from Odyssey 7Q.
As you can see, I got a quite visible magenta cast/banding on skin tones (and also on wall) on 7Q footage.
This very noticeable magenta cast is very hard to correct in post.

Graded version with FilmConvert Plug_in (left image is AVCHD and right is from O7Q)
A.jpg

Ungraded version (left image is AVCHD and right is from O7Q)
B.jpg
 
Last edited:
7qtest.jpg
I did this test where the top of the image is the 7q footage and the bottom the internal avchd. Left to right: indoor preset, outdoor preset, manual 4500K.
It clearly shows that nothing has changed. Altough not as extreme, the weird color gradation is also clearly visible in the preset WB footage. Also it is even visble in the VERY flat slog2 profile and as you can see in attamburs image it magnifies a lot when graded to proper contrast / saturation.

Mitch I have to say I'm quite disappointed that instead of telling us what's really going on you just tell us to test if anything changed with the update and that those problems are "due to something that you are not at liberty to say at this time". It really makes it look like you guys have no clue how to solve this problem. Also I think you should make your customers aware of this issue so they can test before a shoot if this might screw up their footage.
 
I believe what you're seeing is a fundamental problem with the sensor on the FS700. Sony, has, apparently, dialed a filter into their debayering to mitigate this issue before recording in-camera. If I'm correct, C-D has not addressed this issue in their processing. I might suggest that there is a workaround, in post, to bring the O7Q files inline with the Sony files. Try this:
https://www.youtube.com/watch?feature=player_embedded&v=kYg5OlxrYZs
 
cuervo, I don't think so. Raw footage looks perfectly fine. It's really only visible in the Prores on the 7q.
The workaround you're suggesting doesn't help this issue as the chroma bands can be quite large and not pixel size like the pink fringes on 5dm3 raw.
Even at small radii chroma blurring has quite negative impact on overall quality and makes chroma keying for example impossible.
Btw as of Resolve 10 the pink fringe issue has been fixed for 5dm3 raw footage.
 
Well as I said - even though I was rightfully critisized for it - "some improvements" didn't sound promising at all
and seems like I was right about it. Hopefully this can be fixed soon, cause it really is hard to get rid
of it in post. This should be a Nr. 1 priority over at CD.
 
s---70 said:
cuervo, I don't think so. Raw footage looks perfectly fine. It's really only visible in the Prores on the 7q.
The workaround you're suggesting doesn't help this issue as the chroma bands can be quite large and not pixel size like the pink fringes on 5dm3 raw.
Even at small radii chroma blurring has quite negative impact on overall quality and makes chroma keying for example impossible.
Btw as of Resolve 10 the pink fringe issue has been fixed for 5dm3 raw footage.
Click to expand...

I hope you're right. These magenta issues have been around just forever; and, not just with the O7Q. Magenta fringing, in particular, not necessarily banding. It's not CA, but, some peculiarity of the compression schema that's particularly sensitive to overexposure. Or perhaps that's another issue. It's curious, though, that this keeps cropping up.
 
At least for me the issue appears on certain area of exposure, and the problem is the biggest when skin is slightly underexposed. That is a huge problem if I like to keep for example some part of the face darker and other correctly lit. The gradation is ugly as hell because of this reddish/magenta cast.. And okay, I can get the color cast somewhat corrected but the tonal change from light to dark is quite bad. Again, this was promised to be fixed in the firmware update and it seems that it's not, but hey, maybe I'm just being too critical.
 
If you guys can resolve this issue in a firmware release I will be so thrilled. It's top of my wish list too!

I can also confirm this problem never shows up for me when shooting 4k raw.
 
If this issue with the skin banding is resolved it will stop me wanting to sell the whole deal and buy a new camera system altogether! Would be rather lovely.
 
Is there anything that can be done in post to fix this issue? I have 44 days worth of feature film shoot that have been shot with the Odyssey 7Q, with all the firmwares ever released. Im starting to figure out how we do the post work and need this addressed very soon...
 
Guys, give us a chance to regroup. You don't realize how NAB completely takes over a company. When I have information to share, believe me I will.
 
Sorry, this has been discussed in several threads so I lost track.

We delayed our upcoming firmware release as we are directly addressing these issues. We believe we can either fix them or massively improve them. We have considerable progress already and believe that it should only be a short time before we can release the new firmware with the fixes in place.

I must note that the FS7 and the FS700 are different machines with different encodes on the output signal. This means that we have to treat them differently and the results or one may not be as successful as the other. That's not an indication of anything other than we are still in progress.
 
You must log in or register to reply here.
Back
Top