It’s a good tipp, but that will not create the clean rendition like ART. Might provide the base picture as play raw, when there is interest.
I had literally no experience so far and used some modules that made intuitively sense to me. But in the meantime I did grasp at least the basic functions and feel more confident in the sharpening/denoising functions. I really think the capture sharpen module is superior to what DT provides (and I do know the sliders of DoS to create custom settings for a picture). Don’t want to fire up a war here, just sharing my experience.
Yes, that’s what initially triggered me. I find it easy to use and it fits my style/workflow.
I have to explore this more… I know @agriggio did rename the sharping to capture sharpening but I dont think its the same as the RT one… but that would be nice I thought this was mentioned some time ago…
I say this because pixel peeping it was always my experience in the opposite ie that I got a cleaner image from the DT output. This is not meant as a rebuttal…only that I should go back and play some more… i know that setting the default radius lower in ART as was suggested to me was helpful a couple of years ago when I tried to improve my use of the sharpening module and stop generating the worms…
It would be nice to help DT and ART users if there could be some comparison edits in the spirit of sharing tips and techniques and not to diminish either editor. They are both continuing to improve over time… Even if nothing else to see what people consider a sharp image and to see how much that varies between people when they serve up the final edit
Some of this can also come from the fact that most users of DT are throwing filmic and or sigmoid on every image… This comes at a price and if you don’t need it for DNR… try editing with only the tone eq and local contrast or even a curve if you do need it… You will notice that you lose quite a bit of detail with because of the compression with filmic and in skin and portraits for sure with sigmoid… just another nuance when comparing…
I support that very much as I think that’s the attitude that will take aus forward.
I concur. By not using filmic/sigmoid, you will also learn, how to use it I mean, it will push the understanding of what tonemapping does and then it’s easier to apply a tone curve (be it with filmic, sigmoid or by hand)
Here is another example. The pictures in my original post where from the JPG export. This here is from the editors (but I compared also the exported JPGs, with your above mentioned icc from color.org).
Actually I can’t believe it. It seems like the sharpness is from another level!?
The difficulty with maybe trying to compare these two is on one hand with ART is that it can use DCP files and they have a tone curve and color table… I think in a recent post to demonstrate the range of possible looks I just applied all the profiles for I think a Nikon Z50 +/- the look tables and shared the exports. Then if you don’t use that and you use the camera profiles you might have the normal one or the auto matched camera profile… then further to that in a manner similar to older versions of DT with those filmic color preservation modes that offered different looks you have the different models in the ART tone curve from Neutral , Std, Std weighted , film , perceptual… all which look quite different…
Its a great range of options but might make attempts at a meaningful comparison a bit trickier…
Yes thats right. And I didn’t try to match these two perfectly, but I used in both edits a manual tone curve, a reasonable matching white balance and saturation.
But, I mean, just look at the fine details that ART manages two bring out. DT almost appears pixelated or blocky in comparison. It might just be a bit exaggerated by sloppy use of local contrast in DT.
Its hard to say from screenshot on my end but for example I noted in your close-up of with the eyes…the pupil in the DT export had that sort of black that was not crisp whereas the ART one seemed nice a “black” This is what I found I needed to do in DT was either do a small tweak to the black or moving to relative renders using the color.org profile… Then for me it was night and day and I was much happier with my DT edits … This might all be subjective on my part … There is also the thing in DT where setting the high quality reprocessing does seem to make a difference in your exported jpg in DT… Setting it to no seems to produce an export that more closely maintains the look of the full screen preview where as setting it to yes will export an image that is more faithful to what you see when zoomed in 1:1… so just another possible nuance when trying to compare…
If I look at the screenshot in full size, I find it clearly visible. But yeah, this stuff is pretty subject, I guess. Maybe someone else can judge?!
I sure appreciate your opinion.
Yes, I know this option. But above is just a screenshot from the editor, though.
I downloaded the image, and opened it on a 4K monitor in Geeqie at 1:1 (so without scaling, as the image is ‘only’ 2440 pixels wide). Then I asked one of my children, to whisper to me which one she found sharper. Finally, I asked my wife. All 3 of us thought the the image on the right, so darktable, looked sharper. That could of course be due to contrast, not sharpness.
I would not trust any kind of preview / thumbnail in an editor, BTW. I think exporting at full size and comparing the exported images is the right way to go, if you really want to pixel-peep.
Yes, that’s quite impressive. Obviously I don’t know what setting you used in DT but I don’t like the sharpen demosaicing AA filter preset. I always start with the no AA preset and tweak from there.
Of course, I’m not trying to convince you to go back to dt from ART But I am impressed like you from that comparison. I’ll have to do some experimentation…
I have worked more with RT than ART but they have a lot of the same tools so I think my observations still apply. I was always amazed at how sharp my RT exports were compared to the blurred image in editor. Not sure if that’s a quality setting that can be changed or the downsizing algorithm used, but you certainly can’t compare an images true sharpness at anything other than 100%. I did however find it trickier to properly edit RT’s blurry images.
Capture sharpening was one of the main reasons I preferred RT over DT for a while, but DoS (along with details mask) negates that advantage. I find it more versatile and capable of superior results, and have created my own presets for quick use, however capture sharpening is far simpler and can yield better results out of the box.
Indeed I was refering to the first post, not #7.
Though I hae to say that even viewing the original image at 1:1, to me the left one (so the one from ART) is sharper, or with finer detail than the one on the right.
Cool isn’t it? But yeah, it did take a while for me to get the hang of it. I only touched two sliders in my example though, after starting with the demosaicing no AA preset.
I don’t use/know Darktable, so I had to look up on the web what DoS is. Form what I understand, some other modules in ART could maybe do some parts of what DoS: Dehaze, and Texture boost/Sharpening.
But yeah, it did take a while for me to get the hang of it. I only touched two sliders in my example though, after starting with the demosaicing no AA preset.