Idea for something that could make CRTs truly obsolete

[atheistgod1999]

I got a cool idea that sounds good in my brain, but I'm not sure is possible:

Why not make a display with OLEDs or something that actually emulates a CRT? Like, have tons of wires connecting different OLEDs, and have something read and convert the sync (which would already be separated to H and V, like a real CRT) to signals a computer that sets up the virtual scanlines? Like, have the computer separate the rows into respective scanlines, with how many there in are in each depending on how many there are.

Then, for each virtual scanline, the thing would quickly open and close connections to small spots of OLEDs, from left to right, at a speed depending on the horizontal frequency. These connections would have the analog RGB signals flow to the OLEDs, lighting them up like phosphors.

If the sync can be digitized and the rows and horizontal refresh can be set up faster than the retrace, I don't see why this couldn't be done. This could work with fucking light gun games.

[Unseen]

Then, for each virtual scanline, the thing would quickly open and close connections to small spots of OLEDs, from left to right, at a speed depending on the horizontal frequency. These connections would have the analog RGB signals flow to the OLEDs, lighting them up like phosphors.

Phosphor light output slowly decays over time, an OLED turns off almost instantly - so that decay would need to be emulated to get a similar light output.

Second problem: Since your LEDs are only on for a short time compared to the duration of the entire frame and dark otherwise, the percieved screen brightness gets reduced by a factor of (on time / frame time).

(Edit: fixed ratio)

[Fudoh]

RetroUSB recently showed a 256x224 matrix display. Basically the same idea.....

https://www.facebook.com/retroUSB/posts ... 3064290761

https://www.youtube.com/watch?v=cYf1Me3K3nI

[blizzz]

You could make the individual pixels brighter and adjust the hold time, but wouldn't the whole idea just make the display flicker and be more expensive?

[Xyga]

Wasshoi 2012 (iirc) had a tate'd 10ft LED display, maybe even bigger it's hard to tell from the pictures.
Clear view for the audience outside and in plain day!

[atheistgod1999]

Then, for each virtual scanline, the thing would quickly open and close connections to small spots of OLEDs, from left to right, at a speed depending on the horizontal frequency. These connections would have the analog RGB signals flow to the OLEDs, lighting them up like phosphors.

Phosphor light output slowly decays over time, an OLED turns off almost instantly - so that decay would need to be emulated to get a similar light output.

Second problem: Since your LEDs are only on for a short time compared to the duration of the entire frame and dark otherwise, the percieved screen brightness gets reduced by a factor of (on time / off time).

Couldn't capacitors solve that problem?

[kamiboy]

I had a similar idea a while back. I think I even discussed it here briefly.

In any regard, emulation is not likely to make anything obsolete. Because at the end of the day it is just that, emulation, and not exactly like the real thing. Otherwise playing on actual retro hardware would have become obsolete a long, long time ago.

[atheistgod1999]

I had a similar idea a while back. I think I even discussed it here briefly.

What was it? Can you link me to where you discussed it?

[Taiyaki]

RetroUSB recently showed a 256x224 matrix display. Basically the same idea.....

https://www.facebook.com/retroUSB/posts ... 3064290761

https://www.youtube.com/watch?v=cYf1Me3K3nI

Interesting. Wonder what they plan to do with that. Maybe just for use on their sales stands?

[vol.2]

RetroUSB recently showed a 256x224 matrix display. Basically the same idea.....

https://www.facebook.com/retroUSB/posts ... 3064290761

https://www.youtube.com/watch?v=cYf1Me3K3nI

Interesting. Wonder what they plan to do with that. Maybe just for use on their sales stands?

RetroUSB is basically one guy who develops retro gaming stuff. I saw him on RetroRGB's podcast the other day, and it sounded like he's constantly tinkering and coming up with new stuff. Very cool. It's also possible that he's just made that wall for personal use. It would be a different sort of product for him to sell than is typically in his store.

Here's a link to the interview: https://www.youtube.com/watch?v=Yd0o0l4 ... -&index=10

Oh, I like the "wall of pixels" idea BTW. Though I also agree that it won't "replace" CRTs. I believe that there is an element of attachment that people have to specific equipment. However, if it looks just right, it would be interesting to have these as replacement panels for broken old TVs that have cool cabinets. Especially if they could be laminated to glass on a slight curve?

[kamiboy]

I had a similar idea a while back. I think I even discussed it here briefly.

What was it? Can you link me to where you discussed it?

It was years ago, can't find it now. Basically I envisioned a software solution that would emulate the electron gun scanning pattern, phosphor decay and shadow mask/grill sub pixel structure of a CRT on a 4K or higher res OLED capable of very high refresh rates. Hundreds of frames per second.

It was a pie in the sky discussion as the display necessary does not exist. Not to mention the rendering hardware.

Making such a thing is waste of effort anyway since as I said, nothing can do a better job of being a CRT than a CRT. Or equal for that matter.

[bobrocks95]

The biggest appeal to me RE CRTs is that they aren't fixed-pixel displays and don't require scaling. Which this wouldn't fix.

[Xyga]

Still, a 'low-res wall of LED/OLED' under a slightly hazy~smoked~curved glass coating would be nice.

[atheistgod1999]

Even CRTs technically have a resolution; the number of phosphors. This thing would have OLEDs as more like phosphors than pixels. I'd make some algorithm for it to set up which rows are what scanlines without it having to scale, and it scanning back to forth wouldn't be. The RGB signal wouldn't be processed in any way, and sent instead sent to the OLEDs that are connected at the appropriate time for when the beam would be scanning over the phosphors at that point on a real CRT. It wouldn't be low-res; I'd start off with a TVL count comparable to that of a mid-range consumer CRT, and something like 700 rows of physical phosphor rows. The thing wouldn't even detect horizontal resolution.

After the computer division, any non-integer scaling would be dealt with by combining rows where 2 scanlines meet, with the OLEDs on those rows being connected for multiple lines. Here are some examples:

Spoiler

5 lines and sent a signal of 2 scan lines:
1
1
1&2
2
2

3 lines:
1
1&2
2
2&3
3

4:
1
1&2
2&3
3&4
4

This would become a lot less of an issue with more rows. For example, if you did the 4 lines thing with 10 rows:
1
1
1&2
2
2
3
3
3&4
4
4

See where I'm going with this?

[vol.2]

Even CRTs technically have a resolution; the number of phosphors. This thing would have OLEDs as more like phosphors than pixels. I'd make some algorithm for it to set up which rows are what scanlines without it having to scale, and it scanning back to forth wouldn't be. The RGB signal wouldn't be processed in any way, and sent instead sent to the OLEDs that are connected at the appropriate time for when the beam would be scanning over the phosphors at that point on a real CRT. It wouldn't be low-res; I'd start off with a TVL count comparable to that of a mid-range consumer CRT, and something like 700 rows of physical phosphor rows. The thing wouldn't even detect horizontal resolution.

After the computer division, any non-integer scaling would be dealt with by combining rows where 2 scanlines meet, with the OLEDs on those rows being connected for multiple lines. Here are some examples:

Spoiler

5 lines and sent a signal of 2 scan lines:
1
1
1&2
2
2

3 lines:
1
1&2
2
2&3
3

4:
1
1&2
2&3
3&4
4

This would become a lot less of an issue with more rows. For example, if you did the 4 lines thing with 10 rows:
1
1
1&2
2
2
3
3
3&4
4
4

See where I'm going with this?

You have obviously put some thought into this, and your idea is intriguing, but I would conceptualize it differently.

Imagine the OLEDs as you have said, like the phosphors, but then you have the signal controlling them behave exactly the same way the electron gun would. You draw the lines on the screen in an idealized manner mimicking CRT operation. The "drawing" of the lines is just turning the OLEDS on and then off. Then you would need to have some kind of "persistence signal" that controls how long the "phosphors" stay lit for, and what "staying lit" means electronically. This way you have a hope for achieving some backwards compatibility with sources that were designed with the intent on reaching an electron gun.

[atheistgod1999]

You have obviously put some thought into this, and your idea is intriguing, but I would conceptualize it differently.

Imagine the OLEDs as you have said, like the phosphors, but then you have the signal controlling them behave exactly the same way the electron gun would. You draw the lines on the screen in an idealized manner mimicking CRT operation. The "drawing" of the lines is just turning the OLEDS on and then off. Then you would need to have some kind of "persistence signal" that controls how long the "phosphors" stay lit for, and what "staying lit" means electronically. This way you have a hope for achieving some backwards compatibility with sources that were designed with the intent on reaching an electron gun.

Yeah, that's basically what I was saying, except for the "staying lit" thing; I'd just use capacitors for that, and solid-state variable resistors to have control over how long the persistence is. I already said that they would be drawn in that way.

Here's a visual example with each phosphor triad having a number, representing when it gets the signal. It has 10 rows and 5 columns, with a 3-line signal being sent to it:
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1&6 2&7 3&8 4&9 5&10
6 7 8 9 10
6 7 8 9 10
6&11 7&12 8&13 9&14 10&15
11 12 13 14 15
11 12 13 14 15
11 12 13 14 15

If I could see you in person and draw out examples, this would be a lot easier to explain.

I'm wondering something about how sync works with CRTs, because I want it to be digitized where every detail of the analog signal has a corresponding digital signal; it would work so closely that it would naturally distinguish 240p from 480i, for instance. Anyway, do CRTs have preset vertical resolutions and frequencies, or can they be anything that's within the horizontal scanning frequency? Like, if I take a PC monitor intended for 480p 60Hz minimum, and feed it 240p 120Hz, will it display that? If I wire composite pre-separated to RGBHV from a PAL VCR straight to the pins on the chassis of a North American SD TV, will it display 576i 50Hz? Does (S)NES sync give the info that it's slightly faster than 60Hz?

What do you people think of my idea so far? I'd like replies. Just "It sounds good" or "I don't like it" will suffice.

[neorichieb1971]

Are you building a screen, something that plugs into the screen? Or a piece of software that runs on a PC that your running games through?

The issue is more to do with compatibility with different systems. Its all very well putting an idea out but to truly mimmick a CRT you need the resolution to fill the screen regardless of its native resolution.

And pots for HV and Stretch would be nice too. Its better to have an analogue control of things (if your talking hardware)

I had an idea a few weeks ago that would involve TV manufacturers to produce dynamic pixels that got larger or smaller depending on the native resolution of the signal.

[atheistgod1999]

Are you building a screen, something that plugs into the screen? Or a piece of software that runs on a PC that your running games through?

The issue is more to do with compatibility with different systems. Its all very well putting an idea out but to truly mimmick a CRT you need the resolution to fill the screen regardless of its native resolution.

I'm building a new screen. And horizontal resolutions would naturally be handled the same way as a real CRT. It wouldn't detect native horizontal resolution.

While it will use OLEDs or LEDs, the internal circuitry will be completely different.

[gray117]

The major thing for me with any of this is getting a 4:3 screen replacement for cabs. I fully expect that crt emulation will progress from acceptable to great eventually... but we'll need a panel size/ratio for our cabs.

[atheistgod1999]

Every (O)LED would only be on briefly; nothing would be stored. I'd use capacitors so it "decays" instead of instantly going dark after the diodes are disconnected from the RGB lines. It seems like people don't get that. The raw RGB signal would be displayed without being processed. Do you people even know how an RGB signal sends the info? The electron gun paints the intensities the signal is sending for each color at that very instant.

[bobrocks95]

Every (O)LED would only be on briefly; nothing would be stored. I'd use capacitors so it "decays" instead of instantly going dark after the diodes are disconnected from the RGB lines. It seems like people don't get that. The raw RGB signal would be displayed without being processed.

As voltage drops off with a capacitor, doesn't the color of an LED change first, rather than just its luminosity?

EDIT: Nevermind, I guess OLEDs are still an array of RGB diodes.

[atheistgod1999]

Every (O)LED would only be on briefly; nothing would be stored. I'd use capacitors so it "decays" instead of instantly going dark after the diodes are disconnected from the RGB lines. It seems like people don't get that. The raw RGB signal would be displayed without being processed.

As voltage drops off with a capacitor, doesn't the color of an LED change first, rather than just its luminosity?

Which is why I'd prefer to use OLEDs if possible. Is there any way to get them individually, like LED bulbs?

Can someone answer my questions about sync?

[Guspaz]

Not really, but you're talking about millions of OLEDs in a single display here, so discrete elements would be impossible anyhow. Ultimately what you're going to need to do is take an existing OLED panel and design a custom controller for it that simulates a CRT raster.

If you want to talk about simulating a 600TVL display using OLEDs, and you're going to assume that each TV line is represented by one pixel with square pixels, that's an 800x600 display, or more specifically a 2400x600 subpixel display. That's 1.44 million subpixels.

[atheistgod1999]

Is there any way my idea will work? The way it would do sync stuff would be based on what a CRT would do; if CRTs don't detect the horizontal frequency before scanning, then it won't do it and instead do it however CRTs do it. Someone told me they worked like that, which is why I said that. It wouldn't start drawing the whole frame; it would wait for new line pulses or whatever is used when the sync is fully separated to HV. Is there any OLED panel where I can interface with the pins or whatever on the OLEDs, so I can run analog signals into them? If not, how expensive would it be to make one that does? I wouldn't start this until I'm 18, anyway.

[bobrocks95]

As with many other ideas throughout history, it seems to work in theory but will be nearly impossible in practice, unless maybe you've got a few hundred grand lying around.

[Guspaz]

You're not going to be directly sending voltage to OLED pixels, no... You're probably going to want a degree in electrical engineering, a large sum of money, and a few years of spare time before you approach this.

[atheistgod1999]

You're not going to be directly sending voltage to OLED pixels, no... You're probably going to want a degree in electrical engineering, a large sum of money, and a few years of spare time before you approach this.

Alright. I'll do that.

[BazookaBen]

Alright. I'll do that.

Well you're super young, so if anybody can do it on this board, you can. Shoot for the stars mang.

[Ed Oscuro]

The only good things about CRTs, in comparison to alternate technologies, are that it's relatively resolution agnostic (not framerate agnostic; phosphor decay times seem to limit how even performance is at extreme framerates - and also in traditional CRTs, going even to 31KHz requires boosting voltages, and 4K resolution is probably completely impractical) and a built-in anti-blur method - which personally I think should be an optional feature, as it triggers headaches in some people.

Unseen mentions most of the OLED issue, but there's a bit more to it: Driving OLEDs at high brightness without wearing them out has been a long-standing problem, and what I read cites electron mobility and attack of materials by free radicals as the major causes. So if you are having to drive OLEDs much brighter than they are now, you will have to look at materials breakdown and you will have to look at external quantum efficiency. I don't know what exact relationships hold between voltage and lifespan, or continuous stimulation vs. flashes and lifespan, but it seems that short flashes require far higher currents, leading to more heat generation, and more free radicals. I don't think switching off the OLED goes far to make up this problem.

The main meta point I'd like to make is that retrogaming devices and even the basic classic TV formats weren't designed with forward compatbility in mind - planned obsolescence from the start! If we get improvements in technology that will make this happen, people will no doubt work to apply them to the problem and we need people to continue to advocate for this. But be cautious of saying to yourself "I want to make a CRT replacement, so I will learn about television technology (or circuit design, etc.)." It's like saying "I love video games so I will learn programming." Most people only have success going forward if they really are passionate about the discipline and the journey that will ultimately lead to something, and if you don't find that discipline rewarding you won't stick with it. Thankfully there are a lot of things involved in display design, including the consumer feedback like I mentioned.

[Royal]

I apologize for the wall of text, but what about Laser Projection?

http://forums.blurbusters.com/viewtopic.php?f=2&t=1156 (i'd highly recommend reading the first 7 pages)

Great Colors, Infinite Contrast, no fixed pixels, zero motion blur, AND zero lag. It operates by modulating several laser diodes (at minimum 1 red, 1 green, and 1 blue) to a "mixer" of sorts before striking a MEMS mirror that displays the image in whatever scanning frequency and pattern the manufacturer chooses.

A company named Microvision already made a 854x480 front projecting pocket projector based on this technology years ago. They outsourced some of their technology to companies like Sony, who now makes the Sony MPCL1, which is 1920x720. If I was suddenly put in the shoes of one of their engineers, I'd suggest putting a version of this with 480 scanning lines and all analog circuitry in the back of a monitor, being capable of both 480i and 480p together (something you can ONLY achieve with select PVM/BVMs...literally no consumer CRT offers this without improper digital conversion and lag despite both resolutions having the same dot pitch).

There are some issues that need to be resolved though; the immediate one being laser speckle that is noticeable in bright scenes, but the guy (Light23) on the linked BlurBusters forum claims to have found a way to eliminate it by using a "holographic filter".

The other issue is that lasers have nearly zero persistance, which is responsible for zero display motion blur. Great, but on the other hand it means that 60 hz and other low refresh rates will produce very noticeable flickering. Increasing the refresh rate to 72-90 hz will easily fix this, but that will produce tearing that comes from mismatching framerate and refresh rate. Increasing it to 120 hz will be a perfect fit, but frame doubling (60fps@120hz) on a low persistence display is far more noticeable than display technologies like LCD (which seems to mask intense frame doubling like 30fps@60hz with motion blur and display blur).
http://forums.blurbusters.com/viewtopic.php?f=2&t=342

The reason this matters is because every device many of us WANTS to use with a CRT replacement is limited to a 60 hz output and nothing higher (and many games old and new aren't even 60 fps).

This means that a crt replacement using lasers need phosphors to increase persistence, reducing flickering at 60 hz while still having less display blur than your average LCD or OLED (compare 16 ms to 4 ms in this chart: http://www.blurbusters.com/persistence-vs-motion-blur/). But another problem is that there isn't any phosphor I've heard of that is perfectly clear and stimulates to laser light of any color.

However, there ARE phosphors that respond to ultraviolet lasers as this company demonstrates:
http://www.sid.org/Portals/sid/Files/pd ... Hajjar.pdf
Unfortunately, it is also a digital device that is used as part of giant video walls for different companies.

There are also miscellaneous issues that can arrive if any of us want the display to have a convex curved screen, in which MEMS may not be the best way to scan laser light across a specialized screen such as that (another guy on the blurbusters forum mentioned using voltage to bend laser light to whatever direction you want: http://forums.blurbusters.com/viewtopic ... 6&start=80).

Something like this should have its own thread, but I like OP's idea in theory, as it has several advantages in comparison to the laser projection idea but may be significantly harder to implement. I'd certainly use a raster scanning OLED (or better yet, self emissive quantum dots) for multisync with higher resolutions, as I don't have to deal with potentially decreasing quality by increasing dot pitch size or requiring an unrealistic amount of "depth space" for the rear projection. But for a 480i/p 4:3 curved CRT replacement, the right lasers and phosphors can probably outperform OLEDS in brightness for the purpose of raster scanning. Lasers are not only perceived brighter than other technologies at the same rating, but remains at the same quality and brightness up until the point it dies (Light23 claims 20,000 hours of lifetime).

I'm currently studying Electrical Engineering right now at a university. Hopefully, I will learn enough about circuits and optics to make something like this myself. It's not like any Microvision or Prysm or even Sony is going to make either technology for (in their eyes) a selective and picky demographic clinging to old obsolete consoles and games instead of "moving on". We have to make shit for us ourselves.