![]() Otoh, for v4 I remain unconvinced, for my photography. That overlaps the function of basecurve.c, so fair competition. I eventually started using it because its parametric nature made it easier to edit a batch of photos. ![]() See the second photo, which was exposed for the and I had this discussion with filmic V3. None of that yellow is really there, it's result of the blue blowing out and leaving the red & green channels. Sometimes it just doesn't matter: if you have a lot of blown-out highlight, you are in the realm of creating a psychological reconstruction: see photo below that I took yesterday. and for someone who works mostly in B&W, they matter much less. and in colour, they may well take a cast because of limiting occuring in a single channel. That's an oversimplification, of course: quite often there are small specular highlights. If your photo doesn't have blown highlights, there's nothing to reconstruct. PS: has anyone verified that process_cl does the same thing? open_cl is still a mystery to me. well, I'm not really a button-tamer (dresseur de boutons ?), but if I can be useful in the next month or so, I shall do my bit. Now that reminds me, I can remember from Lightroom that there was one slider for saturation and one for hue. Or that is so far as I understand, which in the case of HSL is not very much. Hence my proposal would be to drop the CLIP commands entirely.įor the other colour (or BW) modules, the same is true à fortiori: the only edits affected would be those that pushed against clipping. ![]() so that the edit actually stopped doing anything. It would mean someone pushed red so that it alone clipped. Although, "that being said", it would be a peculiar edit that depends on this. because I wanted to see if I could make it work better in the space of a few days.įor the clipping, I think we can agree to disagree, and you're the boss here. It's just a coincidence that I fell into this conversation as I was in the middle of editing channelmix. I'm still a beginner around here, so doing things off line is part of my way of learning. I can work around this by switching off filmic and making sure that the histogram is not clipped before engaging the filmic and channel mixer modules but I find it much harder to recover shadow detail when working like Interesting questions :-) I know there are alternative ways to get a monochrome image but channel mixer really is the most powerful. It appears that the module is instead clipping the values and then passing those clipped values to the filmic module. This (loss of highlight detail) is something I've noticed mosly when using channel mixer to make my image monochrome but the above is the minimal way to reproduce the issue and it seems to me that without any parameter changes, the channel mixer module really should just pass the RGB signal unadjusted. ![]() I can persuade it back to white again using the 'white relative exposure' slider in filmic but I've definitely lost some highlight detail: I assume by the peak on the right of the histogram, that I have lost some detail here. I switch on the channel mixer module without changing any parameters and it appears that my 'white' is no longer at the far right of the histogram and is bunched up into a 'light grey' peak. Using the default scene-referred workflow I adjust exposure to get the middle-grey values looking reasonable while keeping the highlights well exposed (note the histogram), with filmic auto-enabled.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |