Sony Megatron Colour Video Monitor

RetroArch Additions Shaders
#3195

May i ask why you want to use v1 over v2?

The current v2 does have color accuracy issues in my experience, but the AzMods version i linked is based on a version of v2 that predates those issues. (And that is the version i still personally use, tho i have since made some additional updates and improvements to decoupled-guest-advanced-ntsc and decoupled-guest-advanced-pal, previously called ~crt-guest-advanced-ntsc-decoupled and ~crt-guest-advanced-pal-decoupled.)

AzMods also includes modifications that (seem to) make v2 work on older, non-nightly builds, if that is a concern. (Tho i would strongly encourage using a nightly build if at all possible. There have been gobs of HDR quality of life improvements, like the UI brightness setting.)

Other than that, the only substantial difference i can think of is the old tone mapping. I should be able to quickly and easily roll up a modded version of v2 with the old tone mapping if that is something you really want, but i think that the v2 tone mapping is strictly better in terms of results.

1 Like
#3196

Ok i understood now. There are V1 and V2 version, so i stick with V2, i don’t need to use V1. Since i tried your derived V2 and they seem working, i tried also V2 present under shaders preset folder and some works some not.

I read a post of Cyber user, that stated troubles with RX 6600, since i own RX 6400 maybe there are drivers issue with V2. If i well remember PVM shaders are “red”. Now i remove drivers and try again and then i install older version in order to see if old drivers fix this.

I jumped into these new shaders (for me) because now i own an OLED LG (48") and i can try a lot of beatiful hdr shaders. Unfortunately RX 6400 is not so powerful for reflection, but at this time i would be happy with working drivers and no reflection

thanks for your support

1 Like
#3197

Within Mega Bezel there are several performance tiers, try them all before giving up on reflections. There’s also a Sony Megatron Mega Bezel Base preset that you can try as well.

Besides Mega Bezel, there are other more efficient reflective Bezel Shaders like Koko-AIO and uBorder.

I almost forgot about that Radeon bug. Strangely enough it doesn’t seem to affect CRT-Guest-Advanced which is a shader that you can coax into having very similar output characteristics to Sony Megatron Video Monitor.

Also, Sony Megatron Colour Video Monitor has never been a requirement to use CRT Shaders in HDR mode since HDR is built into RetroArch so it could be applied to any shader.

From what I remember, the Radeon red screen bug only affects certain Mask Type so you can work around it and it also only affects the Vulkan video renderer, so you can just switch to DX11/12 and you should be fine.

Running DX11/12 requires you to enable HDR in Windows as well as RetroArch.

#3198

The issue was resolved briefly before returning after a nightly update and remains still.

So does Vulkan and has for a couple of years now for me at least. I remember HDR used to engage without having to enable it in Windows first, it was quite convenient

1 Like
#3199

Red color issues with V1 for Amd Rx 6000 series started after driver 24.5.1 and is broken ever since then. V2 was working fine at the beginning but recent updates broke it again. Red color issue is only for aperture mask, slot and shadow are working as intended.

1 Like
#3200

@Wilch1 @bigretrofan

If you have time, please do a completely fresh install of the latest nightly, enable HDR, extract this archive to the shaders_slang folder, and check which presets in the base “megatron red tint bug bisection” directory exhibit the red tint issue.

https://www.mediafire.com/file/mr9qwo8cap4kodg/megatron_red_tint_bug_bisection.zip/file

1 Like
#3201

@cybermat Whoops, missed tagging you. Please see previous post.

#3202

Both hdr158x files have the bug. All hdr26x/27x versions don’t have the bug. All hdr28x versions have the bug.

So hdr261-1 fixed the bug and hdr280-20260227 reintroduced it?

1 Like
#3203

Yep.

Technically it was probably hdr260 that fixed it, but hdr260 had broken, dimmed output even on release, so i excluded it from the test.

Just to be fully upfront: i’m not going to make an attempt to fix this myself at this time. hdr280 radically overhauled the shader in ways i have not even begun to investigate, and i don’t really have much interest in doing so given that hdr265 onward all have inaccurate color (on at least my setup) anyway.

1 Like
#3204

What about hdr27x? Same colour inaccuracy? I’d like to stick with vulkan as dx11/12 has some graphical glitches when using shaders

1 Like
#3205

Yep, all variants of hdr265, hdr266, hdr267, hdr271, and hdr280 have inaccurate color (again, at least on my setup, but i’ve seen no reason to think it isn’t universal to at least all modern RWBG WOLEDs.) Everything is a smidge desaturated and has a slight reddish tinge. The tinge is particularly noticeable on a grey ramp test pattern.

Compare shaders off, hdr261/hdr262, and hdr271-20260225 for yourself using the test suite and see what you think.

I also use Vulkan, and i have personally continued using my modded version of hdr261, a slightly older version of which is linked here. I should probably post an update to that soon with the decoupled-guest-advanced-ntsc/pal updates.

3 Likes
#3206

I don’t know if anyone will find these posts again in searches, but in case anyone has the same problem and finds this: it’s getting worse.

Now I can see these brighter vertical lines in some other shaders as well (thankfully not in all, yet), and I have the impression that I glimpse them even without shaders now, although without shaders they still almost imperceptible (at least for now).

I don’t know if I was just unlucky enough to get a defective unit and didn’t notice these problems in time, but I think it’s best to just avoid this monitor if you plan to use it for retrogaming with shaders: LG Ultragear 24GS60F-B.

Thank you again to everyone who helped to try to understand this problem.

#3207

You know, subpixel layouts aside, all this preset probably needs is an increase in Saturation to look good.

It should be able to easily match the screenshot from CRT Guest Advanced below it.

#3208

Sorry about the pictures thing, it takes a lot of physical energy for me to do that with the way I have my setup (+old) and it’s basically waiting for the stars to align on physical energy + working on shaders at the same time.

2 Likes
#3209

It’s okay man. I know there must have been some good reason. No pressure.

#3210

In hindsight, just the one white subpixel.

1 Like
#3211

I’m new here. My JVC CRT recently died, so I have re-entered the world of emulation. I am amazed that we now have phosphor-level emulation. I have a Trintron here for comparison, and shaders like Megatron and Royale look just like the CRT when I look closely as the phosphors.

So, thanks to everyone involved with making this dream come true.

Also, the discussion here is fascinating. It has warned me to never buy a panel with a white sub-pixel! I currently use a QD-OLED that works perfectly with these types of shaders. I am considering a new monitor but will wait for RGB stripe or a newer QD.

I haven’t noticed any macro shots of the shader running on a QD-OLED, so I’m curious if anyone can link some. These labels use a triangular sub-pixel formation. Has this been known to cause issues? Did development of these shader take this into account? It seems that the origin of these shaders assumed the striped pattern of LCD displays.

1 Like
#3212

You havee this mixed up. You should read more then you’ll see where we “solved” the white subpixel issue and that QD-OLED turned out to much worse than originally hoped when it actually landed. QD-OLED is the worst current display techology fror subpixel emulation because it’s triangular subpixel layout doesn’t in any fashion lend well to 1:1 mapping with a CRT’s vertically striped RGB phosphor patterns unlike every other modern display technology, including WOLED with the white subpixel. The original issue is that none of these things were documented until through sheer trial and error and determination by members of this community, we figured out how OLED subpixels work and how to get great CRT emulation out of them (at least the WOLED ones). QD-OLED will never be accurate in it’s current form, there’ll always be a little “bobble” or something worse.

Second thing is brightness, Only the highest end OLED displays of any kind can compete with even good midrange miniLED TVs on the market when it comes to Peak and especially when it comes to sustained full screen brightness, which is a staple of accurate CRT emulation, especially if you want to improve motion clarity by the use of BFI.

You’ll have to ask the CRT shader enthusiasts who also use QD-OLED for that. The only one I can identify is @MajorPainTheCactus. Striped QD-OLED just like regular triangular QD-OLED remains to be seen in order to be proven worthy.

In all of the promotional material, it shows the red stripe being taller than the green one and the green one being taller than the blue one. How that would work or map to RGB phosphors which are supposed to be of the same height, only time will tell.

#3213

I took some macro shots of my 32" 4K QD-OLED monitor. At extremely close distance, the camera can see the individual dots. So, for example, a green “phosphor” which would be a solid stripe on CRT is comprised of several tiny green vertical dots. That said, using my actual eyeballs at a distance of six inches or more, I can no longer see the dots. It just appears to be a solid stripe.

The only real “problem” is the triangular format. I’m not even sure it matters for 480i/p content or lower. The virtual phosphors will be large enough that the sub-pixels aren’t visible. This is all in regard to the extreme pixel density of a desktop monitor. I suspect it is a different story on a large OLED TV. I also suspect it would become visible if using a more detailed CRT mask (rather than aperture grille).

#3214

It also depends on how pedantic or picky the user is or their particular goals.

We can always accept compromises but if someone wanted the display with the ideal subpixel structure or characteristics for CRT emulation, QD-OLED would be very low on the list if not off of it completely, especially given the fact that there are already superior alternatives (in this niche of a context) which are readily available on the market which would provide that near 1:1 mapping between modern display subpixels and emulated CRT phosphors as intended by the CRT shader.

Even on any other display type using the wrong mask layout for the subpixel layout may not cause highly obvious appearance issues from typical viewing distances.

However, there would be anomalies caused by the inaccuracies regardless of how “acceptable” one may deem the output to be.

If you already have what you have then you can and should continue to make the most of it though.

I’m curious to know how QD-OLED will render something like this:

This is a test I made which you can run if you want to share what QD-OLED displays are capable of on the CRT-Shader emulation front: