Jump to content
Curt Vendel

AMY Chip has been found...

Recommended Posts

wow i thought it would never emerge!

well done for tracking it down and respect to John Hardie too.

hopefully it can be brought to the community in some way in the future. its such a senseless loss that it was prevented by all that corporate wrangling

 

  • Like 1

Share this post


Link to post
Share on other sites

Who was the third party company that finished the chip? And why were they so adamant against licensing the chip to Atari?

Share this post


Link to post
Share on other sites
15 minutes ago, gozar said:

Who was the third party company that finished the chip? And why were they so adamant against licensing the chip to Atari?

I'm guessing it was either they incorporated other technology that made the licensing agreement impossible,

 

or they just wanted to extract more money

Share this post


Link to post
Share on other sites

Great news!   The samples sound like FM synthesis on a Sound Blaster.  There is no A8 software that takes advantage of it of course.  If it had been ready for the 600/800XLs that would have gained adoption on the platform.

Edited by Sugarland

Share this post


Link to post
Share on other sites

I studied the documentation and I'm quite disappointed. It's certainly better then AY. But I'd expect some modulations, filters .. this is too basic. It might all be just lack of my imagination, but imho there is a reason why synths didn't go this way. Pity they didn't make some sick demo.

  • Like 2

Share this post


Link to post
Share on other sites

Filtering would best be done externally.  From the short look I had at the docs, it only does the sound generation to the point of obtaining a 16 bit amplitude sample and the output is to be done by an external D->A converter.  Proper filtering could also be done externally.

I guess also given that we have just that one sample value that 2 chips would be needed to obtain stereo.

Share this post


Link to post
Share on other sites

I've not looked in depth but given that the biggest hurdle with SID is the filtering I'd say it'd probably emulate pretty easily.  Though with the number of voices and envelope generators maybe a bit resource-hungry.   But in the modern day emulation of 20th Century tech doesn't raise much of a sweat on anything above low-range gear.

Share this post


Link to post
Share on other sites
11 hours ago, Stephen said:

How easy do you think it would be to emulate (software) this?

Very easy. The math is simple. There is some trick to remove multiplication, but you don't really need it in emulation.

 

Let me first sum in short what Amy does:

- there are 8 voices, using sine wave. They are called fundamentals.

- there are 64 harmonics (also sine waves), which can be assigned to individual voices, and their frequencies are multiples of the fundamental frequency. The assigning is done in pairs, so the voice can have 2,4,6 .. up to 64 harmonics. Each harmonic can only be assigned to one voice, so if one voice has many, the rest of the voices has fewer.

- harmonics (and fundamentals) have simple volume envelopes, 1 target value / slope pair, the fundamental has same system for frequency.

- both volume and frequency slopes are correctly interpolated in exponential space.

- the harmonics have no other parameter, just the envelope and channel assignment. If channel has 4 harmonics, it's 2nd, 3rd, 4th, 5th. Their phase is computed directly as multiplication of the phase of the main voice.

- there are also 2 configurable noise generators, which can be somehow combined with the voices. This part is not very well described.

 

I'm not even interested in full chip emulation. I'm interested in what can be done with this approach, because I don't know any other synth doing it like this, and I'm rather skeptical.

I've played with DX7 simulator, you can do 4 harmonics with envelopes on it. And indeed, I had troubles getting interesting sounds out of it.

 

Yes, the samples we have sound "like FM" .. but that's just because Amy syntheses is based on sine wave, and so is FM. But in FM the other generators modulate the fundamental generator. They can have widely different frequencies, or just slightly. On Amy the harmonics are just added, there is no control over their phase, their frequencies are always just multiples of the fundamental. In other words, there is much less variation.

Also to get many popular synth sounds, you need lots of higher frequencies. On bass it means lots of harmonics. On Amy you also have to use all harmonics. You can't use 2nd, 4th, 8th and 16th .. which would be just 4. You have to use all 16, and mute those unused, limiting other voices. To make something as simple as PWM square wave bass would require voice with many harmonics which all would have to be carefully manipulated. Sid can do it with one poke. It's not as simple on FM, but it is still simpler then on Amy.

 

There is obviously great room for experimenting and different tricks, and it's IMHO easier to predict what it will do, compared to FM. It would be really interesting to see (I mean hear) what people might be able to do with it if they had 30 years, like they had on Pokey or Sid. But at the moment I don't see it. Way better then AY, that's for sure.

Edited by R0ger
  • Like 8

Share this post


Link to post
Share on other sites

Holy crap.

Even those two little samples gives some idea of what this chip could do compared with the other sound chips on the market at the time. Such a shame we never got to see them in the ST. Would have been far better than the crappy Yamaha that got thrown in.

Share this post


Link to post
Share on other sites
4 hours ago, R0ger said:

I'm interested in what can be done with this approach, because I don't know any other synth doing it like this, and I'm rather skeptical.

I've played with DX7 simulator, you can do 4 harmonics with envelopes on it. And indeed, I had troubles getting interesting sounds out of it.

You might look at formant speech synthesis. That's probably the closest to what AMY is doing here. You can use it for voices and melodic instruments pretty easy. The noise sources are for more fricative sounds.

  • Like 1

Share this post


Link to post
Share on other sites
5 hours ago, R0ger said:

 

I'm not even interested in full chip emulation. I'm interested in what can be done with this approach, because I don't know any other synth doing it like this, and I'm rather skeptical.

 

Kawai K5000S was the most popular synth that was based on this approach (additive synthesis). I'd say great for ambient stuff, bell-like sounds, pads... You can find some sound demos on YT.

 

Still... it's much easier to produce rich waveform and filter out some stuff to get something interesting than to add up sinewaves to get the same result. 

Share this post


Link to post
Share on other sites

K5000S has very nice sounds. But it's combination of additive / subtractive. And it seems to have way more voices.

I found video where guy praised additive synthesis greatly, but from today's standpoint, where you have unlimited memory a cpu power. You can take any sound, analyze it to 100 harmonics, and then recreate the sounds, while being able to modulate the process. Like for example having the harmonics differ for low and high notes. He even mentions AI approach. You take hours of recording of some instrument, like cello, which is very hard to reproduce well. You FFT, throw it into AI. AI can then recreate the sound from just notes, even adding expression and dynamic. For something like that, additive synthesis is obviously the best.

Could be also used for MP3 decompression :-D Not sure how the lack of phase control matter though.

Share this post


Link to post
Share on other sites
On 11/8/2019 at 5:57 PM, Sugarland said:

Great news!   The samples sound like FM synthesis on a Sound Blaster. 

You say that like it's a good thing! 😄

 

In the 80s, I suppose it would have been,  but by the time I got my SB16 in the 90s,  I thought the FM sounded horrible and dated.  I quickly got a wavetable card.

  • Like 3

Share this post


Link to post
Share on other sites

It was based of work done by "Bell Laboratories" so not too surprising... But still very exciting stuff. The Casio keyboards in the mid to late 80s used PCM and Phase Distortion technology for their CZ synthesizer and shelf keyboards, but the AMY seems like it was way ahead of its time. Someone would've figured out how to put an attack and decay the speech synthesis over time before Casio came out with their stuff. I contemplated a CZ emulator on a fast 8bit micro-controller, but the "AMY" is probably well suited for wave sound manipulation via speech synthesis. It's too bad Atari used the chip for anything; I know it wind up in the hands of another company who got it working where Atari sued them and scared the daylights out them. I hope the chip manifest like the "Pokey One" chip did; I'd love to take a stab at it with a sequencer... Maybe make a 5200 game and stick the sound chip on a cartridge perhaps. Or better yet put it on an Atari ST cartridge for Cubase sequencer.

Edited by philipj

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...