Notice of Meeting - Output Transforms Architecture VWG - June 14th, 2023

Output Transforms Architecture VWG Meeting

Wednesday, June 14th, 2023
1pm PDT Add to calendar (.ics)


Dropbox Paper homepage for this group:

NOTE: We will be using Zoom for this meeting.

You may join via computer/smartphone (preferred) which will allow you to see any presentations or documents that are shared, or you can join using a telephone which will be an audio only experience.

Please note that meetings are recorded and transcribed and open to the public. By participating you are agreeing to the ACESCentral Virtual Working Group Participation Guidelines.

Audio + Video
Topic: ACES Virtual Working Group Meetings
Time: This is a recurring meeting Meet anytime

Join Zoom Meeting

Or follow this link: bit.ly/acesvwg

Meeting ID: 993 4048 4145
Passcode: 972426
One tap mobile
+16699006833,99340484145#,*972426# US (San Jose)
+16694449171,99340484145#,*972426# US

Meeting ID: 993 4048 4145
Passcode: 972426
Find your local number: Zoom International Dial-in Numbers - Zoom

To save everybody from my bad and slow spoken English (which I guess was a torture for ears) in the upcoming recording of the meeting, here is the summary of my thoughts after using current ACES 2 candidates on a few projects.

I’ve used it on a few commercials and music videos with variable success in terms of skin-tones rendering. One music video was shot in front of white walls and a white floor and was lit by a giant soft box located on the ceiling. It was lit by either neutral or pinkish light and the director wanted it to be graded to various colors to simulate different lighting conditions. I did it by white-balancing and/or creative RGB mixing in linear “gamma” and native camera gamut. I said “it was lit” just to not keep speaking even slower and longer than I already did. But in fact it was done in post.

For scenes that had to become blue-magenta ACES 2 DRT looked really good. The white walls still looked the color I wanted, not collapsing towards either magenta or cyan depending on the color tint I was adding, compared to using per-channel LUT (ALF2) which made it harder to control these skews. In this case ACES 2 DRT looked unusual on the waveform analyzer, showing the blue channel clipped at the top end instead of being soft-clipped, but the actual image looked nice.

For scenes that had to be strong yellowish-orange ACES 2 DRT looked really bad though. While it kept the white walls being yellow without any skews, the skin tones looked dead and dark hair even felt greenish. While ALF2 looked very nice making dark hair to look more reddish and the walls more yellow. It actually had the same unwanted behavior, collapsing highlights again, but to bring it back from greenish feeling yellow to a warmer orange-yellow was way easier than to make skin-tones and hair look good under ACES 2 DRT. So for warmer looks I used ALF2.

For neutral lighting I also used my custom made LUT, which even though has sort of a path-to-white (also intentionally non-linear as hell), still uses per-channel tone-mapping.

So, whenever the skin-tones had to be placed on the warmer side of the vectorscope, I always used per-channel on this project. For me the weakest part of ACES 2 DRT is the skin-tone rendering that is hard to sell to a client. I think ACES-provided stock look LMT is a must. Happy coincidence made per-channel tone mapping perfect for skin-tones in SDR delivery. But I don’t know how to emulate this under ACES 2 DRT by regular grading tools without adding artifacts. 99% of the directors, DP and producers I work with have the same requirement: “Please make it look like it was shot on film, but colors only, not so much grain and blur”. I’d say not “shot on film”, but “printed on film”, since the print is what mostly forms colors everybody likes.

What I also like about ACES 2 DRT is its gamut compression after display gamut conversion. It feels like a path-to-black. It’s forgiving for quick and dirty setting per-clip flare level using the offset wheel in linear “gamma”. It’s nice to see on the waveform analyzer how the saturated colors are not clipping before the neutral colors reached 0.0 (which is typical with most of the common DRTs), but soft-clipping instead.

1 Like

The recording and notes from meeting #106 are now available.

1 Like

Some of the desmos graphs I showed are here: