Thanks Tobias, I don't have access to Lightroom (other than downloading a demo version and try to run it in Virtualbox).
TL;DR: I think exposure compensation (Exposure2012) should be left as it is in the LR-compliant xmp files, because it is the only parameter (maybe color temperature is another candidate) that has a physical meaning that can be interpreted the same way by all raw converters. Scaling might indeed be needed to get similar results between LR and DT, but shall not be done with the exposure parameter. The long and verbose version: I did a small experiment (worth 2 cents, but still useful to convince myself). I took 3 pictures 2EV apart: 200ISO / f8 / 1/10s, 200ISO / f8 / 1/40s and 200ISO / f8 / 1/160s, of the same scene lighted with a constant light, then for all 3 softwares: - imported them in Darktable, Aftershotpro and Rawtherapee (the 3 raw converters I have access to and use). - deactivated all parameter (auto blabla, local contrast, highlight recovery, etc.) which could modify the luminosity of pixels - used only the exposure compensation with 0EV, +2EV and +4EV respectively - and finally checked visually (which is subjective and prone to errors) and quantitatively by measuring the RGB values with the sampling tool. With the 3 softwares, the three images are identical in terms of exposure/luminosity/RGB channels . I'm not talking about the noise, obviously. This is logical to me (and to you also I guess :-) ) and means that the 3 software define the exposure the same way. This was my point in previous messages when I said that it is an 'absolute' physical value. I can check, but I'm very confident that it is the same for LR. For this reason, the exposure compensation dumped in the xmp files by the ramperpro controller are not optimized for lightroom, they are based on real exposure (in terms of light measurement, ISO/aperture/time settings of the camera, etc.)**. When I asked the developer xmp files for Aftershot, I sent an xmp file and he coded the software so that a file is written with xmp fields readable for the software. If I ask for Rawtherapee now, the controlled will generate basic pp3 files with a parameter called "Compensation" and so on. Let's do another virtual experiment: somebody takes 1 image and use LR on it. For some reason he/she wants 2 versions of the same image, one with a normal uncorrected exposure (+0EV) and one with a different exposure (say, +3EV). When he/she imports them in DT with the corresponding xmp files, the first one is +0EV and the second one is 2.7EV (more or less) because of the tweaking done in DT. The //relative// result will be different in DT, with the second image relatively less bright than in LR (by 0.3EV). When the relative value between 2 images is important (as in my case), it doesn't work. So I fully agree with you when you say that a scaling is necessary to match DT and LR results. But I think it shouldn't be done with the "exposure compensation" parameter because it has consequences. Or, as Tobias seems to suggest, maybe with an additional instance of exposure compensation. As you may have guessed, it is not important for me if the images once imported in DT don't look like they do in LR. Not because I don't use LR, but because there are so many other elements that won't be imported in the process (local corrections being the main one) that I don't see how one could expect to import images in DT and not have to work on them again. In my case, I don't use the LR xmp files with the objective to match the images in LR and DT, but only to recover the Exposure2012 value. If I was clever, I would code a small script to convert the LR or Aftershot xmp files in DT xmp files. Or use another txt file that the controller also generates and which lists the exposure compensation for each image. But each time I have a look in DT xmp files with that idea in mind, I remember that they are unreadable (at least to me), and that I'm not smart enough to do that. ** example: during the timelapse sequence if my camera takes 100 images that cover a change of ambient light of 1EV, first one being 1/50s and last one 1/100s, the only exposure times that my Nikon camera has mechanically available are 1/50s, 1/60s, 1/80s and 1/100s. When I build the film/timelapse, there will be a flickering effect for every change of exposure time, small but noticeable. What the controller does is to generate 100 xmp files, with 100 different exposure compensation values, and they are calculated based on the exposure time of the camera. Sorry for the long message, I really appreciate having the opportunity to discuss that topic directly. Cheers denis Le 11/03/2016 13:47, Tobias Ellinghaus a écrit : > Am Donnerstag, 10. März 2016, 22:10:32 schrieb Denis Testemale: >> The only setting present in the xmp files is the exposure compensation, >> nothing else. And for that matter, you can also get xmp files for Aftershot >> program (that is something that I asked to the developper and he did it). >> The only differences between those 2 xmp files are the xmp fields (so that >> they can be read by the target softwares), the exposure compensation value >> is the same. Which makes sense to me since exposure is an absolute value, >> not dependent on the program. It even has a unit (eV) which itself is not >> dependent on the program. So when I say that the xmp files are "for >> Lightroom" it only relates to the xmp fields. > > [...] > >> xmlns:crs="http://ns.adobe.com/camera-raw-settings/1.0/"; >> crs:Exposure2012="0.29048"> > > As the namespace of that value indicates, it's an Adobe Camera Raw setting, > which is most likely using the same exposure math as Lightroom. Therefore I > assume that the value put there is meant to be read by those programs. And > iff > these need some scaling to be getting identical results in darktable then > that > should also be applied here. Of course, there are several assumptions that I > can't verify as it would need access to Lightroom on the same computer (so > the > screen, display profile, ... doesn't affect the results) and some time and > determination: > > - Lightroom settings really need the scaling. From what I remember having > read > in the past in comparisons between dt and Lr was that their default exposure > differs. So it might very well be some extra exposure that we should add and > then using the Exposure2012 verbatim ontop in a second instance. > > - The people who created the timelapse thing that creates your XMP files > actually tweaked the settings to give correct results in Lightroom and not > just dumped some numbers into that field. In that case we would have to take > the numbers verbatim, no matter if Lightroom actually needs scaling. > > [...] > >> The exposure compensation is the same. Once imported, the luminosity of the >> images is smoothed and the timelapse doesn't flicker. And it doesn't matter >> if the image imported in Lightroom, Darktable, or Aftershot looks >> different: that can be corrected differently (levels, curves, basecurve, >> etc.). > > Do you have access to Lightroom to try if the values as given result in a > constant exposure? > >> Hope it's clear. >> Cheers >> denis > > Tobias > > [...] > ____________________________________________________________________________ darktable user mailing list to unsubscribe send a mail to [email protected]
