VSE and x.v. color


I’m quite beginner with these so this might be stupid question but:
I have AVCHD videos (.mts) recorded with x.v.Color (aka xvYCC; a color space that allows 1.8x bigger color gamut than sRGB) enabled and I would like to edit them with Blender VSE. The problem is, that I don’t know if Blender(or OpenColorIO+ffmpeg) will handle the colors correctly, or will the color space be reduced? Video preview looks ok but in chroma vectorscope view colors seem to be somehow blocky, is this normal?

How should I input and output the videos from Blender to preserve the extra gamut?
Should I play with OpenColorIO LUTs/codec settings or something?
Currently I don’t have a screen which would support x.v.color so unfortunately I can’t test this in practice. :frowning:

Can’t answer core issue but the vector does look aliased to me (meaning missing or truncated colours). You can ask the VSE to process each strip as a specific input type, see strip properties (N key):

And apply linear modifiers per strip for colour correction etc.

Blender uses the swscale component to convert the video to sRGB at 8 bits per pixel. So you aren’t looking at the raw data format.

As above.

You cannot. You may be able to deal with it, but it will require a little knowledge.

It won’t make any difference as the footage is sRGB as above.

You don’t need one. You only care about data fidelity for this.

While I am not familiar with this specific codec, it shouldn’t be too complex to peel out the low data and recompose it at full bit depth.

You could either use FFMPEG to dump deeper already sRGB footage as a start, or dump raw YCbCr and handle the scaling and transformations yourself.

If you can figure out how to dump the raw YCbCr to still formats, and assert the lower level nature of the particular footage, I might be able to help you figure out how to get it into viewable sRGB, while leaving it in a wider gamut space.

The typical process would be to dump the raw planes, figure out the scaling on the planes, scale them all up to the same resolution and combine them using a proper transform. At that point you can leave your footage