Unfortunately some people are just focusing on the ACES color rendition issues in 3d apps, meanwhile missing the fact that ACES is not just about rendering a picture but ACES is a full blow color pipeline between camera suits, compositing, editing, texturing, rendering and delivering shots.
Again the definition of this pipeline has been satisfied by OCIO alone. ACES by large has two parts, the color spaces and the view transform. The colorspaces are all based on AP0 and AP1, both of these are outside of the spectral locus. AP1 in particular is extremely similar to BT.2020, but with added problem that AP1 can no longer be considered to be âcolorâ, and thus unable to move away from old CIE 1931, and move forward to CIE 2012 standards etc. BT.2020, on the other hand, is defined spectrally, therefore it is one of the easiest spaces currently available to use CIE 2012 standards. This is the reason why we choose to use BT.2020 in the end for AgX, we need the updated CIE 2012 standard for some luminance correction, where the 1931 standard failed.
The view transform, as someone has already said, the ACES people are well-aware and they are actually developing several candidates for ACES 2.0, but at least for the current official release, they are still using the old one with problems.
3 Likes
I understand the inner workings. Thereâs not that much there other than nonsense.
This is flatly false.
Folks who actually run TD pipes will verify this. It doesnât work. Everyone knows this. And the TDs who work on large projects where the interchange is mandated by one of the larger two proponents, will be the first to tell you it does not work in any way, shape or form.
Not sure where you are getting your information from.
Stop with this nonsense.
5 Likes
Lets both receive RAW files from 4 cameras, lets both render on our choice of linear rendering space, lets export in our choice of colorspace, lets composite in any space we chose, and letâs deliver back in a colorspace that the client prefers which might be BMDFilm, ARRILogC or a straight output to sRGB or Rec .709. Then compare and contrast the effectiveness of ACES as a color management pipeline. I will be using Clarisse and not the broken OCIO implementation of Blender. I completely agree the ACES fucks up the color. No question about that and I have told you that time after time but at least itâs the right step toward a color pipeline. Scattering my workflow by using LUTs or CC24 shots is not ideal for shots that are on tight deadlines and/or large in quantity. If you are not willing to go through the challenge or find someone that will go through this challenge with me, then you are not seeing it from my side of the world. Anyone thatâs not in VFX should stay away from ACES; FOR A DIFFERENT REASON. But anyone that is sane enough to attempt Visual Effects in the sense of integrating CGI Elements into Live Action footage across multiple cameras and shots without ACES is either old schooled and is cruising on muscle memory, has an already predefined pipeline that has been around for more than 10 years and isnât obligated to orâŚis just outright arrogant. For younger folks like myself, itâs very hard to understand color science since there is a massive inconsistency in the industry. There are smart people like you who are hating on ACES and there are smart people that are WORKING on ACES, and even smarter people hating on the entire digital image acquisition, processing, compression, transfer, and display. When AgX is available natively in DaVinci Resolve and Nuke I will trust it enough to stand my ground against clients and say, âWe wonât use ACES, we use AgX, it works better and I will not take no for an answerâ. Even then it has to live up to the hypeâŚnot just slightly better for the technically savvy âcoloristsâ that are mad about decimal points in the hue shift for a shot where the compositor will absolutely destroy using passes and edit the colors anyway.
1 Like
If this were the case, Iâd support the effort. It is simply false, however.
Cameras are unique observers. Being unique observers, it means that their lack of their alignment to the CIE colourimetric Standard Observer never leads to actual alignment. Ask anyone trying to render CGI mixed with footage; ACES does not help here. Origin encodings do. Non uniform alignment processes do.
And this doesnât even scratch the surface of actual control over the picture formation, which is another issue altogether.
Folks I know in the actual pipelines also seek to stay away from it, and are acutely aware of the abject failure. Thereâs a reason there is a frantic attempt at a version 2.0. Folks know it doesnât work.
You make it sound as though this wasnât properly negotiated in all of digital. It has been. ARRIâs pipeline rules the roost as industry standard, and all of these pipelines have had folks actually construct alignments for digital cameras.
The problem remains unsolved. ACES does not make this any better, and in fact worse. The removal of the origin encoding sacrifices the known limits and as such, a ColourChecker24 might barely hold up, but literally everything else falls to crap.
These are real problems faced day to day with integration. ACES doesnât help here.
From the broader âBetter transformation managementâ perspective, I would say there is always room for improvement. I suspect no software handles it properly currently, and again, because things are so broken with respect to ACES etc., folks just donât know or care.
After all, thereâs a difference between paying the rent and generating quality imagery. I speak to the latter.
I really try to focus on design from the standpoint of the software. Blender is unique, and it should remain focused on the primary audience. Folks are already confused enough with the existing configuration, that had ACES nonsense in it forever, and it fell stale and led to countless problems for folks who thought it was working.
The key point; if one is doing actual work with Blender, the only solution is to use the configuration with the show LUTs. Nothing will change this. The other option is to download and install the configuration, which is already doable.
As someone who knows more than a healthy amount about this nonsense garbage, only pain comes from attempting to accommodate all parties, and the folks who are knowledgeable enough to be contracted out to deliver nonsense in ACES, are also capable enough to understand how the show configuration needs to be installed. Not that it matters, because itâs a tell if a show uses ACES; they just donât give a shit.
4 Likes
wondering how I should go about setting up before I start a project, I intend to work in Resolve. Is there an accurate and dependable guide or video? It seems the topic is deep and nuanced, donât want to sift through youtube to find iâm following the wrong advice in the end.
To date, there is no existing pipeline that seamlessly integrates CG and footage of the box. Achieving a well-integrated result requires a complex skill set that extends beyond color accuracy. Compositors are professionals who specialize in this area. While the use of ACES is not without its challenges, it is designed to provide consistent color management across DCC platforms. Comparing the results of ACES with live production assets would be a more accurate measure of its effectiveness in real-world scenarios. Instead of sending belligerent remarks at ACES, a more constructive approach would be to create informative content that showcases the drawbacks of using ACES (specifically since you want everyone to stay away from it, you have to show why it is utterly a deal-breaker). It is important to foster an environment where people can experiment and learn on their own without fear of judgment. There are VFX schools and studios of all sizes that use ACES as part of their pipeline.
4 Likes
Itâs just very sad how everyone looks at your status and contributions and uses it as an excuse for listening to every harsh and toxic comment you make about ACES instead of going out there and discovering for themselves. I personally learned most of my color science and digital image processing from Charles Poynton from FXPhD. He went very deeply into color science on DCT - 102 and âAdvanced Color Theory and ACES Workflowâ. It is very hard to imagine that everything I learned was bogus and false to the degree that you are describing ACES to be. Anyway have a great day! Looking forward to AgX being the industry standard one day or you working with the team at ACES to improve ACES 2.0 or 3.0
1 Like
Only dipped my toes into this topic but I believe most of the criticism here towards ACES isnât because of the wide collection of input LUTs ACES comes with (theyâre useful), but the display transforms.
There is no way to sugarcoat it, those are shite.
Nah, this is not just affecting very few âextreme casesâ, especially since most VFX-heavy Marvel/DC shows are suffering from this. Most of those movies or shows have a lot of shots with neon lights, lasers, explosions, or sunsets.
Godzilla vs. Kong, how in hell can this be in a trailer like this:
some other examples from the same movie:
Obi-Wan Kenobi:
The first step is always to educate the people.
Because clearly, most people (working as professionals in the VFX and animation industry) have no clue about this topic whatsoever.
Every single pixel gets overanalyzed in endless pixel-fcking sessions but huge eye-soring artifacts survive completely untouched until the end when people see it on tv or in the cinema.
And the VFX people that actually see the problem are prohibited to do anything about it because:
- the use of ACES is forced by the client
- the old VFX mantra of ânever change the plateâ and therefore matching the CG to it
As someone working in a VFX studio, the only way to combat this is to secretly and slightly desaturate the highlights of the CG before it goes to the client to avoid those artifacts of the ACES display transform later in the pipeline.
As someone outside the VFX industry (where the use of ACES is not mandatory) we thankfully can use better display transforms like AgX.
8 Likes
Jesus wept , this person i am replying to is impossible, all of your concerns as well as informative resources that you are looking for to show the pros and cons of ACES were already linked, in previous posts, MULTIPLE TIMES (No, i will not re-link them again myself, But i will refer you to troyâs post, post #23 within this topic, as a start).
As well as the usage of ACES to transfer data around and actually using it in production pipeline, read above the whole 270 posts, i did, so can you, enough with this circular back and forward to things that were answered 7 times in one topic.
Nobody thinks that way, not even Sony with OCIO nor hollywood with ACES, they both tried to solve different problems CG and filming industries were facing.
You donât wake up one morning and say : âToday, i will create a standardâ, things get adopted as fitting to become a standard out of demand, cost and licensing, in our context, the demand is to solve technical issues regarding color management (Even standard institutions are formed to solve problems).
It seems like you have a lot to say in this regard, how about some links as references for the rest of us to study?
I repeat again, because it seems to fall upon deaf ears:
It does not do this and does not work. Full stop. End of story.
It is a broken and poorly designed conceptual framework.
3 Likes