There’s at least one preset with a much shorter filename in the same folder and that loads correctly. That’s in the log as well. Also in the log, I can see where the longer filenames are being truncated.
I’ll post the entire log for you. Also note that this pack hasn’t been updated in a very long time and all presets were tested and working a long time ago.
Update: @Hunterk Seems like I could definitely optimize and update these paths to get the presets to work.
not including " 185 characters limit
[WARN] [Shaders] Could not read shader preset in #reference line: "C:\Users\xxxx\Downloads\RetroArch\shaders\Mega_Bezel_Packs\CyberLab\MBZ__3__Standard_Full_Reflections\4K_Optimized_Presets\Console_Specific_Presets\CyberLab_Turbo_Duo_Composite_Slot_Mas".
not including " 228 characters limit
[WARN] [Shaders] Could not read shader preset in #reference line: “C:\Users\xxxx\Downloads\RetroArch\shaders\Mega_Bezel_Packs\CyberLab\MBZ__0__Smooth-Advance_Full_Reflections\4K_Optimized_Presets\Console_Specific_Presets\CyberLab_Genesis_Composite_Slot_Mask_IV_OLED_for_CyberLab_Blargg_Video_Fil”.
Here is a second log after optimizing paths and fixing 15 of the 17 presets:
Is there a character limit which is causing these final presets not to load? They’re referencing presets several folders deep and each of them has relatively long filenames. Is it that the character limit was lowered at some point over the last few months or year?
This is what I like to hear! Feel free to post pics showing how the stuff looks to you. I don’t get much visual feedback from users.
This would have to do with the TVL and sharpness of the preset most likely. You can try presets with “Fine”, “Near Field”, Le’Sarsh", “Smooth” or “PVM/Pro Monitor Edition” in their filenames as well as some of my newer Shadow Mask presets for a less harsh/coarse Mask/Scanline experience. Do note that if you sit close enough to a large enough CRT TV you would also notice the scanline gaps and screen door effect provided by the Mask/Phosphor especially on a lower TVL set.
You can also try lowering the Integer Scale Factors which will give you a smaller window/viewport size. Things tend to look a bit denser and clearer as you scale down in my opinion and you get the added bonus of additional brightness headroom.
Another thing to consider is that due to the triangular subpixel layout of your QD-OLED monitor, you’re not able to experience the emulated Mask and Phosphors rendered correctly and the imperfections and artifacts of this imperfect rendering might be slightly distorting the experience especially when closer to the screen.
For proper subpixel Mask and Phospher Rendering you’ll need an RGB Stripe Subpixel Layout LCD for perfect rendering, BGR Striped Layout LCD for near perfect rendering or RWBG WOLED Display for good but not perfect rendering.
You can see more here:
If you have the time, feel free to run the following tests and post the results. I’m always interested in gathering data about how different panel types perform with shaders:
I finally had a lull in my work schedule and sat down to try out the Legendary PSX Composite shader with SotN.
Not like one of my most cherished games, but it’s such a great game to understand shader differences.
After I set my HDR brightness to 2300 nits things were pretty nice, but VERY green.
Switched to Mask 13, Size 1, BGR, and had to put Base (black) Mask Strength to 0 but now I feel like I finally got to experience a true Cyberlabs Shader.
This is the subpixel order of Mask 13 on a white screen. Interesting.
BXRYWC
Things generally look pretty perfect. At least for my current understanding of what things should look like Shader wise.
Mask 8 Size 2 either order looks just like Mask 13 Size 1 BGR. I believe that’s the panel agnostic mask?
Color correctness wise, this probably makes sense if we think of BWRG to be 3 subpixels wide at 4K Size 1 Mask 1.
B+W, W+R, R+G
Normally I very much dislike when there’s a black line in the Mask. Like RRGGBBX doesn’t look good to my eyes. 300TVL in general doesn’t look good to my eyes, too thick. Mask 12 also has that VERY green tint to it on top of all the solid colors like a blue or red background looking unusually broken up into unnatural TVL’s with extra thick black lines breaking up the TVL’s. Like I’d get two normal TVL’s then an all black one. (to my eyes)
Mask 14 has good color correctness, but this same kind of odd thick black line in solid color backgrounds:
Good thing my CyberLab Guest Legendary Preset Pack has only 1 preset (out of 48) which uses Mask 12, which was only added less than 24 hours ago.
Very good.
A picture paints a thousand words.
Hmm, while I’m glad you’ve gotten things to look good on your setup, if you look around this thread, you might notice that one of the signature characteristics of CyberLab Death To Pixels Presets is being able to go straight up to the screen and see the RGB phosphors in all their glory.
This same pattern would look very different on different panel types depending on subpixel layout. It’s like you’re asking me to translate something which is visual in nature by using theory and science, while I would much prefer to see what the output actually looks like on the particular display in a practical sense, especially with me not being an expert on subpixel science.
More theory but no testing, where are the macro photos which would make this all easier as we would see what is happening with the alignment and subpixel order e.t.c.?
I have no idea. Is this a screenshot or photo? Scaling settings can affect things like that. Are you using Integer Scaling on both the X and Y axes?
I don’t know what any of these Masks look like off hand so I’ll have to refer to the CRT Guest Advanced thread or load up a preset and look at them myself. I know that one of them should be a B&W Mask so yeah, that should be subpixel agnostic.
I’m not trying to pressure or force you to do anything especially since you indicated that for you, taking some photos of the screen showing different known RGB subpixel mask patterns may not be as simple as one might think but I do think that this path seems like beating around the bush and going down a long winding road of workarounds in order to get these things looking good on 4 Stack BWRG Tandem WOLED displays which have been on the market for well over a year now.
I recall when I first became “subpixel aware” when it came to these things, it was thought that the only subpixel mask which could work properly on WOLED displays was the B&W Mask. Even Sony Megatron Colour Video Monitor had a B&W Mask. The more I learned and observed, the more I tested and researched. Until I came across a post by @nfp0 where they said that they used Mask 13 (at the time) RRGGBBx, Size 1 for 4K.
Previously I was using Mask 6 (RGB), Size 2 (RRGGBB) and I’m not sure it looked correct on my RWBG WOLED from the perspective of what a subpixel aware CRT Shader Mask is supposed to do, which is mimc a real CRT’s Phosphor and Mask structure/appearance as closely as possible on a modern display.
This led to the creation of the “OLED Subpixels How Do They Work” thread because I realized how badly some of the masks looked on my WOLED display and wanted to see if better could be done. Through documentation, trial and error, a little luck and sheer determination, I realized that Mask 12 RRGGBBX with the BGR layout (XRRBBGG or RRBBGGX) finally gave a near perfect phosphor triad with excellent alignment.
Prior to this I enjoyed CRT Shaders on my OLED very much but they didn’t look quite right when right up to the screen. I even experienced this when I made my CRT Royale preset pack because that uses exclusively RGB layout masks, of course Hyllian has now made a BGR version available but we now know that even this wouldn’t work for RWBG WOLED as it requires an RBG subpixel layout.
This is ironic because CRT-Guest-Advanced never “advertised” or documented an “RBG” supixel order mask as being available so it seemed like a mere coincidence that that was the missing link. The evidence adds up as well because before my advocacy and inclusion of IV OLED presets, the consensus among the CRT Shader crowd was that WOLED could not be properly supported due to the white subpixel getting in the way of things.
Just to demonstrate how accidental this might have been, Guest.R attempted to update the “BGR” Subpixel layout for improved alignment and it completely broke RWBG support. Things went from something that looked like CRT phosphor triads to the mangled mess it used to be before. I mentioned it to him and he kindly and promptly reverted the change.
Right now I’m having some issues using “BGR” Mask Layout 1 on my TCL QM751G which is supposed to be BGR so I’m actually using RGB Mask Layout 0. I know very well not to bring it to Guest.R’s attention before I have gathered all of the requisite photographic empirical evidence and formulated exactly how I’m going to describe my “issue” and what I might want to ask him to do about it. I know that I most certainly need to start by verifying the subpixel order of my display on my own as although the model has been tested and stated to be BGR, who knows, maybe my size uses a different layout? Only proper testing would show. I’m having fun though the emulated phosphors do look “RGB” with proper alignment although the slot masks sometimes look a little strange using the naked eye, they look like the horizontal slot isn’t going straight across, it looks like the slot is slghtly higher over the red subpixels. It could be an illusion though. Maybe glasses or some macro photos might enable us to better observe exactly what is taking place.
One thing for sure is that “BRG” (RBG) Mask Layout 1 does not look correct on my current display.
Of course if only finer masks with higher TVLs are your cup of tea or if you never go right up to the screen to peep at the RGB phosphors you might think that everything is above board which is exactly what happens with my CRT-Royale or my older presets which used Mask 6, Size 2 on my RWBG WOLED display because this is not something you can see with the naked eye at typical viewing distances. So ignorance is bliss and I think that this is what is happening with the vast majority of users who are buying OLED displays for CRT emualation because they know about OLED’s fast pixel response times and their true blacks and perfect viewing angles. They might know about the CRT-Beam-Simulator as well.
Most of them may not know about subpixel mask emulation and what it entails or requires so they might go out and purchase a QD-OLED display thinking that they got the best of the best tech not knowing that QD-OLED is actually much worse than RWBG WOLED for subpixel aware CRT Shader Emulation and they may not know that OLED’s may not be the best displays for accurate CRT-Shader emulation due to their relative lack of brightness compared to other technologies.
Why do you say that? I was very happy when I made certain breakthroughs and first got Sony Megatron to look good on my setup. It took me years of trying before I did so don’t feel so affected by my apparent lack of enthusiasm. Remember my perspective might be different.
It’s like you’re on the inside “struggling” to get these things to work, while I’m on the outside knowing how they’re supposed to and how simple they should be to work because all of the work has already been done and I wish I could help you more but my hands are tied and I’m blind so I can’t see what you’re seeing. Then I read that you have to do this workaround and that workaround and I know that none of those are in the playbook so there’s nothing for me to celebrate when it comes to that as I already have my vision on where I want support for the new subpixel types to be and I know it shouldn’t be too difficult to get them supported but a don’t have access to one of those displays to do the groundwork and testing on my own like I did with RWBG/WOLED.
So I’m still waiting and I know that even without what I’m hoping to see being implemented people can still see an image on the screen and think it looks amazing from far enough and once the TVL is high enough.
It is only slightly disappointing to read that you found the image to be “green” and to hear you making judgements on RR-GG-BB-X Masks when the display that you currently own maybe isn’t even displaying them correctly.
Did I say something wrong by letting you know that I don’t remember all CRT-Guest-Advanced Mask Layouts off hand? Or by admitting that I’m more comfortable with practical methods, trial and error, reverse engineering and problem solving as that’s what I can do with my limited academic knowledge of these things?
Don’t mind me at all. You celebrate! Remember I can’t see what you’re seeing so maybe that might have been a little more exciting if I could share in what you are experiencing in that way. I do envision the potential of one of the best displays on the market though so I look forward to another “OLED Subpixels How Do They Work” moment like this one:
Yeah it has definitely been an interesting time with different Megatron and nesguy presets.
I would try maxing my nits out (2300 is my 10% window max) testing Megatron and nesguy presets.
But I don’t think I cycled through the Masks with the NESguy presets in combination with the Nits higher than 500. I legit felt so lost looking at what feel like near monochrome or at best, overly green colors.
I’ll try some photos later. As long as work doesn’t ramp up again I’ll try and get some close-up pictures while I’m re-focusing on this.
This even let me finally experience BFI in action without detracting brightness drop. Cause OLED motion on its own certainly isn’t Plasma I must say…very curious about these pulsar displays and what the state of $2500 TV’s will be like in 2032!
Didn’t even have to turn up my Luminance past where it needs to be for the gamma curve on the Luminance tracking chart.
Whatever technology comes next is going to be really surprising if it’s not capable of all the CRT emulation aspects that are super important. Look at them chase 1000hz refresh rates with such enthusiasm cause of those Blur Busters hype. Things were so stagnant in the display industry when they introduced 4K even though nothing could handle the bandwidth needed for it. It’s refreshing to see advancing technologies again.
Peak Luminance, Paper White Luminance and HDR Brightness are not absolutes which correspond to the specifications of your display. They are all relative and conditionally based on the many other setings which can affect Luminance. So try setting these values based on eyeballing if the preset looks bright enough, rather than a specific figure. If you push it too far you need to learn how to spot clipping and oversaturation. Running your favourite preset through the 240P Test suite can help with that if you aren’t sure what to look for in your favourite games.
You have to get used to dialing in the de judder and blur reduction settings to find the sweet spot but that doesn’t apply to gaming scenarios though. For, that you need higher refresh rates and CRT-Beam-Simulator type technology which works best on OLED so Pulsar isn’t really needed.
A few posts up I left a link for Shader Beam. You should really be using that or prepending CRT Beam Simulator or koko’s Apaptive Strobe BFI instead of old school BFI. They provide better motion clarity with less brightness loss and perceived flickering than traditional BFI.
We already have TVs that can do much better than what I’m using which does an excellent job:
Folks who may not be following may not quite appreciate how far ahead of any other LCD TV to date the TCL X11L’s contrast ratio is. Of course OLED is still superior at least on paper and in absolute terms but with the sheer number of zones (20,000+) and with a high degree of dimming precision (26 bit) which means much more granular control of the brightness level of each zone and the additional color gamut awarded by SQD technology which improved upon the Quantum Dots used in previous QD Displays as well as the improvements to the LCD Colour filters themselves to allow more light to pass through (11,000+ nits) while preserving more colours makes that display a serious contender and I haven’t even talked about the new processing chip with 4x 48Gbps HDMI 2.1 Ports. It also has a strobing backlight like the rest of the TCL lineup.
If only it also came with a native 180Hz panel/refresh rate at 4K RGB 4:4:4 full colour that would have been the icing on the cake.
The QM8L should also be no slouch even though it’s contrast ratio is way lower than the X11L.
They both have a true RGB subpixel layout.
The G5, G6, C6H and Panasonic Z95B all have great potential because remember the entire screen isn’t active when doing accurate CRT Shader Emulation. The scanline gaps aren’t using any power, the pillar boxes at the top and bottom of the screen aren’t using any power, the black borders at the top and bottom of the screen aren’t either. The white subpixel is off, leaving more power budget available for the red, green and blue subpixels. So what you might see in RTINGS and other traditional reviews may not take all of these factors into account.
This is what MajorPainTheCactus the creator of the Sony Megatron Colour Video Monitor CRT shader and the person who implanted HDR in RetroArch has to say on this topic:
Yeah I follow it all via HDTVtest’s website. HDR settings in RetroArch being simplified just to the one setting maybe helped a lot here too.
HDR is a lot to learn and take on, so it’s good if you can find which settings to leave static and calibrated.
Like can’t wait for ELMB to be in that 2032 set of flashship TV’s with 480hz+ etc. - we see where we are now with Super QD LED and this stuff, things are finally advancing again. The next decade should be better. We can dream!
The bad news is that Mask 13 was only working in that one game. When I hopped over to Sonic 2 on the Genesis, Mask 0 at Size 3 is good, but tons of deconvergence artifacts. I guess, fringing?
Mask 13 doesn’t work here the same way.
The good news is I’m a lot more motivated to take these photos since I’m understanding a lot more what I’m supposed to see vs. what I don’t see.
I’m also salty here, (Philadelphia word, means feeling stupid) because Mask 13 size 1, Mask 6 size 2, and Mask 0 size 3 are all 300TVL, correct?..sometimes my brain just gets a little frazzled these days. :x
Guess I’ve outgrown 600TVL finally without realizing it and it’s 150TVL that looks bad to me. lol
Which of course right? I was thinking Mask 0 Size 4 and forgot it’s for 8K. (6k?)
@Cyber - OK I’m back with another question. You had said that you would like to see a few screenshots and how you don’t get much visual feedback from users.
I wanted to show you how I have it all setup now and how it’s looking really good to me, and see how it looks to you.
But I have no idea how to get actual screenshots showing HDR on here. The .JXR files from the screenshots taken are way too big to post, like 25-30 MB. I can just change them to JPG and they keep the HDR brightness but they are still the same size. Compressing them converts them to JPG and brings the file size down to 2.5-3ish MB, but they lose all the brightness.
Did I say screenshots? If I did, I meant photos of the screen. Screenshots aren’t going to show the impact the display’s subpixel layout and brightness has on the preset and that’s the type of feedback I’m interested in.
On another note. You can convert *.jxr to *.jpg using IrFanView. Also using IrFanView you can set a custom size limit for *.jpg files and it will adjust the compression/quality settings to fit within that limit.
If you want to post full *.jxr files, you can use a cloud storage and sharing service but I repeat, *.jxr files will only show me what I already know about the presets so I’m not really interested in those.
How does something like this look? I’m no photographer and I’m sure there’s all sorts of settings I could use to get better photos of a screen, it’s definitely hard to get it just right, and I had to compress this from my original photo as that was still a little too big.
For this one I’m using one of the Legendary presets, this specific one is from the Experimental folder, but I’ve been going back and forth on this one and the regular Super Smooth one:
CyberLab Guest 4K HDR Game BFI SNES Composite CyberTron Super Smooth Advanced S.slangp
It looks really good to me! Feel free to take some close up shots as well. Have you began experimenting with BFI/CRT-Beam-Simulator/Shader Beam yet?
I posted some tests which we (the community) can use to evaluate different display types’ performance using subpixel aware CRT Shaders. Feel free to give them a go if you have the time. There are many tips on taking photos of CRT shaders on the screen which also work for taking photos of a real CRT screen!
I took a look at the BFI options in Retroarch but when I tried turning it on it said it disabled the hardcore retroachievements which I tend to like to try for (makes me try some of these older games without things like Save States) so I just let it go for now, I’m definitely interested in learning more at some point. I feel like I’ve just learned a whole lot about shaders in general and how they work with this new monitor and I’m going to just try to actually play some games for a while without falling down even more rabbit holes!
Thank you again for all your help and advice. I’m a child of the 80s, and the early-mid 90s was where I really fell into PC and console gaming, so it’s been a lot of fun trying to do things to reproduce what screens were like back then. I’d actually love to go hunt down some actual CRT monitors and TVs but my wife would probably kill me.
No problem. Shader Beam is independent of RetroArch so it might work with RetroAchievements if you eventually decide to try it. Enjoy it and spread the word to others.
By the way, what HDR Brightness/Peak Luminance/Paper White Luminance settings have you settled on?
To calculate the TVL of a subpixel aware CRT Shader we take the number of vertical pixels of the display then divide it by the number of subpixels in the CRT Shader Mask Layout.
So for Mask 12 RRGGBBX @4K that would be 2160/7 = ~308TVL
for Mask 10 RGBX @4K that would be 2160/4 = 540TVL
for Mask 6 RGB @4K that would be 2160/7 = 720TVL
for Mask 6, Size 2 RRGGBB @4K that would be 2160/7 = 360TVL
for Mask 10 RGBX @1440p that would be 1440/4 = 360TVL
for Mask 6 RGB @1440p that would be 1440/3 = 480TVL
for Mask 6 RGB @1080p that would be 1080/3 = 360TVL
