Proper Conversion

Hi guys, I’m new to color grading, I have a problem of not knowing the camera that was used to capture a scene. What’s the best approach towards getting a proper conversion?? Thanks.

Ideally this would be in the file metadata, but it is often not. Can you say more about the file type (EXR, DPX, MXF, Prores)? Can you tell if the file is in log? Best of all, can you ask the person who created the footage?

In a pinch, if you can determine that the file is in log (you can usually tell because log footage looks characteristically washed out), then you can pick an IDT to convert from the camera native log space into the ACES working space, in the case of grading ACEScct. You just need to try them until you find one that looks the most plausible.

1 Like

Hi, Thanks for the reply, the files are in Prores 422HQ format and it’s not log as the files are not totally washed out.

If previewing them in explorer/finder already gives a ‘normalized’ image with decent contrast and saturation it could be that some Rec.709 LUT was either baked in the exports or captured like that if they are the original camera files.

If you have no way of getting the files in camera log space you need to tag them as Rec.709 and grade from there. It won’t be as flexible as log so heavy looks or balance fixes will be tricky but it’s all you can do at that point.

1 Like

Thanks for this Shebbe, I will try out some few things.

If I don’t know what color space was used and client also can’t exactly tell me that, I usually find a frame with something really dark, to get an approximate level of lens cap black. Then I try different log curves and different data levels (full/video) to see what combination gives me the blacks closest to 0 on waveform analyzer, but not crushed (or crushed not too much) below 0 level. For monitoring the black level I use Color Space Transform OFX and set its input to different popular cameras logs, and its output to gamma 2.4 and untick OOTF checkbox. This lets me to see blacks in detail, because they are stretched vertically by a power law function (gamma 2.4) and lens cap black is mapped to 0.
When I found transfer function, I just try to choose the most natural looking conversion for primaries. There are usually from 1 to 3-4 options for primaries to choose from, when I know the transfer function, so it’s not that hard.

2 Likes