Thanks for the kind words! It is indeed one of the sharper shots I’ve managed to get, I think. Just a slight movement forward or back can dramatically make a difference between “tack sharp” and “sharp enough”. Profile shots like this allow for more sharpness across the whole body. I’ve sometimes found that in more head-on shots it’s the ones where the eyes are just on the edge of focus that I find best because if you get the eyes totally in focus on those shots, not much of the body is and it can look weird.
However I do think the unmodified RAW file can make the fine details a little blurry. This was the first image I tried the demosaic capture sharpening on and it blew me away how much it improved things compared to just using diffuse or sharpen or contrast equalizer. The structures in the eye specifically really come out much more defined. I’m not sure how much it really matters for print or viewing on the web though.
Indeed, I usually only spend a few minutes on edits. I actually do enjoy editing but I don’t like fiddling around too much, like squeezing blood from a stone. The benefit of my setup at least is that the lighting is consistent (flash with diffuser) so it makes the process of culling and editing consistent.
Your first one reminds me of something like a GameBoy camera. This one reminds me of the Playstation game Vib-Ribbon. I’ve always thought about trying something like this after seeing an exhibit on Charles Gaines’ Numbers and Trees series
Tile mosaics like this can easily be done using GIMP presets (i.e., Mosaic filter under distorts). If I were to go to a higher rez, I could have gotten a better result, but too lazy (took this weekend off as vaca; lol).
I did not spend a lot of time on this, but because I have an older version of Topaz, I thought you might find it interesting as a comparison.
For Darktable, I used capture sharpening, AgX, Contrast Equalizer and Sharpening. That’s a lot more sharpening than I would ever normally use, but I wanted to see how far we could go.
Thanks for sharing the Topaz results! I’ve always wondered how well it works in practice vs the marketing of extreme cases and extreme results. It looks like it does a pretty good job and in this case didn’t make things too smudgy. Though the greens in the fly’s body do look noticeably more smoothed out, which I don’t love.
I’m a bit of a purist, so I don’t use Topaz much at all. What concerns me a little is that I don’t know how much generative fill it is doing versus correcting what was captured. Can we trust that the eye detail is all from this particular fly’s eye? I don’t know, whereas I know that Darktable isn’t adding anything that wasn’t there in the first place.
I don’t have topaz to do definitive tests but I do feel that DT would be able to match the topaz image pretty closely. I am certainly satisfied with the sharpening options in DT if I actually get my focus correct.
EDIT: Just compared my 100% DT edit posted previously to the topaz 100% edit and in my view my DT edit is sharper and has more detail and texture. DT wins for me.
Ya really I was just commenting on this particular pair… I don’t have Topaz either just ON1 photoraw and Affinity photo for commercial options and I don’t use either much. In the past the ON1 had done some amazing jobs and some trash jobs but that could also just be me not having a feel beyond defaults for those ones….
Its just really important to at least exposure match these sorts of comparisons to compensate for any perceptual impact of the contrast and exposure….
Yeah, I wasn’t actually trying to match the Topaz image. I was just trying to get as much detail as I could, especially from the eye area.
I could do a much better job with the Darktable version but I think I’ve already got a feel for what both can achieve. Topaz produces a cleaner image in less time, but some aspects of it are too clean. In other images I’ve tried, it can produce some very waxy looking results. But for nature, it’s always been pretty solid.
This is why I gave up on Lightroom denoising (the tool that converts to a DNG). At first I used it because it must be better than not (i.e. I had no idea what I was doing or why), but eventually I realized how unnatural it was making things look.
Of course now I just use darktable denoising which is pretty subtle but only does what’s on the tin, so to speak.
not doing over-sharpening because the photo is already sharp, and I like the smooth aspect. Also over-sharpening quickly lead to aliasing in the eyes of the fly and also giving an overall “crispy/noisy” aspect I don’t like.
Run this through paid Canon Neural Network Image Processing tool with -1 for de-noising, because default is too strong to my liking and then fine details preset from Canon Digital Photo Professional and got this kind of results.
I know this forum is aimed at FOSS but since there are a lot of us Canon owners here I am wonder about your review of the Canon’s Neural Network Image Processing. I haven’t ventured into it since I am a dedicated DT user, but maybe there are some functions DT can’t do but the Canon software can do.
