I've spoken to a couple of professional colour graders about this issue, or at least what I think is this issue. They both use, and staunchly defend, what Resolve does. Both have indicated that the correct interpretation of the source material, its colour and data range etc. is misinterpreted by many programs. I've stopped using DPX sequences as my source files, principally because I can't see any tangible advantage over them in the real world over a well-rendered ProRes 4444 full data range, and because, for a feature- length film they are extremely unwieldy (too many files in a single directory for even the most modern file systems to handle efficiently). DoM, for whatever reason, misinterprets at least the data range (I think) of the source material when created from a Prores 4444 UHD output from Resolve 14 from two completely separate professional grading set ups (both using Resolve), one based in the UK and one in the USA. I cannot produce a satisfactory DCP from the 4444 UHD Prores supplied by either facility. The blacks are crushed and overall the picture looks too dark. If I use a 2K 422 Prores HQ made by Resolve, re-rendering from the same UHD 4K 4444 master, to make a DCP in DoM, I get relatively acceptable (but not perfect) results. There is more detail in the blacks than from the 4444 and the picture is overall lighter.antonarap wrote: ↑Thu May 17, 2018 11:49 pm Independently of project color space, in Davinci 14 render to TIFF RGB 16 bits => In DoM encode with transformation set to Rec 1886.
Said transformation, I select it manually. DoM automatic recommendation is Rec 709.
The question is "why is this so"? Is Davinci 14 in fact exporting all TIFF RGBs in Rec 1886 independently of project settings? Or is it DoM doing something strange under the hood?
Both graders have said to me that the issue is not Resolve - it's the way other programs interpret what is being given to them. Both have said that the converter (DoM, in this case) must be told what the source material is (data range, colour depth, gamma etc.) for it to interpret it correctly and that it must not rely on what it 'thinks' the source material is. Obviously, to be able to do this the receiving software must be capable of being told what the source material is. It was even pointed out to me that Resolve can misinterpret the material it is receiving, but that this can be easily overriden within Resolve.
I've tried overcoming this now on two films. Both were graded professionally in two separate facilities and both produced UHD 4444 Prores full data range files as source material. I can't get these to produce acceptable DCPs from DoM. I realise the OP is using TIFFs or DPXs as source material but I think the issues being experienced are potentially the same. I'm going to try @antonarap's recommendation for colour transformation set to REC1886 to see if this helps at all. From what @carsten says, it is not an exact transformation.
I've done a few transformations from earlier versions of DoM using 2K 4444 full data range as source material (using earlier versions of Resolve to render out) and DoM interpreted those correctly, producing perfect results. It's only recently with a higher resolution UHD 4444 full data range source files that these issues have arisen. This could mean something has changed in DoM, or it could be that the issue only arises with UHD, or it could be Resolve is doing something different in later versions.
I believe that the only definitive way of getting this absolutely right, is to have settings in DoM that the operator can change to match what the source material actually is, if DoM has misinterpreted the source. And so I would like to ask for that as a feature request.
I'm willing to try anything to resolve (sic!) the best way forward, but I am not an editor, nor particularly adequate on Resolve, so I may be a little useless. I can, however, ask others who have skills to help in finding a resolution or more information, but that will be subject to their goodwill and availability. If there is something I might try to cast any further light on this, please let me know.