I already see it and try, IIRC I cant get Rainbow & Dot Crawl effect, and its so slow at least with 4x upscale of ps1 game mixed 240p with 480i using gtx 1650
anyway, from first look, Complementary Gamma Correct
seems work perfectly, thanks!
I already see it and try, IIRC I cant get Rainbow & Dot Crawl effect, and its so slow at least with 4x upscale of ps1 game mixed 240p with 480i using gtx 1650
anyway, from first look, Complementary Gamma Correct
seems work perfectly, thanks!
Some possible speedup can be achieved if you lower the ntsc-pass1 scaling to 2x. My speedup with 4x internal resolution is like 2.5x.
Man this is another level, the new complementary gamma correct is amazing!!! Thank you!, this is fantastic!..
Definitely I will update my settings with this, I’m really impressed…
Here is another beautiful example, this game helps me to identify the errors, but it’s all so aligned and everything pops out.
Testing some new presets. Primarily tested with 1440p.
guest-advanced parameters:
m_glow = "1.000000"
m_glow_cutoff = "0.080000"
m_glow_low = "2.999998"
m_glow_high = "0.700003"
m_glow_mask = "0.700000"
FINE_GLOW = "4.000000"
glow = "0.200000"
mask_bloom = "-0.700000"
gamma_c2 = "1.200000"
brightboost = "1.200000"
brightboost1 = "1.000000"
clips = "0.350000"
gsl = "2.000000"
scanline1 = "-0.500000"
scanline2 = "14.000000"
beam_min = "1.700000"
tds = "1.000000"
beam_size = "0.800000"
scans = "1.300000"
scan_falloff = "0.400000"
spike = "0.900000"
h_sharp = "3.600001"
s_sharp = "1.200000"
shadowMask = "6.000000"
maskstr = "0.850000"
maskboost = "1.250000"
mask_zoom = "1.000000"
mzoom_sh = "0.300000"
mask_gamma = "1.700001"
slotmask = "0.850000"
slotmask1 = "1.000000"
smask_mit = "1.000000"
Can we merge these latest updates into MegaBezel? I’d like to try some of the new features.
Does anyone have or know any good “accuracy” focused (ntsc) presets for 1440p SDR? Preferably contained in guest-advanced so I can bring it to reshade for use in dolphin and what-not. I’ve been trying to use reference photos and the like but I’m not 100% satisfied with what I’ve been able to do.
If you want more pronounced artifacting and color bleed, then you can lower ntsc resolution scaling. Supposingly that a game triggers interlacing, then you are best off with interlacing modes 4 or 5 unless you decide to merge fields. 2-phase and interlacing have conflict with temporal effects since both use the “same approach”, resulting in a strange sync situation.
Meanwhile i’m doing some magic glow and sampling improvements. I’m quite happy with the SF2 results, since it’s a special picky game.
Is there any recent port in reshade?
The recent port is quite solid, missing only improved rainbowing, edge color preservation and some minor tweaks with ntsc shaders. Otherwise other changes to mainstream shaders are not yet incorporated, but the main composition is in a good shape already. @DevilSingh has been doing the Reshade ports, you can check his thread here:
With most mask mitigating techniques and guest-advanced shaders they work the way brightest colors get the most mask mitigation, which can be an ok scenario, but maybe not with all games.
Some games have the majority of their graphics drawn in mid-colors (regarding brightness) and the feel of overly pronounced masks in quite notable if a preset looks best with brighter graphics.
A solution to these situations could be 2-point (0, 1) to 3-point (0, 0.5, 1) mask strength parameters, but this would obviously break existing presets and be a bit confusing to use.
Currently i’m tweaking something in the ‘lab’, which looks like a quick and simple fix for this.
original:
with fix:
It’s ofc. purely optional fast code, nothing to worry about in general.
I like that you actually used the word “lab” haha. This sounds and looks interesting, I’ll be keeping my eyes on this development.
Wouldn’t that break the output gamma uniformity for games that use full palette?
I belive no, since it’s only mask strength application. Now it can be configured to set mask strength for very dark colors and very bright colors, the distribution is then interpolated between these two settings for other color brightnesses. 3-point interpolation would only add more control (like from 0 to 0.5 and then from 0.5 to 1). It would definitivelly break existing presets and confuse more users.
New Release Version (2025-02-09-r1):
Notable changes:
Download link:
https://mega.nz/file/MxJjDASZ#48DNWpHhO4k-OqkeAskgZ4Pz1pyYPXKHI3-loDAbFhk
Wow I can’t believe this update has been here all these hours and I’m just now seeing this, shame on me lol.
Really curious on what this is about. I’ll check this out and report back after playing around a bit.
Also:
What exactly does it mean by raising the taps color restriction from “8 to 12”?
The Preserve Edge Colors effect still works out-of-the-box for like 14 TAPS in 2-phase. So if the #TAPS is set greater, it will get limited to 14. The limitation only applies to 2-phase. You can choose lower #TAPS numbers and they will apply accordingly.
Example: #TAPS is set to 16. There will be 14 TAPS with 2-phase and Preserve Edge Colors enabled for 2-phase, 3-phase will operate wit 16 TAPS regardless.
Possibly dumb question time - I often like to use square pixels, sometimes even with uncropped image (for example SNES is I think 1:1 PAR/256:239 DAR) even though it’s not entirely representative of how it would have looked. Are shaders flexible enough to accommodate this, without breaking, or do I need to be more rigid in my selection?
As long as you don’t use doubled input resolutions you should be fine with basically all crt shaders. Guest advanced shaders (like most other crt shaders) can handle a variety of input resolutions and viewport sizes without even needing to re-tweak presets used (i.e. sonkun’s, Hari-82’s, Cyber’s…presets).
Guest advanced shaders can also handle multiplied input resolutions, especially the HD version.