News Blur Busters releases authentic CRT simulator shader for high refresh OLED and LCD screens— 240 Hz+ OLED recommended for the best experience

[Admin]

A truly high-end, authentic CRT filter (in the form of a shader) releases, but only for high refresh OLED and LCD displays.

Blur Busters releases authentic CRT simulator shader for high refresh OLED and LCD screens— 240 Hz+ OLED recommended for the best experience : Read more

 

[YoyoMandrake]

But we still can't play light gun games. That's the real reason CRTs rock.

 

[bit_user]

I was glad to leave CRTs behind. I will say that when I got my Plasma TV, which was supposedly one of the best models ever made (Panasonic ZT60), it didn't look better than my Sony GDM-FW900 computer monitor (but the plasma was a huge step up from the 29" CRT monitor I used as a TV). The most annoying thing about that Sony was the floating blacks, at least later in its life (I bought it used). You could adjust it to have a good black level, but the black level would sort of wander as it heated up and over time. Maybe it was due for a tune up.

 

[helper800]

I was glad to leave CRTs behind. I will say that when I got my Plasma TV, which was supposedly one of the best models ever made (Panasonic ZT60), it didn't look better than my Sony GDM-FW900 computer monitor (but the plasma was a huge step up from the 29" CRT monitor I used as a TV). The most annoying thing about that Sony was the floating blacks, at least later in its life (I bought it used). You could adjust it to have a good black level, but the black level would sort of wander as it heated up and over time. Maybe it was due for a tune up.

Maybe a new set of tires and an oil change while you're at it. 😉

 

[Fox Tread33]

I was glad to leave CRTs behind. I will say that when I got my Plasma TV, which was supposedly one of the best models ever made (Panasonic ZT60), it didn't look better than my Sony GDM-FW900 computer monitor (but the plasma was a huge step up from the 29" CRT monitor I used as a TV). The most annoying thing about that Sony was the floating blacks, at least later in its life (I bought it used). You could adjust it to have a good black level, but the black level would sort of wander as it heated up and over time. Maybe it was due for a tune up.

Happy holidays. Yes.. I really loved my 50" Panasonic Plasma TV which started to get fine Sharpie like lines across the screen and in the most annoying places. At least one YouTube video, which was really great, showed how to replace the parts. I wasn't doing that, and just moved on to an LG 55" OLED TV.

 

[Co BIY]

But we still can't play light gun games. That's the real reason CRTs rock.

The shooting light gun games were always my favorite.

 

[mdrejhon]

But we still can't play light gun games. That's the real reason CRTs rock.

They do!

That being said, it's a simple Arduino microcontroller project (photodiode on Zapper, a muzzle accessory), with a wire leading to the composite output of the Nintendo. You'd use a different method of triangulating light gun position such as a Wiimote bar or IR beacons at the corners of your monitor. The Arduino controller would triangulate where it's pointing, and flicker the right part of the zero-lag composite signal into the photodiode. Voila.

There's already a GPL3 project that does this, AFAIK.

And by the way, I'm adding lots of bugfixes and quality improvements to the CRT simulator in Winter 2025:

- Fix chroma divergence behaviors
- Fix banding (some crappy displays are clipping RGB(0,0,0)-RGB(10,10,10) to fullblacks, interfering with Talbot Plateau Theorem brightness-preserving calculations in the shader
- Add a global mode (like an instant-scanout CRT, a phosphorescent BFI), for those who prefer that effect as a hybrid between BFI and CRT
- Etc

There's a neat trick that the CRT simulator does, where midcolors (e.g. midgreys) are brightened and peaked/pulsed out in the first refresh cycle of the CRT simulation, but bright colors cascades over a few Hz. This cleverly reduces motion blur more than the average BFI algorithm because of this variable-MPRT trick. This does mean a bit of extra ghosting on bright colors, but zero ghosting on dim colors.

You can adjust the "GAIN_VS_BLUR" variable, or RetroArchs' "Brightness vs Blur" variable. The chroma ghosting will have its own adjustment later in winter 2025, if you're getting color-tinted ghosts (I know why it's happening, it's from a math quirk...)

 

[mdrejhon]

I was glad to leave CRTs behind. I will say that when I got my Plasma TV, which was supposedly one of the best models ever made (Panasonic ZT60), it didn't look better than my Sony GDM-FW900 computer monitor (but the plasma was a huge step up from the 29" CRT monitor I used as a TV). The most annoying thing about that Sony was the floating blacks, at least later in its life (I bought it used). You could adjust it to have a good black level, but the black level would sort of wander as it heated up and over time. Maybe it was due for a tune up.

The nice thing is that 480fps 480Hz OLED has much less motion blur than a plasma now, and grandma can see 120vs480 OLED better than 60vs120 LCD ... That's the 4x geometrics and GtG=0 for the win, much like 1/120sec camera shutter and 1/480sec shutter (for 120fps vs 480fps non-strobed).

For people who don't understand the blur mathematics:

Motion blur of impulsed = pulsetime (or largest part of phosphor decay)
Motion blur of sample hold = frametime (so, spray lots of framerate at it).

Currently,

- Early LightBoost = 2.4ms MPRT
- Plasma = 3ms MPRT
- 480fps 480Hz 0ms OLED = 1/480sec MPRT = 2.09ms MPRT

So, there are now times where strobeless has less motion blur than strobed. It's amazing to see sample and hold displays that now has less motion blur than some early strobe-backlight displays. It's freaking amazing to see sample and hold having less motion blur than even a Pioneer Kuro home theater plasma (I used to work in the home theater industry...) The big fat bleeping red asterik is simply the framerate (480fps 480Hz) and the lack of good BFI in OLEDs, so that's why I am a giant fan of BYOA (Bring Your Own Algorithm)

1000Hz OLEDs are already in the laboratory, going to be demoed 2026, and on commercial market 2027. Remember, 4K cost as much as a car. Tomorrow's Hz won't be expensive forever. Look at how some 120Hz hit some office monitors, cheap chinese TVs, and a bunch of other phones/tablets. 240Hz is the next mainstreaming frontier, and our open source display algorithm potential exponentially expands from there upon.

I have formulas to emulate plasma on a future 600Hz OLED, and formulas to emulate 1-bit mirror flipping on a future 1440Hz OLED. I have separate components already (TestUFO Interlacing, and TestUFO Black Frames, and TestUFO DLP Color Wheel (view only at 240Hz+, or the low-Hz flicker is epileptic), and TestUFO Variable-Blur Black Frames (view only at 240Hz), and now soon TestUFO CRT (I'm porting the shadertoy to it early 2025), I'll have TestUFO Plasma by 2025-2026 too.

Generic Hz is freaking amazing for BYOA (Bring Your Own Algorithm) approaches, when we can now open source our own display algorithms, in today's dawn of open source display technologies. Yes, yes, yes, before you ask, yes, yes, yes, I tried to convince display makers. Meet brick walls and wasted time. So, it is BYOA from now on mostly for me. Much easier.

For some of these, display makers can come and buy service from me, instead of the other way around. Or download the open source algorithms, port it to their firmware, and put the required credits on the box, as per open source license (which helps market these accomplishments also). They know how to contact me.

I can even do it in a box, I'm helping Retrotink add the CRT emulator to some current or future device, so it can even be done in a video processor (just supply generic brute Hz). The display makers can supply Hz, and I can do the algorithms instead before the signal-input jack of the display.

Remember:

GtG = pixel moving between colors = like a shutter slowly opening/closing slowly.
MPRT = pixel static visibility time = lik the shutter fullopen time.

So, you want GtG=0.000 (not GtG less than refreshtime, GtG must be near zero refreshtime). That's why LCD ginormously throttled Hz differences (240Hz vs 360Hz looked only 1.1x different because of nonzero GtG).

TL;DR: Refresh rate geometrics and GtG=0.000 for the win for the humankind benefit, as seen at 120vs480 example.

 

[bit_user]

The nice thing is that 480fps 480Hz OLED has much less motion blur than a plasma now,

FWIW, my plasma had a 480 Hz base frequency, or something like that. I'm not sure exactly what that meant, but it seemed to me that perhaps Plasma works a little like DLP where the pixels are binary and they just use something like FRC to approximate different levels of intensity. It would explain the insane dithering noise that forced me to sit far enough from my plasma TV that I could no longer distinguish individual pixels. Once I did that, the image looked great!

Anyway, thanks for stopping by and dropping some knowledge on us. I won't pretend I followed everything in your posts, but I love the passion and your ethos of open sourcing this! Congrats on getting it out there and hopefully it's getting the kind of reception it deserves!
: )

 

[mdrejhon]

You're welcome!

There is a lot of variables that muddies Hz quality geometrics.

- It's why LCD seemed to have only ~1.1x difference between 240Hz and 360Hz (the nonzero GtG), instead of the expected 1.5x

- In plasmas' case, it's the 3ms phosphor turning its ~600+ Hz into something slightly less than the Hz number would suggest under the BlurBuster Law mathematics (3ms is more than 1/600sec MPRT).

This is why 480fps 480Hz OLED has less motion blur than a Pioneer Kuro plasma display now; it's very linear pure Blur Busters Law mathematics (which becomes perfectly true at GtG=0ms with no phosphor ghosting between frames). So in that case, a 480Hz OLED manages a darn near nigh perfect 1/480sec MPRT in sample-and-hold mode.

OLED and direct-view discrete LED, have absurdly fast pixel response, to the point where refresh rate differences are much more human visible, where during eye tracking situations, OLED frametime can perfectly equal the motion blur of equivalent camera shutter (120fps = 1/120sec shutter blur, 480fps = 1/480sec shutter blur), so you really see the 4x blur differences.

So the Hertz headroom, gives me a lot of room to reduce motion blur of low frame rates, like for retro material.

Rule of thumb for display upgrades: Geometrics AND 0ms GtG.