Repository navigation
Do we need slide mode? #1167
Replies: 11 comments 1 reply
|
CC @thetalkingdrum since you were just working on some slide bug fixes. |
|
Lol. Yeah, I don't even shoot slides. :) The reason I started working on it was that there seemed to be so many bugs and things that just weren't working quite right. I figured if we're going to have it, it might as well work. It has always felt like more of an afterthought compared to the negative pipeline. I don't have any strong opinion on this. But people seem to use it at least... |
|
I was working on it for similar reasons. I do shoot slides, but I don't use NegPy to edit them generally. It just annoyed me that the feature was there and didn't work well. But every improvement/fix to it feels like swimming upstream. There are a lot of assumptions baked pretty deep into the codebase about negative conversion, and as we add more features I think it might be smart to look for places to also trim and simplify. That said, if a lot of people are getting value from it, maybe we should keep it. It would just be nice if someone who actively uses it wanted to take the lead on maintaining it... |
|
Agreed on the swimming upstream part. Feels like every time you try to fix something, something else breaks. I would actually like to shoot a bit of slides tbh, its just so expensive and unavailable these days:) |
|
I don't shoot slides and never use that part of the pipeline. Dropping it completely would simplify the pipeline A LOT |
|
overall as we are disussing on discord, way fwd should be refactoring the pipeline to have clear decoupling of negative and positive pipelines in backend, that way it should be much easier to maintain, I'll start with the refactor |
|
I guess the main problem that NegPy still can't fix color problems in Slides. It can't create complex correction curves. For my opinion, currently the best tool for slides is RetroShine https://github.com/yellowdolphin/retroshine But it is still missing few important features like custom detection area and 1D LUT export, because developer was on Summer vacation. So yes, current Slides mode in NegPy is sort of useless. |
|
I am no longer shooting slides, but I have quite a few old slides that I am slowly processing. NegPy's slide mode has been really useful there. Even with just a single image it better recovers the look of a slide w/o to much manual tweaking. The easy use of multiple images for an HDR merge make that even better. |
|
Here are some faded test film scans examples https://brianpritchard.com/wp/british-motion-picture-laboratories/lad-and-lady-tests/ They are sRGB JPEGs but even so they still could be color corrected. And this is how color cast correction looks in NegPy. Cast Removal slider just can't handle complex color casts. |




Uh oh!
There was an error while loading. Please reload this page.
Perhaps a sacrilegious question (especially since I have contributed to the feature), but... does slide processing even belong in NegPy? If we're trying to mimic traditional darkroom printing, it kind of doesn't fit (yes, I know about ilfochrome/cibachrome). And even setting that aside, the process for manipulating a slide scan is drastically different than manipulating a negative. Slide mode keeps breaking as we improve things for negatives, and I just don't know if it's worth either the development overhead or the user-facing complexity. What if we just said "NegPy is for negatives" (as the name implies), and let people process slides in other apps?
All reactions