Shaders for various emulators

[Dr.Venom]

Originally Posted by guest.r

I updated the shader, now masks 5 and 6 are added, should have a bit of a more trinitron look (especially mask 6).

This definitely looks promising .

For me mask 5 is the best out of the 2, personally I think mask 6 has unrealistically high spacing between the darker dots when I make a comparison between shader on 15" laptop LCD and my 14" Trinitron.

Mask 5 though comes very close, but has some issues. The gray ramp in the 240p test suite is not very good with mask 5. I need to bump the Gamma and Gamma Tweak both to the maximum to have the darker grey tones to show at all. The white tones with mask 5 are overblown quite a bit. All in all when using mask 5 it's a bit as if contrast and brightness are off with whichever setting you try.

Is there a chance you could create a middle ground between Mask 0 and Mask 5?

Or another great option would be if the dot spacing for the single color dots (the level 2 and 3 dots as per previous post) could be made user configurable, it would be a great way to accommodate for the different types of users' screen sizes, resolutions and preferred simulated dot pitch.

Since Gimbal is asking for it, here's another one for his nostalgia belt

Guest.r 's shader with alternate scanlines (taken with version 1) and mask 0:

[guest.r]

Originally Posted by Retro-Nerd

Any chance that you could add the Sony WEGA mask from their flat CRT TV series? Looks pretty nice too. (as always: a second click to enlarge). Shader file attached.

The Analog shader pack brings me in a good mood everytime i look up the presets. Unfortunately the mask is calculated by stacking four crt shaders with their masks on as it seems. May look good, but hard to reproduce with a single shader.

@Dr. Venom

For me mask 5 is the best out of the 2, personally I think mask 6 has unrealistically high spacing between the darker dots when I make a comparison between shader on 15" laptop LCD and my 14" Trinitron.

Yeah it requires a special taste/preference to use it below a 1440p display. It gets better with larger resolutions i think. Unfortunately it's the only way to reproduce some mask effects. I tried combining a width-2 and width-3 mask in the same setup, but glitchess happen. It's definitely on my TODO list to invent a shader mask which is accurate and dense at the same time...

which bring us to mask no. 5. I made it far brighter, passed my "suite" test it seems without gamma corrections, so it should be ok in general.

Is there a chance you could create a middle ground between Mask 0 and Mask 5?

Mask 5 is related with Lottes maskDark/Bright parameters, so altering those might help a bit.

To sum up the changes, i concentrated to improve mask 5 a bit and brightness should be better now. Alternate scanlines 2 are also tad brighter.

Or another great option would be if the dot spacing for the single color dots (the level 2 and 3 dots as per previous post) could be made user configurable, it would be a great way to accommodate for the different types of users' screen sizes, resolutions and preferred simulated dot pitch.

As i look up to some shaders, i see some efforts to bring the 1080p mask look to 4k, but it's much harder the other way around, unfortunately.
Yeah, it's 2 or 3 dots, smart mask which does something like **-**-**-** for middle values is possible, but far from trivial. I'll keep my eyes/mind open.

[Dr.Venom]

Originally Posted by guest.r

Yeah it requires a special taste/preference to use it below a 1440p display. It gets better with larger resolutions i think. Unfortunately it's the only way to reproduce some mask effects. I tried combining a width-2 and width-3 mask in the same setup, but glitchess happen. It's definitely on my TODO list to invent a shader mask which is accurate and dense at the same time...

Yes it will probably be a very different ballgame with 4K and HDR... I do look forward to owning say a 4K HDR GSync monitor someday and maybe experience a further step up in CRT shader realism, but that'll probably be something for 2020 or so. Until then 1080p will have to do

which bring us to mask no. 5. I made it far brighter, passed my "suite" test it seems without gamma corrections, so it should be ok in general.

Mask 5 is related with Lottes maskDark/Bright parameters, so altering those might help a bit.

To sum up the changes, i concentrated to improve mask 5 a bit and brightness should be better now. Alternate scanlines 2 are also tad brighter.

Thanks, this looks better indeed. I did play around with the mask settings and it helped, but there is something still not yet right.

I think there are two issues remaining (related to each other):

1. Inconsistency of mask appearance: The mask is much stronger / more visible in some parts than others. Because of this there's a discontinuity in the "trinitron" style across lighter and darker single color areas of the screen. I.e. for some adjacent areas it goes from having no or very light mask to a very strong mask appearance. This kind of breaks the authenticity for me when comparing with real trinitron.

2. Dark gets too dark: Because darker single color areas have the mask applied too strongly (black pixels in the mask) these areas appear even darker than they are by themselves already, creating a wrong contrast between dark and light.

One of the more clear cases that illustrates the points above is with the game Super Puzzle Fighter II Turbo on PS1 and specifically the gems. I've made a cut-out and enlargement of a single green gem in the picture below. You can see that the mask is very strong for the two darker areas of the gem, forming a discontinuity compared to the very light mask in the brigher areas.

To make the look more consistent I tried to lower Mask Dark to 0.0, but then the mask is still too strong for these single color darker areas. On real Trinitron this difference is much less apparent. It's only slightly stronger for the darker single color areas (these are very subtle effects when looked at from normal viewing distance).






Hopefully there's a way to recalibrate this Mask Dark strength. I think it would provide a more consistent mask look if the current 0.5 strength for mask dark (i.e. fully blackness between pixels) would only be reserved for mask dark at max setting of 2.0 (max). Everything below that (and thus also default of 0.5) should provide a much lighter application of the mask (i.e. no full black between pixels)

A good test case would probably be to have the four shades of green in the green gems show a gradual and subtle increase in the mask strenght when going from the lightest to darkest parts (when at default setting of 0.5 and 1.5), and for the darker areas the mask doesn't have fully black pixels when Mask Dark is at default of 0.5.

All this while still passing the "test suite" tests of course .

I hope this is among the possibilities, because this new mask holds much promise to be a very good if not better alternative to mask 0!!

[guest.r]

Well, the "cutoff" (which produced black gaps)was hard coded into the shader, i changed that with the new version.
Unfortunately i can't access a trinitron crt, so i have to relly on user feedback...
I added a new parameter and considerately reduced the default value (now 0.1, previously 0.45).

I'm sorry to say this, but the brightness profile changes a bit with masks/scanlines used, so it's best to optimize the favorite setup and save the preset then.
I compared mask 5 / scanline 2 defaults with crt hyllian and it's there-there considering the screen lightness. But i think it's much better compared with a few versions back.

Personaly it's hard for me to go back with mask 0 / scanliens 0 now, hehe...
Feedback is welcome.

[Retro-Nerd]

Mask 5 probably looks very close to a real Trinitron TV. But somehow i still prefer mask 6 for some reason. I don't know, maybe it's because mask 5 is thinner/finer. And due to it looks a bit like generic black scanlines from the normal viewing distance on my monitor.

Anyway. Both are great, can't imagine a better CRT shader for now. We need this in WinUAE (i know, not in this state).


Mask 5, alternate scanlines 2






Mask 6, alternate scanlines 2





edit: Ok, on my 4K TV mask 5 looks perfect. For sure the best choice there.

[Dr.Venom]

Originally Posted by guest.r

Well, the "cutoff" (which produced black gaps)was hard coded into the shader, i changed that with the new version.
Unfortunately i can't access a trinitron crt, so i have to relly on user feedback...
I added a new parameter and considerately reduced the default value (now 0.1, previously 0.45).

Thanks, that last tweak on mask 5 is a bulls eye

Just a question on alternate scanlines 2.0. I have to use sort of "inverted" alternate scanline 1 values for "scanline bright" and "scanline dark" to calibrate to real trinitron.

So where with "alternate scanlines 1" I would for example use:
-scanline dark 1.2
-scanline bright 0.75

now with alternate scanlines 2 I need to sort of ïnvert" the bright and dark values to get to accurate results:
-scanline dark 0.7
-scanline bright 1.15

otherwise the bright areas get way way overblown (remember on real trinitron scanlines remain even visible for a tiny part in very bright / white areas).

I'm fine with setting these values for alternate scanlines 2, because the result is pretty much perfect, but it sort of feels inconsistent with the settings for alternate scanlines 1.0 (always been used to setting the scanline dark to a higher figure than bright..). Just thought I'd mention it.


Some last meticulous detail comparison...

In the image below a comparison is made between some details from the high resolution shot from Lomax from real Trinitron (as attached in this post) and the current shader.

In the top half of the image is real trinitron, in the bottom half the shader upscaled from 1080p source (4K would have been nice, but alas..). Two -minor- things stand out:

1. With alternate scanlines 2.0 the fall-off grade from very bright dots to something adjacent less bright is less gradual than with real trinitron. With the shader this makes some highlights to have a very slightly "artificial" standout.
2. In some parts of the image the "fall-off curve" for scanline height seems to follow a non-linear function, i.e. more concave , making some parts more elliptical, compared to the more linear (less "rounded" ?) fall-off in the shader.

Especially with regards to point 1 is there maybe some room to improve, or possibly even a configurable fall-off grade?

Just to be clear, I understand this is really about small details, so I'm totally fine if you leave everything as it is .


Please click for large and then click larger picture again for true size:

[guest.r]

Hello there,

i worked on the shader past few days ... added new scanlines, adjusted brightness a bit to work better with defaul settings...

Just a question on alternate scanlines 2.0. I have to use sort of "inverted" alternate scanline 1 values for "scanline bright" and "scanline dark" to calibrate to real trinitron.

now with alternate scanlines 2 I need to sort of ïnvert" the bright and dark values to get to accurate results:
-scanline dark 0.7
-scanline bright 1.15

I don't want to bring too many parameters into the shader, and scanlines are now plentiful. This results that same parameters and their default values are used in different formula and such things can happen, but dark and bright parameters still function in the same direction, even overlap but still work as intended.

otherwise the bright areas get way way overblown (remember on real trinitron scanlines remain even visible for a tiny part in very bright / white areas).

With mask 5 i lowered the maskLight value a bit, since it was a bit too much for the defaults and this influenced scanlines too.

I also introduced a new parameter called "Increased bright scanline beam" which can control the "growth" of bright lines/pixels. Before isolated bright lines/pixels swelled a lot and it's lowerd a bit by defaults. Comparing with other crt shaders some endorse "swelling", some keep it in bounds. So i rather made this optional. Can also influence beam fall-off.

1. With alternate scanlines 2.0 the fall-off grade from very bright dots to something adjacent less bright is less gradual than with real trinitron. With the shader this makes some highlights to have a very slightly "artificial" standout.
2. In some parts of the image the "fall-off curve" for scanline height seems to follow a non-linear function, i.e. more concave , making some parts more elliptical, compared to the more linear (less "rounded" ?) fall-off in the shader.

I added a parameter to control the fall-off grade. On 0.0 it will work stronger, on 1.0 as before. But in general, it's hard to influence this since the whole image get's altered and most scanlines become thinner.

It can be controlled by the "scanline dark" parameter also, since dark lines get thinner and the fall-off is steeper.

What i wanted to say, yeah some changes with "alternate scanlines". A very nice new version is now under 1.0, the replaced one is now 3.0. It's hard to find an equation which doesn't produce uneven scanlines / banding with 1080p and non-integer scaling and therefore i'm very pleased with the new no. 1.0. Old 1.0 (now 3.0) failed here a bit, so it's a good question to keep it or not, because it slows the shader.

It's also a good question whether to keep the fall-off parameter, since it's another chunk of code and i don't know if it helps to solve the issue in a satisfactory manner.

Anyway, the defaults are working decently with this version (as far i could test them) and i feel a bit bad if you need to test some things again.

[Retro-Nerd]

I noticed some image output issues generally in PC-Engine games. In some, like R-Type or Legend of Hero Tonma, your shader cuts a few lines from the bottom. Thought first it has something to do with my video settings in Retroarch but it doesn't happen with CRT-Royale (exact same option settings). Strange thing.


Cut bottom with CRT-Guest-Venom:










Correct Image with CRT-Royale

[nobody]

Guest.r, is there any chance the shaders work on dosbox? Meaning crt-guest and crt-trinitron. Also this crt-trinitron comes from retroarch?

PS I updated glsl shaders on retroarch and I see your crt-guest shader is included there now. Pretty good but maybe too much bloom/glow, it's a bit stressing to look for long time.

[Retro-Nerd]

Take the crt-guest-dr-venom shader attached here (including the preset file). This is much better what Retroarch uses officially atm.

[gimbal]

Originally Posted by Dr.Venom

Since Gimbal is asking for it, here's another one for his nostalgia belt

Guest.r 's shader with alternate scanlines (taken with version 1) and mask 0:

Thank you sir, may I have another!

[Anubis]

What happened to poor game?

[guest.r]

Originally Posted by Retro-Nerd

I noticed some image output issues generally in PC-Engine games. In some, like R-Type or Legend of Hero Tonma, your shader cuts a few lines from the bottom. Thought first it has something to do with my video settings in Retroarch but it doesn't happen with CRT-Royale (exact same option settings). Strange thing.

I can imagine two reasons for some missing lines. First could be that Smart Y Integer scaling is enabled and some overscan happens.

With second one there were some problems with the previous version of the shader that were fixed later. So it's best to update the RA GLSL shaders and then copy "our" shaders over it.

I tried the same game and it worked normally.

@nobody

Guest.r, is there any chance the shaders work on dosbox? Meaning crt-guest and crt-trinitron. Also this crt-trinitron comes from retroarch?

PS I updated glsl shaders on retroarch and I see your crt-guest shader is included there now. Pretty good but maybe too much bloom/glow, it's a bit stressing to look for long time.

Yeah i ported the "trinitron" version to Dosbox (gulikoza, Daum). Looks about the same as the WinUAE version.

With the RA version, as Retro-Nerd mentioned, you should overwrite the shader in /crt/guest folder with the version posted here. It's still a bit experimental, but nothing has changed to worse and new things like masks and scanlines are added. You couldalso change the parameters to make glow go away...

[Retro-Nerd]

edit: Ah, now it works. Deleted the whole "CRT" subfolder, updated GLSL and copied your latest "crt-guest-dr-venom.glsl" file. Great.

Absolute best shader for dosbox by a mile.

Yep, that's true. But i miss the selection for mask style 5 or 6.

[nobody]

Originally Posted by guest.r

Yeah i ported the "trinitron" version to Dosbox (gulikoza, Daum). Looks about the same as the WinUAE version.

Absolute best shader for dosbox by a mile.

[Dunny]

Still hoping for a PAL TV+RF Signal shader some day... All these monitors are very nice but I never even saw an Amiga running on a monitor until I got a PC. Nobody had them around our way.

14" TV FTW!

[Retro-Nerd]

You never had a PAL TV with RGB scart in the 90s? What a shame. I hate this RF antenna crap. Probably used it last around 1987 for my C64.

edit: There are a few shader that mimics a crappy RF connection (including all the analog artefacts like flicker, color bleading, distortion etc).

This one is from Retroarch/Shaders/NTSC folder.




A different PAL variant:

[Dunny]

Good heavens no! No RGB scart around until we all got ADSL TV boxes, which was around... what? '99 or more?

And only four channels on the telly too.

Neither of those simulate the colour inverse smearing very well at all. Is it not too much to ask for a shader that at least looks retro instead of imitating a box monitor from 2006?

</sorta sarcastic>

[Retro-Nerd]

These are default settings. There are tons of parameters to set. And probably other shaders to try.

[Retro-Nerd]

The Afterglow effect looks quite nice. But in motion/scrolling on a pure black background you can see ghost images/smearing.






Is this adjustable tweaking the Afterglow settings (which ones) without degrading the effect too much?