The 7800 Minnie In-Cart Sound Chip

[RevEng]

As most of you are likely aware, TIA wasn't the first choice for 7800 audio. Originally the 7800 Maria chip was supposed to include advanced audio within it, but that was removed when die-space was needed for other functionality. At that time GCC pivoted to in-cart audio, coupled with TIA audio.

 

Over the years there's been some speculation as to what GCC's proposed in-cart sound chip solution would have been like, and how cost-effective it may or may-not have been to have an in-cart chip. To help nail this speculation down, I reached out to Steve Golson. (GCC chip-designer extraordinaire, and friend of the community) Steve kindly searched through his archives for all of the docs he could find, and shared them with me. (and now you)

 

When I received these docs, I was astonished to read about Minnie. This thing was no pokey-redux. Here we have a chip capable of basic WaveTable synthesis while still keeping things lean and cost-effective.

 

I've written a 7800 Minnie Sound Chip article over at the 7800.8bitdev.org wiki, which also contains the original GCC docs. Steve is still seeing if there's anything else he can put his hands on, so the article may be updated as time goes on. If it is, I'll post an update here.

 

Enjoy, and feel free to ask questions here, or discuss what may have been.
 

[Trebor]

This is phenomenal and incredible!  Thank you for sharing this information with the community and putting up a devoted section to it under 7800.8bitdev.org.  We're well beyond speculation and just chit-chat at some board meeting, this is really a thoroughly fleshed out design.  Such a shame it never saw the light of day back in the 7800's heyday. 

[RevEng]

8 minutes ago, Trebor said:

Such a shame it never saw the light of day back in the 7800's heyday. 

Honestly, reading through the specs made my heart ache more than a little, for what could have been. 

[+Muddyfunster]

Thanks for sharing this Mike.

[+DrVenkman]

35 minutes ago, RevEng said:

Honestly, reading through the specs made my heart ache more than a little, for what could have been. 

Yeah. But reading carefully can also bring a hearty chuckle.

 

[Trebor]

4 minutes ago, DrVenkman said:

Yeah. But reading carefully can also bring a hearty chuckle.

 

For those having a hard time reading the section...

 

"...it had damn well better not dissipate more than 500 mW or we're all gonna die..."

[+DrVenkman]

2 hours ago, Trebor said:

For those having a hard time reading the section...

 

"...it had damn well better not dissipate more than 500 mW or we're all gonna die..."

Well, even before that part, it gives a maximum airspeed over ground of "23 furlongs per fortnight," and an operating environment roughly equivalent to operating 100 feet underwater, lol.

 

EDIT: Shame on me, I did the estimation wrong in my head. Ten atmospheres is about 335 feet deep at standard density of water. Sea water is denser, so it would be shallower. ? 

Edited March 5, 2021 by DrVenkman
Engineering nerdery

[+sramirez2008]

3 minutes ago, DrVenkman said:

Well, even before that part, it gives a maximum airspeed over ground of "23 furlongs per fortnight," and an operating environment roughly equivalent to operating 100 feet underwater, lol.

Further proof that Atari employed quite a few characters.  Thanks for sharing Mike.?

[fiddlepaddle]

Snickering quietly to herself, the marketing secretary inserts test language to prove that management doesn't read the middle of paragraphs.

[SlidellMan]

Things sure are looking hopeful for the 7800.

[ZylonBane]

Quote

2 arbitrary 64-byte waveforms stored in on-chip ROM, which could differ per-game

 

Err, did you mean on-chip RAM here?

[+batari]

Everyone knows that one furlong per fortnight is basically 1 cm per minute. That's a pretty damn slow airspeed if you ask me.

 

That said, this looks like a job for HOKEY! hehe.

[RevEng]

2 hours ago, ZylonBane said:

Err, did you mean on-chip RAM here?

Nope, the plan was ROM. It's an in-cart chip, and the idea was they could have customised it per-title.

[CPUWIZ]

24 minutes ago, RevEng said:

Nope, the plan was ROM. It's an in-cart chip, and the idea was they could have customised it per-title.

 

Ignore him, he is pedantic, you said "on-chip ROM".

[ZylonBane]

37 minutes ago, RevEng said:

Nope, the plan was ROM. It's an in-cart chip, and the idea was they could have customised it per-title.

You should emphasize this point then, because the concept of a sound chip that's customized on a per-game basis is utterly bizarre.

 

Also, ignore CPUWIZ. He is 17 corn muffins and a paperclip.

[RevEng]

13 hours ago, batari said:

That said, this looks like a job for HOKEY! hehe.

This was also part of my motive for getting the Minnie information from Steve. It would be a good alternative to pokey, and something unique to the 7800 scene.

 

There's certainly enough information in the docs to do a sound-alike clone. Probably worth it to deviate from the original design and do the samples in updateable RAM though, since 64 bytes of RAM per sample won't affect our cost, like it would have with the original chip.

 

[BydoEmpire]

This would have been amazing, thanks for sharing!

[+Pat Brady]

18 hours ago, RevEng said:

Honestly, reading through the specs made my heart ache more than a little, for what could have been. 

That happens when I read almost anything about Atari after they ran Bushnell off. Just a massive squandering of potential.

 

Quote

Similar to other contemporary sound chips, Minnie relied on the CPU to adjust volume for achieving sound envelopes.

I'll quibble with this. Contemporary sound chips were SID (on the market in 1982), 2A03 (1983), and YM2151 (1984), all of which did envelopes in hardware. 2A03 envelopes were limited but better than nothing.

 

To be clear, leaving it to software was probably a smart concession for a chip with an extreme cost limit.

 

Anyway, thanks for sharing these documents. It strikes me as a clever design with at least a dash of wishful thinking: the second document contains a jarring statement about needing to outsource design of the analog section, and I believe the per-game ROM customization was not feasible at all. But even if they had chosen 2 useful waveforms to be shared by all games, this would have been a very nice option for 7800 games. For that matter, even if it only provided the 3 "free" waveforms it would have been a vast upgrade from TIA alone.

 

EDIT: I *love* the idea of programming this into HOKEY.

[RevEng]

48 minutes ago, Pat Brady said:

That happens when I read almost anything about Atari after they ran Bushnell off. Just a massive squandering of potential.

 

I'll quibble with this. Contemporary sound chips were SID (on the market in 1982), 2A03 (1983), and YM2151 (1984), all of which did envelopes in hardware. 2A03 envelopes were limited but better than nothing.

Fair point. Wiki updated.

 

48 minutes ago, Pat Brady said:

 It strikes me as a clever design with at least a dash of wishful thinking: the second document contains a jarring statement about needing to outsource design of the analog section, and I believe the per-game ROM customization was not feasible at all. But even if they had chosen 2 useful waveforms to be shared by all games, this would have been a very nice option for 7800 games. For that matter, even if it only provided the 3 "free" waveforms it would have been a vast upgrade from TIA alone.

Steve reported the design was complete, and that they had prototype chips manufactured and working. I think the analog section outsource mention in that doc was old news. They found the expertise somewhere.

 

When I read about the rom-based waveforms, I was also thinking that it missed out on economies of scale. Thing is, I think GCC would have been aware of that at the time, having just finished Maria2, so I personally trust that they would know if it was a feasible concept or not. Either way I do think they would have likely had a common Minnie config (or two) and only customised it per-game when the game actually demanded it, and was expected to sell in large numbers.

 

[+Pat Brady]

19 minutes ago, RevEng said:

Steve reported the design was complete, and that they had prototype chips manufactured and working. I think the analog section outsource mention in that doc was old news. They found the expertise somewhere.

 

When I read about the rom-based waveforms, I was also thinking that it missed out on economies of scale. Thing is, I think GCC would have been aware of that at the time, having just finished Maria2, so I personally trust that they would know if it was a feasible concept or not. Either way I do think they would have likely had a common Minnie config (or two) and only customised it per-game when the game actually demanded it, and was expected to sell in large numbers.

 

If they had a working prototype, that's definitely significant.

 

As for custom ROMs, I noticed that somebody at GCC crossed out "easily" from "easily redefined on a per-game basis."

 

"Feasible" was not the right word on my part. It was doable technically. I have to remember, the GCC people did not know that Warner was about to sell Atari to Tramiel, that the 7800 would not see wide release until 1986, that by then Nintendo would dominate the industry. In 1983 they may have projected a 7800 game could sell 10 million. In that context a game-specific mask might have been not unthinkable, if custom samples would have helped sales.

 

But if we're going to emulate this thing, might as well pretend it's RAM, like you suggested.

[Trebor]

3 hours ago, Pat Brady said:

If they had a working prototype, that's definitely significant.

"...prototype chips were produced, and a "Minnie Rag" demo sang out within the walls at GCC"

 

Shame we don't have one of those prototype chips or a sound clip.

Still, the documents + HOKEY and we may be in for something wonderful down the road.

[+batari]

21 hours ago, Trebor said:

"...prototype chips were produced, and a "Minnie Rag" demo sang out within the walls at GCC"

 

Shame we don't have one of those prototype chips or a sound clip.

Still, the documents + HOKEY and we may be in for something wonderful down the road.

A prototype would be great, but with enough specs I am sure I can get it going with HOKEY hardware. It's a pretty straightforward design.

 

This chip uses a 32 byte address space and the registers are basically write-only. Because the POKEY chip can only address 16 bytes, the registers that control this chip should also be tweaked to fit within that address space (if this is to interface with current hardware.) I think it can be done, but some changes will have to be made to the register layout. One way is to maintain the register layout as described by GCC and bankswitch it. Though, that isn't necessary.

 

As for the 64-byte sample registers, I can see a mask ROM being done as these would be for mass-produced games and so it's OK that each chip would be unique to the game. But for our purposes, it would have to be RAM, of course. I vote for creating three serial registers for setting these up, stored as a queue, in addition to the registers the chip currently has.

[R_Leo_1]

I can only imagine having a homebrew game that uses the Minnie sound capabilities. That would be amazing to see!

[Jinks]

1 hour ago, R_Leo_1 said:

I can only imagine having a homebrew game that uses the Minnie sound capabilities. That would be amazing to see!

The way things are going, it looks like that day will come! 

 

[Lynxpro]

Did they try selling the chip to Tramiel's Atari Corp during - or after - the negotiations for the 7800 and its original launch titles which finally wrapped up in August 1985? Maybe Leonard Tramiel might remember for Corp's side of things. He certainly was impressed with the AMY and lamented how that project didn't work out [he's mentioned that a few times over in the Atari Museum Facebook group in the past].