ACES workflow in Photoshop
-
Hi Dr Sassi, I was previously using the C4D ACES to Photoshop workflow here in the redshift manual using the OpenColorIO plugin for Photoshop:
https://help.maxon.net/r3d/maya/en-us/Content/html/Compositing+with+ACES.html#CompositingwithACES-ACESinPhotoshop
However I recently updated all my adobe software (removing all plugins too by accident) and it seems you can no longer get this plugin - it used to say MAC or WINDOWS download on this page but now its just the source code as far as I can see with no plugin file.
https://fnordware.blogspot.com/2017/02/opencolorio-for-photoshop.html
Im guessing this is because ACES was implemented natively in Photoshop and After Effects - I can see the new After Effects method in the Redshift Manual but no updated Photoshop method.
I've changed the color settings to ACEScg (screenshot bellow) but not sure what else to do to get it looking right? A display transform is needed I guess?
https://www.dropbox.com/scl/fi/fvxl4y2lo9ixkhk9rz1f4/Photoshop_XkmcXdMKlV.png?rlkey=r7jy1wapo5n8nvmg275jv2asp&st=wf24cgkk&dl=0
Cheers
-
Hi MaverickMongoose,
The image you shared is the set-up for the "engine" inside of Photoshop, which means how large the color space is in which you work. This will affect all images that you use. But doesn't set any information for eh image you might open, here the primaries (Colorspace) must match 100,00%
The ACES guidelines describe only ACES 2065-1 as being "flagged" inside the file as ACES, meaning the compositing app should be able to read it. I have written to Adobe a few times about it, by the way. Photoshop doesn't read this (at least with my 25.0.0 and anything before that).
There is no rule set for the ACEScg. If it is an ACEScg file, to begin with, you need to make sure that you set it (not convert the file!) to ACEScg.The Display stuff should be taken care of by Photoshop. For a decade+ or after Linear light workflow was established, that works. Of course, not if the engine is set to sRGB (your image shows the right one. Perceptual Conversion should be fine in most cases for beauty passes, etc.)
But the problem with anything that large (9Gamut) and high in dynamic range is that you need a monitor that can show you at least 1000 nits and DCI P3 as color space. If your monitor doesn't have that, you might not get exactly what the screen shows. As a side note, for best practice, the monitor should be calibrated, and the light in the room should be as it was during calibration.That being all set since I write in a forum: there is a lot of nonsense on the web about ACES, that it looks more filmic, etc. This might look that way when the RRT/ODT produces an REC 709 and 8bit/channel result. Then Gamut Compression and Tonemapping deliver that, where all colors move. If you deliver the same content as ACEScg to UHD REC 2020 and they have a laser projector close to cover that color space, the tone mapping, and the gamut compression are not typically applied. The result will look different. Here is the most confusion about ACES-Look. The idea of ACES is to keep the cinematographer's intention, not to apply a look.
Sorry for the long text. That point is often unclear, and then the most adventurous tips are given to keep the Client logo colors, producing more problems than needed.
The main tip I can give is to clarify what primaries are and why things change so much along the pipeline when one or the other set of primaries is used. It is often not the Display transfer function that causes that.
My background in Color Grading/Color Science:
Beginning with the announcement of ACES in 2010/12, I have taken many color science courses (Dr. C. Poynton, FXPHD) and several ACES courses from Dado Valentic, a Los Angeles colorist.
https://www.colour.training/masters-program-summer-2024/
I took this one above last year and many others, I moved through the whole Summer Academy for professional Colorist last year and finished top-level (3) with over 100 submissions. Earlier this year, I took the ACES 2.0.
Not to mention each book I have read since.Cheers
-
I usually use After Effects or Davinci resolve, set the working space to ACEScg and add a conversion at the end to sRGB so I can see it correctly on my monitor like this https://www.dropbox.com/scl/fi/ayx0wy1k9sx8kwjmtjh1t/Resolve_JKcaqMiNQk.jpg?rlkey=sk06k9bsvonqictx9s19kv74x&dl=0.
However I need to do some image retouching on an ACEScg render in Photoshop, just not sure how to attack it to have the extra ACES color gamet but still be able to export to sRGB.
But are you saying that all the hype on ACES is unfounded and that I'd be better off going back to the old Linear way of working? I thought the quality of my images had improved since I switched to ACES but your right, getting the exact client-supplied colors are super difficult.
-
Thanks for asking, MaverickMongoose.
In short, ACES is great, and the shortest answer to going back to LW, how have you handled gamut-compression and tone-mapping back then?
OR was the screen sRGB, and didn't show all the color and brightness clippings? I hope that doesn't sound pretentious, not my intention. If so, sorry!I hope I can stop myself soon enough. Here is my apology already
Please note that is my personal opinion and not an official idea of Maxon-One. Ask ten people about ACES or LW; you might get 11 answers.The idea and need for ACES were based on the growing use of digital cameras on film sets. Some sets use dozens of cameras, each with different Color Science and Sensor response characteristics.
The core idea was to define an ideal camera and screen while allowing any data to be stored in the ACES 2065 Color Space.
Since then, each camera manufacturer has produced an IDT (Input device Transform) to get the data produced as close as possible to the"Ideal Camera," with the target of mixing and matching any camera used on a set.
Which leaves the constant match of which camera to which discussion.
Following a rigorous and precise conversion into the ACES space, the exposure is expected to be clearly created. Otherwise, we would start to match this everything again.Notice what is missing already? Yes, whatever you do, no light meter, gray card, or even precise Color measuring (Sekonic C800, for example) is given. Virtual MacBeth charts are kind of useless in this regard, as they are based on RGB-three channels, not pigments with the exact color, but that matters. That is a long story, but the reason for many misconceptions I see on YouTube is that people measure, perhaps via Eyeballing or tell that when it looks good, it is good. If that would work, we could print out on a CMYK printer color charts [sic], but that would only show how well we reproduce a specific Cyan, Magenta, and Yellow, nothing in between.
I should stop here, even I have perhaps only touched on a few percent of the story. Yes, ACES 2.0 will be more supportive here.
But a few more points:
In 3DCG, we do not have to match cameras. My observation: no one cares about perfect exposure, comparable exposure, color temperature (matching all parts), most of the time it is eyeballing.
Moving forward, we got our stuff done (LMT, Look Modification transform), and we like to get it out. Then, all the problems start when a small space is the target. DVDs were typically in REC 709, a 30-year colorspace. It wouldn't be my suggestion, but widely in use. However, for ACES, the main idea is as before Cinema/Theatre, with projectors that cost six digits, which get calibrated more often than most people do with their own central viewing devices. So, not all the problems ACES solves, but having all the trouble when going into a small space? What's the point?
I answer that any REC 709 production will look pale and outdated when people get used to having well-timed HDR content when the screaming highlights and bleeding saturation calm down.After this long introduction to your question, "Why not stay in linear light workflow?" Real linear light workflow is XYZ/RGB space (Absolutely not sRGB) and is based typically on floating point with no limitations. Nuke has this sorted out perfectly, for example.
How has that gotten material into SDR and sRGB/REC 709?
Perhaps you never were in real Linear Light space and used just like sRGB, the slightest color space currently used for production. Then the conversion was often less problematic. Sorry if my assumption is wrong here. Go to Cineversity old and check any project file older than five years, the color space is most likely sRGB (99%)! This is not LW.
Run things in DaVinci, the DaVince Wide Gamut, or ARRI C3, which is the typical transformation when people don't use ACES these days, and compare the results while setting up the Pipeline. IF you feel well there, this is a great place for color.
Log footage is more appropriate when you use DaVinci, but I'm not sure if that will work well in the future.
Example
https://community.acescentral.com/t/do-we-need-to-change-acescct/5320
How have you "tone-mapped or compressed the given data from Linear Light Workflow (LW) into REC 709? If that has worked for you, ACES is linear and floating point. So write ACES 2065-1 out, the Storage and Exchange format, and you have (compared to REC 709) very little difference to LW. From there, go the regular route. However, if the LW was based on sRGB or REC 709, it was not LW.I believe that ACES is an excellent concept for teamwork and multi-camera production, but is it the best? I do not think so.
When I develop footage from my RED or my Canon log raw footage, and I do not need to match it with anything, why move out of the native gamut so early?The idea of ACES and 3DCG and Compositing is way more complex than Camera matching in a pool. I am all in if all parts were, from the start, developed in a large gamut and preferably Linear, while also, for all components like textures and HDRI, the color temperature needs to be known. There are so many parts that I could fill days with it, and I believe no one wants to listen after a few hours
The best idea is to use where you feel savvy with, and what will hold up in the next ten years on developing screen. Clients don't like to look old with their visuals.
Chees
