Stephanie Wallace
Member
I, for one, am quite excited about your project… and you never know where it might lead if you catch the bug to do more after getting some audio out of the chips!
"Using" an Atari Pokey Chip with a Teensy 4 or 4.1?
- Thread starter jonathan322
- Start date
I, for one, am quite excited about your project… and you never know where it might lead if you catch the bug to do more after getting some audio out of the chips!
jonathan322
Well-known member
Good point. But, like you said, these pokeys are getting harder, and more expensive, to acquire. Although, it does seem like some in the Atari community are trying to develop pokey replacements, and the like. PokeyMAX (1, 2, and 3), PokeyONE, etc.
jonathan322
Well-known member
@Stephanie-Wallace , do you think I can use these (or similar) with the audio shield? At least to get started?
https://www.amazon.com/gp/product/B07FTB281F/
Yes, there are replacements. But if you use a teensy 4.x you actually don't even need one of these. We are using a Teensy 4.1 to emulate old arcade vector games including the sound chips. A teensy has more than enough raw calculation power to emulate multiple of these audio chips at once and generate a I2S signal for an external audio DAC.
Stephanie Wallace
Member
You certainly can, but the output isn’t going to be—nor is it meant to be—particularly desirable. The speakers will be quiet and tinny, and as testing the quality of your circuit and the resulting output is probably going to be one of your first priorities, I’d suggest sending the digitized audio to your development machine if you have halfway decent speakers either via AudioOutUSB from the Teensy or line-out from the audio shield to a digital input.
Unless you have no higher-quality speakers/monitors that accepts digital input, I wouldn’t waste the money. Sure, they’re ridiculously inexpensive, but you’re probably not going to get much future use out of them either.
Plus… the sheer pleasure of listening to output from these ancient ICs on modern hardware is something exceptionally special! I can’t even begin to convey the unique joy of listening to a slew of YM2149s, quirks and all, on studio monitors.
Stephanie Wallace
Member
I can only speak for myself—though I think @jonathan322 was largely pointing out that efforts are being made to preserve these circuits from total extinction—but there are damn good reasons I’m using the real deal rather than YM2149 “emulators” for my project, the largest of which is the desire to teach an old dog new tricks, something that can’t be done with a simulator, with authenticity, nostalgia, and the challenge of doing so coming in tied at a very close second. In fact, a specific product that uses a simulator rather than real hardware played a small part in pushing my desire to create this synthesizer over the edge.
A large part of the issue is precisely that so-called emulators of old 8 and 16 bit era audio chips that we still haven’t managed to fully reverse engineer are just that: So-called emulators. I used quotation marks above because they are not emulators at all. Emulation is the attempt to perfectly recreate hardware—and if applicable, software (such as a boot ROM) required for it to operate—in a virtual environment.
Simulation, on the other hand, is the recreation of how one interacts with a piece of hardware, and outputting something very similar to the results of doing so.
As an example, in the case of the YM2149, which is basically just a square wave generator with a bunch of shift registers and few PLL clocks along with a pseudo-random noise generator and a few DACS, not a single YM2149 “emulator” works this way. These simulators simply accept the values one would send to the chip, then return the audio the chip would return, in the best case scenario, based on measurements of the chips voltage output or PCM output. None of the code actually replicates the manner in which the circuitry works. Interestingly, the YM2149 is basically a state machine, and a large portion of its circuitry could be recreated using FlexIO, much like accurately implementing circuits using an FPGA.
Even the best emulators of legacy computers, gaming consoles, and arcade machines don’t actually emulate their respective audio chips.
To return to the overarching thesis: One cannot use a simulator to develop new means of programming/controlling a piece of hardware with any guarantee such techniques would actually work on the real thing, which is precisely the goal of projects such as the one I’ve referred to in this thread.
jonathan322
Well-known member
My thinking was that I don't want to destroy any of my good equipment on the first go-around - I really have no idea what is going to come out of these old chips. These little speakers are... disposable
And they can hook right up to those outputs.If they survive, then I can assume it's safe to move on?
As to your other post, on emulation and simulation, you're exactly right - I want to see what I can make the old hardware do. At least the FPGAs I mentioned recreate the circuitry in question and not, as you have said, the outcome of said circuitry.
Last edited:
Stephanie Wallace
Member
Though based upon the nature of the project and our conversation thus far, I can't envision you doing anything that would destroy your speakers... if you're worried about it, then yes, absolutely play it safe and use some throwaway speakers.
Precisely. I almost added "reproduction" as a third category, but it was a little too late at night to include an additional topic.
While FPGAs are't a perfect solution... they're the best solution we have for the preservation of a large a number of items that will soon become either extinct or extraordinarily rare artifacts.
The AY-3-8910, for instance, is the audio chip originally produced by General Instruments in 1978 for which the schematics were then licensed to Yamaha, who released a slightly modified version, the YM2149, in 1982 (along with a long line of chips that incorporated the PSG for many years following its release); yet, despite the existence of an excellent shot of the AY-3-8910's die, which has been converted into a faux photograph of a circuit both by hand and neural network, there are still aspects of how it works nobody has managed to figure out to this day. Not to mention, even in the case of perfect reproduction, there's much to be said for working with a tangible piece of history.
We have finally reached a point at which a full analysis of the such chips is possible, but it's extraordinarily expensive, and vintage computing enthusiasts can't begin to fund such an endeavor, while companies that specialize in state-of-the-art reverse engineering have no incentive to undertake it. Perhaps the necessary technology will eventually become inexpensive enough for a small group of enthusiasts to do so before a number of items are gone forever, and perhaps it won't.
I'm glad, for now, we at least have a means of reproducing very close approximations of the original versions of these chips.
Precisely. I almost added "reproduction" as a third category, but it was a little too late at night to include an additional topic.
While FPGAs are't a perfect solution... they're the best solution we have for the preservation of a large a number of items that will soon become either extinct or extraordinarily rare artifacts.
The AY-3-8910, for instance, is the audio chip originally produced by General Instruments in 1978 for which the schematics were then licensed to Yamaha, who released a slightly modified version, the YM2149, in 1982 (along with a long line of chips that incorporated the PSG for many years following its release); yet, despite the existence of an excellent shot of the AY-3-8910's die, which has been converted into a faux photograph of a circuit both by hand and neural network, there are still aspects of how it works nobody has managed to figure out to this day. Not to mention, even in the case of perfect reproduction, there's much to be said for working with a tangible piece of history.
We have finally reached a point at which a full analysis of the such chips is possible, but it's extraordinarily expensive, and vintage computing enthusiasts can't begin to fund such an endeavor, while companies that specialize in state-of-the-art reverse engineering have no incentive to undertake it. Perhaps the necessary technology will eventually become inexpensive enough for a small group of enthusiasts to do so before a number of items are gone forever, and perhaps it won't.
I'm glad, for now, we at least have a means of reproducing very close approximations of the original versions of these chips.
jonathan322
Well-known member
A "faux-tograph"! (snicker). Couldn't resist.
Where there is a will, there is a way. Here, in the current environment, where there is a severe chip shortage, it is nearly impossible. Yet, if the market reaction to the shortage ends up being an over-reaction, resulting in an abundance of inexpensive chips, in the long run, and if the demand from enthusiasts for replacements/reproductions is such that said replacements/reproductions can fetch a high enough price (without impacting demand), then such a project would be feasible and perhaps even profitable.
jonathan322
Well-known member
How’s progress? JS