ACES errors in PQ ST2084

[Alex Winker]

I am working on mastering a project inPQ ST2084 at 1000 Nits. It is a short documentary shot on a couple different formats. but mostly all very high quality captures, with the exception of a few outliers. These outliers are my issue. Some of the material captured and encoded already in Rec709, when tagged with the rec 709 IDT and the PQ ST2084 1000NIT ODT produce really harsh and garish highlights near areas that would be peaking in Rec709, for example a drone shot with car head lights in it, or some of the archival materials that have clipped whites. When monitoring rec709, and using the rec709 ODT the highlights do not go so disgusting, but running a Dolby Trim for the down convert still leaves them clipped and gross. 

 

Has anyone else encountered similar errors with ACES when working in HDR spaces? Any potential solutions? My current solution is going to be to run a sort of fake aces using CST's at the node level so I can try to correct the footage before the IDT to neutralize the garish clipping then do the rest of the grading in ACES space, however this seems like it will be a lot more work just to handle a few problem shots.

[Stefano Zordan]

Hi Alex,

Highlights issues on ACES are actually very common. This happens most of the way because the colors that are out the rec709 gamut or very near of its bounds are reinterpreted wrong from ACES.

I don't know if this will solve your problem, but one fix I'm finding myself using often is the "Neon Suppression LMT" on AP0 processing mode.
This can probably be fixed with some custom DCTL's too, so I'll be interested to know too if someone else has another solution to it.

Edited September 4, 2020 by Stefano Zordan