@michael_scott , you said in your post you didn’t want to make it about comparing ACES workflow to other color management approaches, which is fair enough, so I’ve made this separate post about ACES, and had a go at reproducing your image FWIW.
You say “Developing RAW files in Resolve using ACES gives me really beautiful color representation - very close to real world.” I’m not at all a colour science expert but I’m interested in how ACES might work so well. If I or someone posts a more “typical” shot - straight daylight, varied colours etc. rather than your image - could you show the ACES/Resolve version perhaps?
I’ve had a play with your raw file in darktable. It looks like you shot with a manual white balance. Darktable thinks 3564K. I don’t suppose you remember what you set?!
You mentioned the candle. Darktable has a Defringe module which improves the look of the candle and lights. Anyway I’ve tried to reproduce your Tiff image in Darktable, result below, with de-noise at default values. Also the defringe. What do you think? I appreciate you like Resolve and don’t want the learning curve of new software.
Your tiff converted to jpeg (hopefully no change) -
I have not gone back to that thread but if I recall there was a comment about comparing DT with Resolve raw processing after only changing exposure and white balance… I’m pretty sure that Davinci applies a tone mapper as part of the default processing or it can and I suspect the OP meant default DT steps but if not DT would be missing say sigmoid/filmic in his comparison and that would certainly be better than without it so that would have to be confirmed… also I think Davinci has a panel to specify what you want done when it opens several types of files including float etc so I think you can specify how to handle something like an EXR or float from an external program but in any case my first point would have to be confirmed followed up by is there really a need for the OP to give up his work flow if it starts with a raw or a linear high bit or float file type as it seem might be the case from his comments…
From this configuration box I think you can set up the raw decoder to apply your desired gamma and color space…if you can’t do that for the alternate external file it looks like the color space transformation nodes have basically the same controls…as the ones found for the raw project settings…
Absolutely There should be no problem in doing so if the camera (which captured the image) has its ACES IDT present in Resolve (for example you won’t find Nikon there).
Yeah, not really, its some older photo I found just for this denoising example.
I just want to say I would happily use darktable if it would provide the ACES-based color management! That would be just perfect in my opinion. Working on photos in Resolve presents many challenges (like landscape and portrait photos - I have to rotate my grading monitor each time orientation changes ), the only reason I stick to it is purely for the color representation.
When looking at the darktable version, I feel like there is still a point at which luminance drops off faster than it should, bringing also this stronger orange saturation with it.
In my opinion the ACES version with sRGB ODT has very smooth highlight transition - all the way down through the candle (without introducing unnatural kick in saturation as the luminance drops below certain point). When thinking about SSS-based light transport, I feel like the ACES version is what is should look like in real life. But maybe I am wrong, its just my opinion on this.
But anyway if we were to step back a little, the main point in ACES is that you don’t have to worry about this stuff. What we are looking at here is just a problem of how to represent HDR data in SDR (in this case pushing whole dynamic range into a classic sRGB monitor). Even though I can dive deep into darktable parameters and tweak it so that the candle would more or less the same, I would be just doing some “tricks” on top of its default color management to make it appear realistic. Once I change the exposure (or the whole photo), it might go really wrong and I would have to tweak parameters again to make it look as before. In my opinion this should be handled by some better color pipeline, not by me. Among other concepts, ACES solves this problem for me - its ODT’s can nicely represent the image in given target space (like sRGB). It doesn’t matter how much I change the exposure, it just holds up (and I am free to work on top of it without worrying about breaking stuff).
More detailed (but still easy to follow) explanation of what I am trying to say could be found here:
For those interested here’s the result RT Selective editing > color appearance gives. Did this using the apse-improv branch but that’s just because it’s what I had running. Brightened the image a notch by setting scene conditions > surround to dark then changed the source data adjustments a bit to control “exposure” and then sigmoid for highlight attenuation and contrast. I’ve not tried to go for anything just ran the tools. WB is the automatic temperature correlation method that gives slightly different results.
There should be no problem in doing so if the camera (which captured the image) has its ACES IDT present in Resolve (for example you won’t find Nikon there).
), the only reason I stick to it is purely for the color representation.
