It seems that the histogram / colour picker / exposure indicator system has been a matter for debate for some time with the developers and probably has developed some bugs. Thanks, I saw that but be aware the settings can give very different results. ![]() Since, in the export module, one needs to indicate a profile.Ĭlipping preview set to full gamut also indicates clipping in saturation. > usually have the histogram set to use "export profile" so that it reflects where I'm using the image, does this sound correct? I get very similar results to previously. I have been checking some images against older export versions. I'm not sure what the 11 indicates apart from more confusion. Also the colour picker for the jpg shows RGB(%) 0,0,0, pixel RGB 0,0,0. The histogram displays as I would expect with a more or less flat line across the middle of a log scale display. I've just exported the result of the curve bottom right of the screenshot as a jpg and a 16bit tiff and opened them in gimp 2.10.22. I usually have the histogram set to use "export profile" so that it reflects where I'm using the image, does this sound correct? Maybe it could be linked to the output profile to auto adjust it rather than a long list of possible values in a tool tip?Įither way I'm not confidant of the clipping indicator and histogram at the moment, though it may be a set up issue, or it may be that I've been making incorrect assumptions for a while. Admittedly a pretty narrow opinion of correct exposure but there you go.įor this reason the new clipping indicator setting where the lower threshold is set in EV's and the upper in percent seems unhelpful or at the very least over complicated. ![]() Black point should be at 0, white at 255 (or within 5% of). This makes it difficult to judge the spread of the histogram for Alamy's QC which specify "Poor Exposure" to be "Flat or washed out. Also the colour picker shows no value lower than 12,12,12 in RGB and 0.00, 0.04, 0.03 in Lab, which I find odd because surely it should be 0,0,0 RGB? I forget what I tried then but I also noticed that although I could use a tone curve to send a large section to black in the greyscale and it would show a clipping indication for shadows under the "luminance only" method, it shows no clipping in the legacy "any RGB channel" mode which is supposed to work like the previous version of dt. Yes, I think, but the surprise was that the aRGB histogram slid off to the right and wouldn't come back to the left edge no matter what I did. That does not surprise me, since the clipping preview mode is depending on the gamut selected, right? When I make one preset of clipping indicator settings, I see different blue/red areas when I change the histogram profile too. You change the histogram profile on each of your four pictures. I've uploaded this to illustrate, a picture should be worth a thousand words after all. ![]() Did something change in the filmic module? I did make some changes to the clipping indicators (a major improvement in 3.4.0, I think), but the issue remains. I also noticed on some images (not all) that I was unable to get a proper exposure using filmic: the image stayed way too dark, and the histogram indicated the same. I haven't filed a bug report as I wanted some confirmation I'm not doing something dumb first and I'm not on github or any of their other channels and have enough logins for now thank-you. ![]() But when I switch the histogram profile to sRGB(web safe) the graph stays on the left side but the min values are 4,4,4. When I switch the histogram profile to linear Rec2020 RGB the graph moves up to the left side and the min values are 0,0,0. Further darkening the image in various modules stacks the darks to the left of the histogram chart, but it never meets the left edge of the graph. I then noticed that the colour picker showed the darkest point at 12,12,12. I have the histogram profile set to Adobe RGB (compatible) and try to get the histogram well stretched to the black and white points to satisfy the requirements for submitting images to Alamy. I first noticed that when using filmic I could not get the histogram to meet the left edge of the graph. I'm running darktable 3.4.0 from on Ubuntu 20.04.1 LTS and seem to be getting unexpected behaviour from the histogram and clipping indicators.
0 Comments
Leave a Reply. |