Nothing’s showing up in the log. The error is just “The operation couldn’t be completed. (OpenEmuShaders.SourceParserError error 6.)”, which isn’t very helpful.
I can’t test it myself, so it’s good to stick to the last working version for now.
It’s probably not a parse/compile error but something in the frontend driver. If you could report it via github maybe it could get resolved.
That’s probably a no-go; the last release of OpenEmu was in January 2021 and I can’t build the current code myself as it requires Ventura (macOS 13) to build despite running on Mojave (10.14) or later.
I see. But 2023-07-17-r1 is fortunatelly quite new. You could also copy/replace the ntsc and bloom slang shaders from the newest version into it. I think it should do and you’ll get most of the updates.
And if then newest ntsc or bloom shaders are causing the error then it’s a good trace of the issue.
Hey @guest.r would it be possible to separate the “scanline saturation/ mask fall off” from each other? I would like to just use the mask falloff feature by itself, or is that a setting that only works with the two of them together by design?
It’s basically the same thing (and same code), but you see it like scanline saturation with dissolved masks and as mask falloff with strong stripe-based masks situations.
Ah ok. I thought they were somehow two different type of settings that you “merged” together or something. From what I noticed at least with my current preset samples and my crazy tweaking is that lowering that setting seems to make colors look a little more natural bringing it a bit closer to how things look when not using any shader at all.
Still testing things out in any case, I feel about 90 percent satisfied with how my current slot mask presets look, maybe one more tweak of something and they should be good to go.
It seems to be the newer ntsc shaders causing the issue. The newer crt-guest-advanced-ntsc-pass2 also causes an error, but it’s “error 1” rather than “error 6.”
It’s best to do whatever works for you. Fortunately, the last working version is quite new. Blind fixing the shaders is somewhat hard, but even more inappropriate if the error might be somewhere else or unavoidable, like related to a specific build.
Bright, saturated but also great contrast, I very like it. Have you tried the same options with NTSC, which is harder to mess with?
Ntsc can also look very similar, but there are some tricks…at least horizontal bloom range should be bigger. Nevertheless, I’m posting some updates soon.
It looks so sharp, sweeeet!
This looks awesome! do you think that is possible to do something similar but @1080?
This is where I’m at with
Sure, it shouldn’t be this different with 1080p and 1440p. Maybe you can use mask 9 with 1080p.
Bit off topic, but how do you achieve this:
I mean, look at the " x 0 0" characters, in the upper image, you manage to shrink its shape. Usually, using blurs, you mix the near pixel along x axis and the result is wider.
I even tried to achieve the same effect by playing with shader parameters (not ntsc version), but without much success; btw I was more interested in the implementation and, though i think it looks really good, if there are any other reasons you pursued that look; i mean, bright pixels tends to grow, not shrink themselves.
Also, the red on the left definitely overflows over the gray, so maybe it is some ntsc specific code that comes into play?
Thanks in advance!
NTSC sharpening uses some gamma tricks to do things differently compared with linear gamma. The overall ntsc interpolation is done in curved space, although post-filtering is linear. But it seems the effect can thin instead of thick regarding bright pixels.
Edit: there is also some additional blur in curved space, so i think this is important.
Speaking of which, is it possible to fully detach the NTSC passes from the GDV-NTSC preset for use by themselves in a preset or appended to other shaders? The NTSC passes used there are already quite a bit more advanced than the standalone NTSC-Adaptive shader, after all. I tried reducing the number of passes up to the 8th pass, which works, but there’s still the afterglow passes before it, and removing those causes the preset to not load.
You can use the shaders in other situations if you set the aliases right. There’s a bit of interleaving between passes.
First ntsc pass should have the alias name “NPass1”, the pass before ntsc shaders “PrePass0”, unless you remove the blend mode functionality in ntsc-pass 3.