Hello,
i post a sunset of a wheat field in Nothern Italy. I use Raw Therapee 5.7 and i would like to better recover highlights, but when i choose highlight compression the sun turns magenta. Suggestions welcome. Taken with a Canon Eos 2000d and 18-55 IS II.
IMG_5749.CR2 (26.3 MB)
This file is licensed Creative Commons, By-Attribution, Share-Alike.
Hey
In the “Raw” tab > “Raw White Points” tool, set “White-point correction” to 1.23.
To improve support for your camera, send us white frames with LENR on and off (sets 1 and 2):
http://rawpedia.rawtherapee.com/Adding_Support_for_New_Raw_Formats#Needed_Photo_Sets
Thank you everyone, so far i must admit my favourite one is the one from @Morgan_Hardwood .
@Haribo nice one too but imho there are too halos and a bit too bright
@yteaot nice too but a bit green around the sun
As for the camera, unfortunatly it isn’t mine and i’ll try to contact the owner for the sample shots.
really nice and natural looking!
@paulmiller very nice editing! I’m studying your pfi to learn phf a bit more.
It’s good to recall something that I had forgotten about the color adjustments module: that it needs changing the color space.
I found this thread that explains it:
Then I switched off your last layer (color conversion to sRGB), and followed the instructions from that thread, that is, before color adjustments, switched to ACES, then back to REC2020 after the color adjustments. The results seem to be the same as yours, and the blown region around the sun is tamed as well.
Thank you.
Thanks @gadolf, looking through the settings people have used is one of my favourite parts of Play Raws (apart from the pictures of course!). Its great to hear that someone is learning from my efforts.
Good point. If you want the numbers to be compatible with an ACES workflow, then you need to have your data in linear ACEScg for the Colour Adjustment. You can do this by changing your working space to ACEScg in the Raw Developer module (Output tab) if you don’t want to have multiple colour space conversions all over the place. Having said that, ACEScg and Rec2020 aren’t very different and the results of applying colour adjustments in either are quite similar (to my eyes at least) - the numbers may be a bit different but you’ll end up with the same result in the end.
One thing that may be making a difference in the blown-out region near the sun – do you have “Clip Overflow Values” turned on in the cc-to-ACEScg and ACEScg-to-cc layers?
The conversion to sRGB at the top of the stack is because I find it useful to adjust the gamut mapping parameters when converting to the output space (the colour conversion in the image export doesn’t have this feature and you couldn’t preview the results if it did).
One subtle point on the sharpening layer at the bottom of the stack:
I’m trying to reproduce the capture sharpening feature recently added to RawTherapee, and it works very well indeed. However, Photoflow’s RL sharpening can’t handle negative values (you get black squares on the output) and it clips values > 1. I put exposure correction AFTER the sharpening to get around the clipping (on some files you may want to put -2 stops exposure before sharpening and 2 stops after to avoid clipping). If I move the Basic Adjustments layer to before the shaprening, then all the detail in the sky gets lost. Putting a defringe layer before sharpening helps if you have lots of specular highlights in your image as otherwise they get nasty halos. If you use defringe, you’ll need a clip values layer to deal with any negative values that the defringe may introduce which otherwise upset the sharpening.
That’s absolutely right. My personal view on this is that the actual colorspace only matters if you want to exchange presets with other people. Otherwise, just stick to your preferred colorspace in order to get consistent results between edits… ACEScg and Rec2020 are basically equivalent in that respect. Don’t use ACES, it is too wide.
That will come at some point, but at the moment it is not at the top of the priority list…
Porting RT’s capture sharpening into PhF is instead very high in my priority list. I still have to decide if to wait until the algorithm is included in Librtprocess, or just go ahead beforehand.
I have been able to use again the 2000d, i opened #5495 with the files. Canon EOS 2000D white frames · Issue #5495 · Beep6581/RawTherapee · GitHub
Thank you.
Quick google shows that the 2000D is analogous to the 1500D, so I loaded Adobe’s dcp profile Canon EOS 1500D Camera Standard v2.dcp into Rawtherapee. Seems to fix the colors somewhat.
Same processing but with the default “camera standard” color corrections which is the color matrix embedded in the raw file:
