POKEY vs. SID

[JamesD]

Ok, I never tried to program one so I can't tell if we're discussing different characteristics, but based on the register set and brief documentation I read, it seems to have different envelopes and volume settings alright, but not so sure about waveforms or pulse width. Even the POKEY is said to have variable duty cycle on the square wave, which I think is baked into the noise register settings so you can get a different pulse width or white noise, but yet again that is something I've only read about.

 

Waveform to me is a timbre, like the SID that has triangle /\/\/\, sawtooth |\|\|\, variable pulse width |_|___|__, combinations of those and white noise. Envelope is a function of how volume changes over time, and usually contains an attack phase (fade in), a decay, a sustain level and a release level. Some more advanced synthesizers may have multiple attack, decay and sustain levels for the same note, and the envelopes on the AY seem to repeat an ADSR cycle multiple times without retriggering the note manually.

 

Btw, as part of the so-called Viznut Waveforms, a couple of years ago it was realized that you can synchronize the internal timers in the VIC-I chip too, and get kind of different pulse widths on the VIC-20. In the old days, people would notice that suddenly the sound got a different timbre and then back again without knowing exactly why or how it could be used, but this behavior was tamed and I think even VICE emulates it reasonably these days. There may be internal timers in other sound chips too that you can sync and cause different pulse widths, I don't know.

Ok, I thought it had square waves and several sawtooth waves (|\, /|, /\) but that is probably some other chip I looked at.

The envelope can repeat or be "one shot" if I remember right.

 

The example I posted is a modern creation which takes advantage of newer trackers I'm sure. Older stuff doesn't sound as good. There are literally hours of older AY tunes on youtube you can check out.

[JamesD]

FWIW, a lot of arcade games had custom sound hardware in addition to off the shelf chips like the AY or POKEY, so don't assume all arcade sounds were from those chips.

[JamesD]

Just so you can compare the AY to other chips

For some reason that doesn't want to start at the beginning of the first song. It starts around the 3 minute mark.

[Nebulon]

I really like both chips. Yes the SID is better.

 

However, it's pretty amazing that POKEY can hold its own in a match with the SID when you consider that:

 

1) the SID chip was designed long after the POKEY was released.

2) the SID designer was an experienced analog synthesizer programmer.

[am1933]

POKEY for that authentic gaming experience.

[JamesD]

I really like both chips. Yes the SID is better.

 

However, it's pretty amazing that POKEY can hold its own in a match with the SID when you consider that:

 

1) the SID chip was designed long after the POKEY was released.

2) the SID designer was an experienced analog synthesizer programmer.

By long after you mean 3 years.

[emkay]

By long after you mean 3 years.

Three years in that time meant ages in computer history ...

 

Also... the 08/15 edit of "Thing of a Spring" got the attention by Youtube and the Copyright thingy...

 

Did I mention that SID "lovers" have a biased taste in music? Well, at the end "taste in music" is always relative

[+OLD CS1]

That is interesting about being able to set the VIC generator timing. I ran into a problem with the 9919 (PSG) in some songs when the same tone from two channels would obviously cancel each other out. Turns out the generators start in unpredictable states and cannot be manipulated. Would be cool to discover they could be manipulated.

[carlsson]

This is Viznut's writeup for the VIC-20. I'm not sure if this can be adapted to the 9919/76489 or other chips, but it could be worth investigating.
http://www.pelulamu.net/pwp/vic20/waveforms.txt

[CatPix]

I am unbiased to both, as I didn't grew up with neither systems; nor with any system from either brand.

I tend to prefer the POKEY. Sure, SID get slightly better at doing music, and some tunes are amazing. But the thing is, we're talking about video game sound chips. So, the music has to blend into the game. And for that, I found POKEY to be better, with his distinctive, more "beepy" sound. When you hear a POKEY tune, it says "welcome to the world of 8 bits chiptunes" much better than SID. But not by much heh. And it's just how I feel about them.

[Mayhem]

By long after you mean 3 years.

 

Not even that long in the scheme of things probably, Bob Yannes started work on the SID in around Jan or Feb of 1981. When was POKEY first created/used, in the Atari 400?

 

As for the capabilities, limitations and costs "restricted" the SID a bit, it was going to have more features if you read some of the background into the chip.

[carlsson]

It seems a number of people are credited as the inventors behind the POKEY chip. Wikipedia mentions two people, Doug Neubauer mentions himself and additional three persons. Perhaps the two mentioned in Wikipedia were heading the development crew and thus got the official credits?

 

In any case, it appears Atari started to work on the Colleen during the summer of 1977 and that the team had a working chip within nine months, which should be spring 1978. The actual computers Atari 400/800 were presented at Winter CES in Jan 1979 and shipped in Nov the same year.

 

http://dougneubauer.com/atari/

http://www.atarihq.com/othersec/library/neubauer.html

http://www.atarimuseum.com/articles/joedecuir.html

 

So if we take July 1977 - January 1981, that means POKEY actually is 3.5 years older than SID.

[JamesD]

So what you are saying is that Atari had a 3.5 year head start to improve the POKEY and they didn't.

[Nebulon]

So what you are saying is that Atari had a 3.5 year head start to improve the POKEY and they didn't.

 

Hmmm.... Interesting way to look at it, although not at all what we're saying.

 

By that logic though, one could say that Commodore had almost a decade to improve the audio in the Amiga, but didn't. However, Atari did with the Falcon.

[carlsson]

I know both chips have additional functions than just generating sound, but one could also argue that the market for digital synthesizers was not that big in 1977 that anyone had imagined a need to make even more synthesizer oriented chips for game computers.

 

However it appers that Yannes' team completed the SID in five months during 1981, while the POKEY team took nine months in 1977. Whether the faster development was due to improved processes, gained experience or other factors, I can't determine.

[Osgeld]

I am unbiased to both, as I didn't grew up with neither systems; nor with any system from either brand.

I tend to prefer the POKEY. Sure, SID get slightly better at doing music, and some tunes are amazing. But the thing is, we're talking about video game sound chips. So, the music has to blend into the game. And for that, I found POKEY to be better, with his distinctive, more "beepy" sound. When you hear a POKEY tune, it says "welcome to the world of 8 bits chiptunes" much better than SID. But not by much heh. And it's just how I feel about them.

 

I didnt really grow up with either as well, and what I played on C64 wasnt anything good enough to have a soundtrack, but I disagree with you on saying were talking about video game sound chips, no we are talking about computer sound chips

 

yes in the real world they most likely were experienced with only video games by the majority of people, but they had more serious applications, like scoring music, or testing hearing or generating waveforms for electronics work, which the pokey falls flat with its off tune sour off beat frequencies.

 

pokey isnt bad, in fact I love its sound for what it is, SID on the other hand is a much more rounded design

[JamesD]

Hmmm.... Interesting way to look at it, although not at all what we're saying.

 

By that logic though, one could say that Commodore had almost a decade to improve the audio in the Amiga, but didn't. However, Atari did with the Falcon.

http://www.thule.no/haynie/research/a3000p/docs/a3000p.pdf

[Nebulon]

http://www.thule.no/haynie/research/a3000p/docs/a3000p.pdf

I'm not sure it counts if it was never actually released.

[+OLD CS1]

So what you are saying is that Atari had a 3.5 year head start to improve the POKEY and they didn't.

 

 

Not what I read into that, at all. I figure the POKEY did exactly what they wanted it to do. Yannes wanted a synthesizer on a chip, so the SID did exactly (well mostly) what he wanted it to do.

[JamesD]

I think C64 games often fell short on sound effects and focused on music which may give people the impression it's not very good at sound effects.
Here are some SID sound effects. They are very different than the POKEY for sure, but it's clearly pretty capable.

[JamesD]

I'm not sure it counts if it was never actually released.

You didn't say released in your original post, you said:

 

By that logic though, one could say that Commodore had almost a decade to improve the audio in the Amiga, but didn't. However, Atari did with the Falcon.

[Nebulon]

You didn't say released in your original post, you said:

 

Even if you did count the Amiga 3000+, it's audio is still 8-bit.

 

Although.... interestingly, despite Paula remaining 8-bit, it looks like a 16-bit CODEC is supported for the DSP. I suppose it also depends on whether 16-bit DACs and I/O ports are included with the build or more hardware is required.

 

Regardless, I still feel that things should be kept to realm of hardware that was actually mass-produced and available to the consumer. After all, both Commodore and Atari had a ton of unreleased prototypes.

Edited January 22, 2016 by Nebulon

[carlsson]

Personally I wonder on which of these chips development began first, and which one was soonest finished:

 

POKEY - Assumed development started in the summer of 1977, finished in 9 months (spring 1978) and demoed in the Atari 400/800 in January 1979. It also strikes me that development of the next generation gaming hardware was initiated before the VCS had been launched to customers, but perhaps constant development is a given.

 

AY-3-891x - Apparently GI had the AY-3-8900 STIC ready by January 1978, but not sure about the 891x sound chip. The Intellivision though is said to have been developed during 1978 and test marketed in late 1979 in California.

 

TMS9919 - It was pre-dated by the short lived SN76477 which only featured one channel of sound that could be either a sine wave, a square wave or noise, although with envelope control. This chip was found in Space Invaders from June 1978, but was soon replaced with the SN76489/TMS9919 with better capacities to match the already popular (?) AY-3-8910. Another time marker is the ALF Music Card MC16 for the Apple ][ that was demonstrated in late 1978 and for sale in early 1979. It is said that the 9919 pretty much is a single chip version of the MC16. Finally the TI-99/4 was demoed in June 1979, so by then the sound chip for sure must've been ready.

 

For further dating, there is a resource on the TMS9918 VDP that is rather vague about dating, but it mentions the hardware sprites were inspired by Atari, which I assume would put development no sooner than September 1977, unless TI engineers got a sneak peak. There is a timing document on the VDP from 1978, which matches the dating for the companion sound chip.

 

Thus, it is a good possibility all three of those sound chips were worked on simultaneously. Perhaps GI actually predated Atari a little on this, which would suggest this order of chips being ready:

 

AY-3-891x

POKEY

TMS9919

(three years later)

SID

[CatPix]

 

I didnt really grow up with either as well, and what I played on C64 wasnt anything good enough to have a soundtrack, but I disagree with you on saying were talking about video game sound chips, no we are talking about computer sound chips

 

yes in the real world they most likely were experienced with only video games by the majority of people, but they had more serious applications, like scoring music, or testing hearing or generating waveforms for electronics work, which the pokey falls flat with its off tune sour off beat frequencies.

 

pokey isnt bad, in fact I love its sound for what it is, SID on the other hand is a much more rounded design

 

I worded my answer wrongly indeed... I meant that given we're on a forum about video gaming, and that, I personnaly care about sound chip mostly in a gaming or gaming related used (chiptune) I think that POKEY is better IMO for that role.

Probably indeed SID can do better in other aeras, and that can even be measured, but as far as how chiptune sounds, it's just a personnal feeling, not a measurable thing.

[Osgeld]

fair enough