Yes, I tried those as well and the issue is the same.
I am going to do more testing - it may be an issue with the TV and the colorspace. I tested on my computer monitor (it is not 4K, but is at least HDR (to a degree)), and it appears much better than my main 4K tv.
Keep me posted as there are others suffering similar issues (such as @BendBombBoom) that I’d like to get to the bottom of.
I’ve been quiet for a little while as I’ve been concentrating on colour and more specifically colour grading and using proper colour systems such as NTSC-U. I havent submitted everything I’ve done just yet (I will tomorrow) but the results are promising.
The biggest thing to note is that I’m using the same ISO, WB, Aperture apeed etc on these photos and so I’ve managed to bring the brightness of my monitor up. We’re still not quite there as my CRT is a little more washed out in its colours than the LCD version - the LCD being more saturated. I’ve turned on ‘vivid’ mode though which saturates the image a little more, as this in person seems to be closer - but I’m not sure whether its just more visually apealing and thats why I prefer it.
Certainly in person I feel the CRT and LCD are far closer together than before in terms of colours and possibly a little closer than these images would show (but then again the camera never lies)
Anyway here are a number of screenshots comparing the two and again remember they are using exactly the same camera settings:
CRT Photo: OnePlus 8 Pro Camera: Pro Mode, ISO 200, WB 6500K, Aperture Speed 1/60, Auto Focus, 48MPixel JPEG.
LCD Photo: OnePlus 8 Pro Camera: Pro Mode, ISO 200, WB 6500K, Aperture Speed 1/60, Auto Focus, 48MPixel JPEG.
That looks great. CRTs always look weird in photos, so i would chalk any small differences up to that. Good job on the blue water instead of purple
This is the main reason I visit the forums now. Looks amazing. I might actually prefer the way the LCD looks now, color-wise.
I think you need a new name for the shader now that you’ve added all the masks in.
Thanks!
So the main reason the water is now blue is because I put the white point up to D93. I find this a bit strange as it kind of says my PVM has a white point of D93. I suppose it might have as Sony is Japanese but the PVM is a PAL region PVM. It’s a bit odd - I reset all the options to ‘NORM’. I have two of these PVMs so I might see if it’s been effectively fused off to D93 by someone ‘fixing’ it over the years.
The other thing it could be is that it’s playing Secret if Mana that was developed in Japan. But this is the NTSC version playing on an NTSC SNES.
That’s a lot of colour systems there!
Thanks for your kind words - it really means a lot to me. Funnily enough Im getting to that conclusion too - I actually prefer the LCD version. I won’t say that in front of my old two girls (the PVMs that is ) though - I will always love them.
In motion they are slightly better - my Eve Spectrum seems to be quite good at reducing blur compared to other LCDs. I think the Blur Busters did a good job calibrating it even without back light strobing - maybe? It’s the poor no. Of darkening zones that let it down in HDR though.
As for the name I think you’re right. The only thing is though that I get quite boring names unless I make it some general name that doesn’t really describe what it does. Hmm have to have a think - any ideas are welcome!!!
Right so another attempt at getting the colours right! I think we’re basically there - but see what you think - we’re a smidgen out.
Ok so I added in the option to tweak the gamma value and between that and pumping brightness and contrast up I could get the CRT to look quite a lot like the LCD in terms of the colours at exactly the same camera settings.
I do think though in person this isn’t as nice as slightly lower settings more saturated settings on the LCD.
What this is causing headaches with is what are LCD pictures and what are CRT (one give away clue by the way is that some photos have the blue ‘on’ light of the LCD below. Its a nice problem to have at this point as it means I’ve pretty much reached my goal. I’ll keep tweaking this though.
One thing to point out is that the camera on the LCD screen gives a red halo around the center of the screen so thats why the door frame is slightly more red - its not visible with the eye.
CRT Photo: OnePlus 8 Pro Camera: Pro Mode, ISO 200, WB 6500K, Aperture Speed 1/60, Auto Focus, 48MPixel JPEG.
LCD Photo: OnePlus 8 Pro Camera: Pro Mode, ISO 200, WB 6500K, Aperture Speed 1/60, Auto Focus, 48MPixel JPEG.
that’s really beautiful, man. I think we can finally put to rest the old “crt shaders are cool, I guess, but they don’t look real” adage. Even the close-close-up shot is indistinguishable unless you know exactly what to look for.
…and just for fun Sonic in the waterfall - this probably fails the transparency test even on my PVM because the signal is too clean.
CRT Photo: OnePlus 8 Pro Camera: Pro Mode, ISO 200, WB 6500K, Aperture Speed 1/60, Auto Focus, 48MPixel JPEG.
LCD Photo: OnePlus 8 Pro Camera: Pro Mode, ISO 200, WB 6500K, Aperture Speed 1/60, Auto Focus, 48MPixel JPEG.
That effect is not just about simulating screen characteristics but also signal characteristics, which can vary depending on the input type used on the display as well as the characteristics and quality of the output circuits of the different consoles themselves.
Now that you’ve conquered PVMs and Slot_Masks you’ve basically created a great foundation. It would be nice to see you tackle some of the other “problems” that still exist in a similar, patient and systematic fashion.
So maybe getting not just rainbow effect or de-dithering but Genesis Composite and RF out right might be a nice next step. It could be modular in that one module can be for the TV input side of things that remains constant across consoles and the other module can be unique to the different consoles’ output qualities themselves.
I think the PC Engine also did some unique tricks at the output level in order to display additional colours which is why they added a Composite/RGB toggle in Mednafen/Beetle.
I’m seeing some small but not insignificant differences here in these comparison shots.
The vertical gaps between the jailbars seem wider and the black background seems darker in the LCD vs the CRT. Also the width of the jailbars themselves seem thicker than the width of the jailbars in the LCD. So there’s a difference in ratio. The LCD looks more 1:0.8 or 1:0.75 between the Jailbars and the gaps but the CRT looks closer to 1:0.6 or 1:0.5 not exactly but 1:(something less than what the LCD is)
The black level seems lower on the LCD also. It could be as a result of this or some other reason.
Be sure to zoom in on the cropped screenshot to better see what I’m talking about.
It’s an elusive goal because the way any composite signal actually looks depends on the quality of the comb filter used by the TV, which varied widely, so choosing a reference is hard and partially depends on what the desired look is. A PVM might be perfect for this as it has no comb filter, meaning all the composite video artifacts are present and dithering is completely blended (If the dithering isn’t blended, I don’t really see the point).
That’s really easy to fix, and CRTs weren’t all the same in this regard anyway, since they were calibrated to have different amounts of overscan and such.
That’s a plus in my book, one of the reasons why I kinda prefer the LCD shots in some of these comparisons. That’s also easily adjusted. In any case it’s within the margin of difference between individual devices, I’d say.
I understand and agree to some extent but these small differences could add up to the difference between dithering effects working properly or not (or not as good) in some instances as well as keeping the shader from achieving what might be perceived as the “natural or analogue” CRT look and feel. Another thing the increased sharpness and contrast might do is cause the LCD version to be ever so slightly more aliased compared to the CRT version. So we’re losing out on some of the natural anti-aliasing that CRTs generally tend to produce. We’re not losing it completely just reducing it slightly.
These differences are so subtle that they could be due to differences in the photography and even though some might prefer the LCD image, the difference is there so if we’re going for the ultimate Pepsi challenge shader with this particular model and sample PVM there can still be a few adjustments to make to make it even closer to that goal.
In this case whoever is up to the task can just select a TV which displays the said effects to their liking and use that as their reference. A set can be selected via many different means. It could be random based on photos, memories, a poll, a competition. Criteria could be supplied and screenshots or suggestions submitted. This could be done in so many ways.
Ultimately it might just boil down to whatever reference set the shader creator has access to that can give the output that we’re trying to achieve.
Another desireable side effect of slightly widening the jailbars and reducing the gaps would be slightly improved brightness.
I think any difference in black level or ratio between the LCD and CRT at this point is well within the margin of difference between individual CRTs.
I’m not seeing any difference in sharpness between the two at this point. If anything I think the CRT might be barely sharper, but again, the variation here is within the range of variation between CRTs.
This is another good reason to choose the PVM, since @MajorPainTheCactus has access to one already.
There are several types of comb filters and several manufacturers of each and they all look different depending on the manufacturer.
Basically a comb filter recovers lost signal and reduces artifacts. A PVM lacks a comb filter and just displays the raw signal. So if we start with a PVM and get that dialed in, we can then emulate the entire range of variation in composite output among CRTs by reducing the composite video effects, however they’re implemented.
This is precisely what I meant when I said this:
To clarify, I meant “if we start with a PVM showing composite output” using the same methodology employed so far in this thread. That would seem a logical place to start, IMO.
Yeah so yes you’re right there does seem to be some softness in the CRT than the LCD. I’d want to rule out the camera focus to be definitely sure it’s the shader but I think you might be on to something. I’ve got the original pictures zoomed in here so we can see in more detail than those high level ones:
Hmm I think some of this maybe over exposure. I’m going to go back and take a lower ISO shot to see what’s what.
According to the service manual of your PVM (2730QM you mentioned), the factory white point is 9300K (D93).
Online: https://manualmachine.com/sony/pvm2730qm/8174091-service-manual/
pdf:
Also I’m reading there’s a “HUE” setting on your PVM: two buttons “GRN” and “PUR” that sets skin tone color more greenish or purplish (button above left according to the manual?) that exclusively affects NTSC input signal. Might be something to look at when you’re using NTSC consoles on your PVM.
Good work on the shader, looks better and better with changes. Can you tell something how it is when actually playing side scrolling games etc. How does gaming for a prolonged time on your CRT compare to playing on your LCD with the shader, does the LCD+shader hold up in prolonged gaming sessions versus the CRT?
Amazing! Good catch! Right there in black and white on the first page! Oh well at least I’m proving my shader matches the real thing! I’ll come back to your other comments.