I have read 1 or 2 recommendations to use DxO PureRaw 3 ahead of darktable in one’s workflow. I’m obviously not fully understanding things, but I don’t see the overall benefit. Could somebody please clarify it for me?
The DxO product cannot write .RAF or .CR2 files, so I assume that, after use, it must export my images to darktable, no longer as ‘camera manufacturer raw’ files but as .tiff or .dng - or even .jpeg. Surely this then limits the ‘raw file’ processing that darktable can do, to the point where you might just as well do all the raw processing in DxO - but it would have to be in PhotoLab 7. This would avoid the resource cost of running multiple raw processors and of having to store an intermediate file between DxO and dt?
The articles I was reading at the time - about how to use certain functions in dt - extolled the performance of Pure Raw 3 in the very narrow aspect of noise reduction, suggesting that the combination of using PureRaw 3 and dt would give a better result. I just can’t understand how this could be true.
I sometimes use PureRAW and darktable. Excellent NR but lens recognition may be bad, like my Sigma 500/4.5 that is identified as 500mm in the converted DNG file.
I actually find noise reduction in DT pretty good. I sometimes wonder if too many people are obsessed with get images where everything is a smooth as a babies bum. I am currently processing numerous images shot at ISO settings in the thousands and am more than pleased with DT’s ability. With such high ISO I limit sharpening to edge details by using the detail slider that becomes available when you select a drawn mask option (no need to draw the mask though). I use the diffuse or sharpen module for sharpening and denoise (profiled). So for my 5 cents worth, I want to bring in the unadulterated raw file into DT. I make no comment about how good or bad DxO may be but I doubt there would be much benefit using more than one program to process your raw files.
Oh, it does work. Normally you just export a DNG from the other software - Pure, Topaz, whatever. Then open it in darktable. In theory it’s still got all the data, except that it’s already demosaiced and probably white balanced. But there’s quite a few if’s and but’s possible.
Some have found the that the “AI” noise reduction gives better results to their taste.
And still, while I am wholeheartedly for Open Source and Darktable, my experiment with preliminary cooking the DNGs with PureRAW continues. Of course, it hasn’t been long, so we’ll see if it has a firm place in my future workflow
i use pure raw for certain noisey images, but not everything. I export the dng with noise correction and sometimes lens correction. I keep the original raw along the export. Currently shooting on an older micro four thirds camera and love the ai denoise for when I need it, but light noise is handled well in dark table.
In instances where I need more noise reduction, I will first process my raw files through ON1 using its linear raw profile, with lens correction turned off and no other changes to white balance, tone or color and then save the file as TIFF or DNG and import those into darktable to take advance of its superior editing capabilities.
The trade-offs are larger file sizes and loss of darktable’s more advanced demosaic algorithms, but I haven’t seen any limitations concerning white balance, highlight reconstruction or any other module. I can’t explain why, but the exported DNG files retain the ability to use darktable’s details threshold slider, however, the DNG files out of ON1 seem soft, so I prefer the TIFF format.
I don’t know anything about DXO Pure Raw, but they seem to tout its integration within Adobe, which obviously is not the case with darktable, so personally I wouldn’t use it as part of your regular workflow unless you had a clear technical objective in mind for all of your photos.
There are certainly limitations for pre-demosaic modules. highlights, cacorrect … either don’t work at all or offer restricted options/quality. (for tiffs even more so than for demosaiced DNG files)
No big ones except they all rely on “best input” so do as you like but if you have camera “raw” files (whatever the vendor is, some give raw-DNGs) use them!
The first question I’d ask, is why do you want to use this non Open Source software?
