Really nice improvements, thanks Alberto!
I have the impression that the tone equalizer black slider has much less authority than the white slider.
I am usually led to use shadow slider instead to pull the black.
Is it normal?
The action of the sliders depends on the exposure of your picture (and its dynamic range). It can be perfectly normal that “whites” has a more visible effect than “blacks”. Anyway, tone equaliser comes after exposure compensation, so if you apply negative exposure compensation then the blacks slider will have a more visible effect.
1 Like
As your tone equalizer is based on an early version of @anon41087856 's tone equalizer for darktable, why not just get some of the updates? He implemented a mask that will essentially spread the tones/sliders over the histogram so that all sliders will have an effect and not only a subset.
honestly, I think it works fine the way it is
How does everyone set their black point? Does everyone put on the black clipping indicator and then adjust blacks until clipping starts to show?
The reason I ask is because I used to use the histogram a lot in other software, but I find the histogram is not so useful in ART/RT. I can’t really tell from the main histogram where pixel values start and end, and there is only one histogram shown on the auto-matched tone curve. Is there a reason why the histogram isn’t shown on the second tone curve?
This assumes you have ‘real black’ in the photo. This might not be the case every time…
Personally, I care about it only if the parts that should be black are not 
1 Like
I agree, this version much simpler and quicker to use and seems to work pretty much as well IMHO
Missing libraries.
I intended to use the compiled ART program obtained on my main machine (ubuntu 18.04) , on a new one under ubuntu 18.04 also. I had to add to libraries to be able to launch it :
libcanberra-gtk3 and libexiv2-14.,I thought that all necessary libraries were include during the compilation process. Where is my errors ?
Thanks for your suggestions, I learn at each step.
Yes, you’re absolutely right. I was talking about photos where there’s black and you want to maximize dynamic range.
My workflow usually involves:
- open the photo
- decide if the auto-matched curve fits or I want to use some Film Simulation (I have a Fuji and use the CLUTS’sfrom internet)
- fix the exposure if needed (i usually try to make it right in camera, but is not always possible)
- use the Tone Equalizer for more exposure tuning
- apply other more or less ‘artistic’ modules to tune sharpness, noise, contrast, color, etc
- crop and save
All the while I am looking at the histogram to check that I do not unintentionally clip highlights, shadows or any of the colors.
I don’t think ART (nor RT) provides any general libraries. The presence of those libraries is checked at configuration time, before compiling.
Yes but RT on the same machine did not ask for those libraries. That is not a serious problem, now I know that I will have to install those two.
RT doesn’t use libexiv2, only ART does. But for libcanberra, I see it’s also required in RT’s CMakeLists.txt, so I don’t know why it doesn’t complain on the second system, unless libcanberra is present but ART was compiled with a different version than that present on the second system?
Last updated french translation. New 1356 line and minor typographical errors.Francais.txt (147.2 KB)
curious behaviour
Using photo from Sony ARW looks yellow in RT and DSC04493.ARW.arp (37.0 KB)
I get curious results.
snapshot 2: neutral ok
snapshot 3 ; applying local contrast with all sliders to 0. should be NOOP. But it blurrs image! ko
snapshot 4 : applying sharpening usharp mask, blank regions appear 
ko
Error message during compilation process of ART 0.3-15 :
/home/serge/programs/code-art/rtengine/ipshadowshighlights.cc: In lambda function:
/home/serge/programs/code-art/rtengine/ipshadowshighlights.cc:67:37: warning: ‘*’ in boolean context, suggest ‘&&’ instead [-Wint-in-bool-context]
const float gamma = 3.f * (hl == (amount >= 0)) ? base : 1.f / base;
Do not prevent from a success of the compilation.
thanks, that’s a rather silly mistake of mine – sorry about that . Should be fixed now
Fixed,
but this one now :
In function ‘_ZN8rtengine14RawImageSource10cfa_linednEfbbRKNS0_24CFALineDenoiseRowBlenderE._omp_fn.0’:
cc1plus: warning: assuming signed overflow does not occur when assuming that (X + c) < X is always false [-Wstrict-overflow]
New updated french translation file. It takes into account the new lines added for de mask.
Serge MoreauFrancais.txt (147.7 KB)