Rebuilding today's emulators - to perfection!

[S1500]

You are right. A lot of the interfaces & overall UX for many emulators, well, aren't that good. MAME & MESS are fun, but to set things up is a huge hassle at times.

[Schizophretard]

If it would be a CRT emulator that you basically "plug" other emulators into then couldn't it be used for anything standard definition and not just classic video games?

[Keatah]

Keatah, on 15 Nov 2013 - 5:06 PM, said:

I had a thread discussing this a while back. It failed to be serious. What I (still) think needs to be done is have a CRT module, much like we have Z-80 and 6502 cores available for emulators.

 

This "module" if you will, that I am describing, is a standalone bit of code that simulates to the best of its ability all the characteristics of an analog CRT assembly and the low-quality video signal coming into it. This would include simulating, in-depth, the following:

 

1- the standard faire of analog circuitry in a real TV chassis and assembly, the amplifiers & filters

2- flyback voltage and acceleration

3- horizontal and vertical amplifiers and drivers

4- electron gun operation

5- beam sweep

6- vertical freqencies of 30Hz to 250Hz with 0.1Hz granulation

7- color subcarrier of 3.57MHz

8- horizontal freqencies of 24kHz to perhaps 500kHz, again with 0.1Hz resolution

9- variable video bandwidth, beginning at 4.5MHz ranging to some big number like 50MHz or something

10- power supply variances

11- the various CRT geometry & picture adjustments we all know and love, like H.hold V.hold, contrast, color, tint, saturation, hue, convergence, pincusion, keystone, brightness, black level, gamma, tone, H.size, V.size, linearity, sharpness, bleeding, artifacting, fringing, resolution, color separation. All these things, and more!

12- a comprehensive examination and study of a selection of perhaps 50 different aperture grilles and phosphor masks

13- the same comprehensive exam of the phosphor chemistry itself, emission rates, phosphor persistence, glow, spill, color tone.

14- the nuances of how and where the input signal degrades and is "rebuilt" or amplified and reinforced

15- magnetic field interaction in the CRT tube, especially with the deflection coils and the point where the electrons leave the gun

16- the timings of the beam movement, the sweep, the retrace, or in the case of vector monitors the x,y movements and accelerations and delays

17- the interactions of the tube as a whole with its accompanying driving circuitry

18- as a catch all, consider the analog noise, hysteresis, instability, variances, all the things that must happen to a signal coming down the wire into the set.

19- noise and voltage levels, all the imperfections, crosstalk, spillovers, harmonics, signal isolation. How the phosphor reflects in the glass, and blooms, and smears and shimmers. The natural background noise introduced into the analog drivers of the beam. A/D, D/A, quantization noise, modulation, demodulation, frequency shifting, randomness..

20- basically the entire bag of analog video imperfections that are part and parcel of the composite/RF/NTSC system.

21- ability to accept incoming incoming video from whatever hardware is in the pc such as usb converters and digitizers or old-school vivo chips.

22- ability to accept digital data from emulators.

 

It's either all that or keep bellyaching over the dwindling supply of real CRTs.

 

This is delving way off topic. You can emulate all that to hell and back but we still need to fabricate a real-time display that works with light guns. A Plasma like display with custom drivers might work. Basically what we need to do is create a display that "follows" the beam. V-hold and H-hold are not necessary with a digital controller. First we create a 480 line ED display capable of producing strobing light that follows the beam. The phosphors will optionally be be laid out in a rectangular grid which will mimic the design of retro TV displays, rather than the square pixels of modern display tech. Secondly, there will be no overscan. The position of the 480 lines of display is predefined according to NTSC standards. The right hand and left hand edge of the display area can be adjusted using controls, to optionally display every pixel available to the video game console without cropping the image or leaving side bars. Various custom presets can be set for different game consoles. The most important aspect of the display is that the "beam" is scanned in real time using a digital controller. Each of the red, green, and blue phosphors has underneath it a tiny chamber filled with inert discharge gas. A pulse of current causes a discharge in every plasma cell which illuminates the phosphor. There is no "beam" per say, but the precise amount of discharge is determined by the instantaneous amplitude of the video signal. Interlaced or 240p video can be determined by counting the exact number of scan lines per frame. 240p video can be represented by doubling the scanlines for true 60Hz display which will prevent the venetian blinds effect often seen on HDTVs when 30Hz flicker is applied to a sprite. An optional 100% / XX% effect can be used to emulate retro scanlines for 240p sources. 480i interlaced video will be displayed on alternate scanlines as per NTSC specification. This custom plasma display will work with light gun games and also naturally display phosphor trails like a real CRT, but have a sharpness and clarity unparalleled by CRT displays. It should also include an NTSC RF tuner with support for both US and Japanese frequencies, as well as composite, S-video, component, and possibly HDMI (480i/p modes only). The HDMI standard still has the required timings for displays that scan the beam so it should work in 480 real-time display mode. The set will be strictly 4x3 aspect ratio since it is primarily meant for being used with retro game and/or video players. PAL models could also be produced which have SCART connectors and 576 scanlines. The PAL models will also able to play NTSC/PAL 60Hz formatted video signals in real time using a letterbox mode.

 

Now we just need some heavy venture capitol, crowd funding, or whatever else is necessary to produce these unique plasma displays that specifically cater to the analog video / retro video game enthusiast.

 

BTW, the concept probably wouldn't be viable since the custom sets will probably cost a minimum ten thousand USD a piece in quantity, when gamers can just go to the thrift shop and pick up quality used CRT sets for like $30. However, it such a display could be fabricated, it would be awesome!

[Keatah]

I could only guess, but I think that there would be a worldwide audience of maybe 5k-10k retrogamers that would spend the thousands upon thousands of dollars needed to have Sony or Samsung create a display as you describe. And even that is generous in both price and quantity. IDK.. My back of the napkin calculations show almost 75k per display. That's big buck for an average gamer.

 

What I do see happening is refresh rates and resolutions continuing to increase in low-end consumer displays. Sony's already got a 4k screen going. And 8k isn't far behind. This is more than enough resolution to depict a CRT glow. Turn down the lights and crank up the contrast & brightness.

 

I could see, cheaply done too, a tiny camera that you clip on to your NES lightgun. In this camera is a processor that looks exactly at where you are pointing the gun. It also looks at the big picture playfield, and black boarders. From these two frames it can calculate the where the target is in relation to the image. And with knowledge of the original setup (crt tv, and game console, and resolution + timing) this little processor could say yeah or nay when you pull the trigger. OR, it could send a time compensated signal when you pull the trigger. Thus behaving like the old-school cds-photocell or photo-transistor.

 

In the case of doing lightguns with emulators, there could be an off playfield counter or one of those funky square barcodes. Generated by the emulator itself. This would be a frame and beam position analyzer. Or the emulator could, as soon as you pull the trigger, change something subtle in the target that the camera can see, but you cannot.

 

Or just do a modded version of what modern head tracking units do. Build it into a tiny module that clips to the old-school lightgun.

 

Lots of possibilities if you put your head to it!

[PFL]

Cathode looks sooo close to whats needed. Hardware is all fine and good but if it's hardware you want you can just get a CRT telly.

 

Surely, with the technical know how in the emulating world there must be enough knowlegde out there to fool old style lightguns into reading something of an LCD or LED display. Like it or not, these displays ain't going anywhere. What Keatah (and myself) suggests seems to be the most economic path.

 

I don't quite get why using a 'virtual' cable as stardust puts it would be a nightmare. Is using any plug-in for any emu a nightmare? Not really. I suppose people would need to agree on a standard (kind of like how Kega Fusion's plugins have a standard and can thus be used in Virtual Boy Advance too) but surely that's about it?

 

Also, this whole "well we all have multiple cores so Stella can fix it" doesn't make Stephen's job any easier. He would still have to code phospor trail etc just like EVERY OTHER EMU CODER out there.

 

I really believe that without this kind of software telly emulator it will ultimately hurt the emulation of every old console. Afterall, CRT isn't coming back and modern TV will keep evolving further and further away from the old standard output devices of countless machines. Sad but true, IMHO.

[Keatah]

Keyword is looks. Cathode looks like a real CRT. But if you look a little closer you'll see that it's on the level of Microsoft Arcade. A simulation, a re-write look-alike; when compared against an emulator of today. I would say there are tons of tiny bitmaps going on rather than mathematical calculations to arrive at what the raster should look like.

 

In other news:

I used a serial port emulator to connect two instances of AppleWin together so I could pretend I was calling my old BBS

[Andromeda Stardust]

You know, this is something where an ultra-HD display would come in handy. When the pixels are so tiny that they are almost invisible, it becomes possible to do stuff like a CRT matrix in 4k resolution where you can get real close to the screen and see the rectangular honeycomb grid of phosphors on screen.

 

Another thing too while most displays still use LCD tech which uses subtractive light mixing, an additive light source such as LED array has the potential to yeild a much better display tech with rich black levels comparable to old school CRTs. Problem is LED tech has not yet been miniaturized enough and is still cost prohibitive except on a giant scale, ie "jumbotrons", billboards, etc. Currently each LED has to be soldered in place as there is no way to print them into a matrix like phosphors or liquid crystals. Given time LED tech will mature and backlit LCD displays will eventually take a back seat to LED matrix display. I use the term "matrix" to differentiate it from LCD displays that are merely backlit by LED tech.

 

One potential issue here with regards to emulating a CRT is that all these post-processing effects have the potential to add lots of lag to the display. That's why I'm thinking phosphor trails would be simple to do in post-processing without adding much lag. Every pixel has a decay signal based on experimental data. That data can be offloaded to a spare CPU thread and summed with the brightness of each frame buffer as it outputs to the display. The summation can be done almost instantly since modern CPU cores can perform billions of calculations per second. A separate thread would be used for special effects from the emulation thread, preventing slowdown. A true "phosphor trails" effect blending many frames together in such a manner would provide a much better overall "persistence of vision" effect than the current Stella hack which simply blends adjacent frames, but although the current implementation doesn't look exactly correct, it does facilitate the production of 30Hz video for Youtube upload without mass deletion of up to 50% of the sprite/playfield data. Adjustment sliders in software could be used to control the strength and persistence of the trails effect.

[raindog]

As I suggested on the other thread, it'd be a hell of a lot easier to just add Wiimote support to emulators (and stick a couple IR LEDs on top of your monitor, or candles on top of or just below your TV) than to produce a flat-panel display that would work with lightguns.

 

<oldman>And in 1973, we played light gun games on hardware so primitive you could point the gun at a lightbulb and win every time, and we liked it!</oldman>

[Andromeda Stardust]

Wiimote light guns, pbbbt. Needs cross-hairs to work, LOL!

[raindog]

Nah. One-time calibration so the emulator knows where the sensor bar is relative to the screen, and then it's no less imprecise than the NES Zapper. It's "Wii Motion Plus" that needs crosshairs due to the drift.