I’m totally desperate. I’ve tried many possibilities, but I simply can’t reconstruct the JPG image of the camera even close to it.
This is the JPG from the camera:
The conventional reconstruction uses base curve, which gives me following result:
The steel framework from the candle is way to thin and washed out. Tweaking the base curve in any direction combined with highlight reconstruction and clipping and so on gave me a better result in the overall image but the framework still is very thin and I get gibbs ringing / dark pixel rims artifacts.
I wonder if the way with the base curve has anything useful at all, so I found a thread here, where the base curve is skipped out.
This gives me better result with the resolution of the high contrast candle wires, but now I cannot reconstruct the overall image brightness.
I desperately need you help, I love DarkTable, but it took me weeks to process a few images and I’m still stuck. I also have years of experience in signal and data processing, so I’m no complete noob
With an image with high contrast you may have to apply modules selectively using masks. Idea is to focus on one region, find best settings for it ignoring rest of the image and apply the modification to the image. Then, mask out the irrelevant parts. Have you tried it?
I don’t think that this quite complicate procedure is done by the automatic reconstruction by the camera. So to my opinion it must be easier to get a bit closer to the jpg, but for the fine tuning, of course one might do parametric or regional masks.
I found a very satisfactory solution for Canon .CR2 files using Adobe .dcp profiles and RawTherapee as described in the following post: Canon camera styles
I’ve no idea if it would work as well for your Sony RAW files but it may be worth a try and you could always export the tiff to darktable for further processing once you have done the initial conversion.
The vary fact that simple procedures done by you for hours don’t give you the same result as the camera is proof enough of a secret sauce. Besides the procedure that I had indicated is not impossible to automate. The camera styles are such attempts done by the manufacturer.
I’m going to go out on a limb here, as I’m finding it hard to actually see the material difference in the candle frame you describe in these screenshots, on my sRGB monitor, but it may be the “washed out” gradation you’re seeing is actually the retained data of the raw file, where the JPEG clipped the high-key parts and the camera processing increased that contrast, locally. I struggled a bit with such when I started to use raw files without the so-called benefit of the camera processing, but came to realize I’d been given a gift that I had to learn to manage.
For your third screenshot, you do have an opportunity to increase the overall brightness of the image without losing the highlights, in the simple application of a curve with a small ‘lift’ in the center, or lower, or higher, depending on your taste. What won’t happen with this operation is to push the highlights past white, as the curve gradually tapers to the upper-right.
From my experience, once you learn the ways of the curve, you’ll probably abandon all consideration of camera ‘look’ profiles and such, and just make your images look the way you want them to…
Edit: I struck out ‘locally’, above, as I don’t think camera processing is that sophisticated, looking for specific patterns in the image and ‘gronking’ them around. That’s a techincal term…
Not sure it will help. Do not shot RAW+jpeg or do not look to the jpeg. The camera embedded algorithm is not darktable and inevitably you’ll get two different rendering. Open your RAW in Lr or any other software and you’ll have more different results. Which one is giving better result? Not sure. If you want to stick to the “reality” you’ll need to profile your camera and this will give you a style to apply (tone curve + color checker).
The camera JPG has a proprietary look to differentiate the manufacturer from its competitors. While it isn’t impossible to emulate the exact appearance, the real question is, as a photographer / artist, would you want someone else to dictate how your images turn out? Personally, I would rather go with my own style and tastes.
That said, RawTherapee does have an auto-matched curve, which makes adjustments to match the embedded preview, which I would say gives you an excellent starting point to your workflow.
Great idea. This is not a great shot, but I have many night shots containing contrast like this and just want to make them as smooth as possible. DSC09850.ARW (23.9 MB)
Why did you published an edit from DxO in a thread that’s supposed to help the author do a better edit in Darktable or, eventually, another free, open source software?
When joining this forum ( a few days ago ) i did it because there was an “noise reduction” thread , and I wanted to show what DxO is capable of. RT has improved a lot with Noise reduction , but I feel there still is room for improvement.
Outside this forum like in Dpreview/forum there is a wide acceptance that DxO PRIME NR is the industry standard when it come to nr.
Have been following this forum for a long time ( years )
There have been many thread about shadow/ highlight recovery.
For a long time darktable/RT has been light years behind Lightroom ( which has been the industry standard).
Some users have posted editing done with Lightroom/ACR - just to show what can be done.
Maybe this has pushed RT to drastically improve shadow/highlight recovery ( which I feel they have done).
If it is against the forum rules to post images edited with other RAW editor than free/open source , I will stop doing that.
Have been using Lightroom for years. After going subscription I moved to DxO ( number two reason for that is PRIME )
Even though i like RT 5.5 a lot - DxO is still my primary RAW editor. Thats why I have username DxO-user
I didn’t see any explicit rule in that direction, and sometimes posting edits from commercial software adds value to the discussion, like the noise reduction thread you’ve mentioned.
But it doesn’t make sense to me posting DxO edits all the time, when there is no particular reason.
But I’m relatively new here at the forum, and maybe I’m being too radical.
Honestly, the impression I get is that you’re trying to prove how better DxO is from other free software. But, again, I’m probably wrong, and if I am, I apologize in advance.
There is no explicit rule and you’ve hit the nail on the head, it is surely welcome where it adds value. Where it adds value is extremely subjective @DxO-user seems to be posting in good faith and had said s/he is trying to learn RawTherapee.
I don’t get that impression at all.
We mostly shy away from posting proprietary software renderings of our images because (1) not everyone has access to proprietary software and (2) we are explicitly for free software. In addition, there are a multitude of other places to share proprietary edited images. We stared this forum specifically because of the lack of places for people to congregate around free software for photographers.
