Hello everybody, I finally decided to join you folks in this amazing forum. Through all these posts, studies, explanations etc. I learned a lot during the past year and first of all I wanted to thank the corona team and everybody else on the forums for their work. Never had that much fun with rendering like I am having with corona (and I am glad that it is still freely accessible for cinema, as I am counting myself to the more hobbyist group of users).
But enough of the flattery. The hard part is that my first question is concerning the good old subject of LWF and color management. Yes, again, another fool trying to understand this whole subject with limited success. I really tried and read a lot, but still some questions remain. I think I understood the basics behind everything, that corona is computing in linear space, but that we need the gamma correction to visualize the colors on screen etc., but there are so many articles and statements about how to use the LWF correctly, reaching back over years, mainly concerning 3ds (which seems to work differently than cinema in setup) that I have to get rid of some confusion resulting from all this mixed up information.
So, what I did was creating two monochrome photoshop images. The first one is a simple square in mid-grey, with a linear RGB value of 0,5 in 32 bit with a 1.0-gamma-profile, saving it as openEXR. From my understanding this should be a perfect mid-grey linear bitmap.
For the second image I converted the first one to sRGB, visually resulting in the same mid-grey square with an RGB value of 188/188/188 and a brightness of 74%, saved as 16 bit in tif-format.
Both were saved with their color profile embedded.
I then created a small test scene in cinema, where I would render a square with these two bitmaps using only the diffuse channel. I used the corona bitmap loader and left the input on „embedded“.
The resulting two images (saved as 32 bit openEXR) were then again opened in photoshop to color pick the resulting pixels. Altogether everything went as expected, the 32bit EXR is identical to the starting image. The second EXR (of course now also 32bit) is slightly off having linear RGB values of 0,5083 respectively sRGB value of 130/130/130 and a brightness of 51%.
Now my questions:
Do these steps describe a correct linear workflow? Since this is my understanding of LWF I just want to make sure …
The slight deviation comes from an interpolation between the 256 from sRGB and the wider spectrum in linear RGB, correct?
Now that both ways appear to have a more or less identical result, do I have to bother with LWF at all? Is this outdated since all sRGB-images are corrected internally to gamma 1.0 by corona?
I guess my problem is (assuming that I understood what LWF means in general) that I don’t know how this effects my images, as the results look the same, at least in this small experiment.
I read about Bertrand Benoit using 180 RGB for his white wall paint in his „classical“ images. In his secret little hideout Dubcat explained this September that Bertrand meant 180 linear RGB with a reflectance (brightness?) of 70.6% because 180 sRGB would have a reflectance of only 46.5% and therefore would be too dark. This I simply don't understand: When I create a square in Photoshop in color space sRGB and fill it with 180 grey, the brightness value is 70.6%, if I do this in linear RGB the values are the same, but of course the color changes. This confused me a lot and brought me to my longterm goal, to finally understand LWF.
Sorry for the long post, I know this has been discussed to the limit, but to be honest I am frustrated about my own confusion and I feel this has to be understood before continuing with things like PBR. Thank you in advance.
