What’s the difference between scene-referred and display-referred grading?
Stefan Ringelschwandtner, on his website Mononodes, explains it very clearly with simple diagrams. He writes:
“The easiest scene referred workflow is to transform the image from LOG to Rec709 color space at the very end of your node tree – e.g. Timeline level. And do the grading “underneath” it in the LOG space of your camera. Before the CST / Color Space Transform.”
Whereas display-referred involves immediately converting the footage into the display color space and grading on top of that.
So what are the practical differences, or is it all just theoretical?
In a video on his YouTube channel where Darren Mostyn explains CST color management, the colorist demonstrates both methods. He begins by converting ARRI Log C to Rec.709 in the first node then, in the second node, he proceeds to increase the exposure using the Offset wheel in DaVinci Resolve. He points out that we’re hitting the boundaries of Rec.709 very quickly, and we can see that the highlights are clipping badly in the parade scope. Mostyn then begins a new timeline with the CST as the very last operation, and, grading underneath the CST, he again increases exposure with the Offset wheel. This time, the highlights magically aren’t clipping, they’re rolling off nicely! Which is the reason Mostyn says he prefers to grade underneath the output color space transform. Which all sounds super convincing – until you look closely at the Offset wheels and realize that not only is there a fifteen-point difference between the two – 62.25 for the display referred, 47.35 for the scene referred – but the display-referred picture is also unmistakably brighter!
How our own scopes look in scene-referred vs. display referred.
Header image courtesy of Stefan Ringelschwandtner. Be sure to visit his website!