11

Two of the more famous and highly regarded sound chips in the 8-bit era were the Commodore 64's SID and the Nintendo APU. Which of the two was overall better, seems to be a matter on which there is a perennial difference of opinion, but this caught my eye on https://www.reddit.com/r/fpgagaming/comments/gxol18/mister_nes_core_apu_rewrite_by_kitrinx/

While the C64's SID chip was arguably better - the NES had 5 voices to Commodore's 3 but the SID was programmable while the NES APU was not - though the NES had limited ability to play sampled audio - the NES was less expensive and purpose built for games and it showed.

Now I'm curious: what could be meant about the NES being purpose built for games, in the context of sound hardware? The SID was good at music, but games in general have music, so... on the face of it, the two machines would seem to have the same requirements for sound generation capability.

What could the SID do that the APU couldn't? Was there some non-game kind of sound the former was particularly good at?

Improve this question
2
  • 3
    I saw this question title right after browsing the aviation SE. My immediate response was "What in the world does a Standard Instrument Departure have to do with an Auxiliary Power Unit?" :)
    – Wayne Conrad
    Aug 8 at 20:33
  • 1
    @WayneConrad You're not the only one. I went WTF? as well :-)
    – Tonny
    Aug 9 at 7:53

2 Answers 2

Reset to default
18

I don’t know that this adds much beyond the information already in the question, but…

A perennial problem with the SID is playing music at the same time as generating sound effects, due to the fact that it only has three phase accumulators, and some effects bind two of them together, giving you only two independent noises. It’s very common for games not to have in-game music or very obviously to multiplex.

The NES conversely has five independent channels, making it more suitable for maintaining some sort of music while also throwing in occasional sound effects.

In the other direction, the SID offers higher-resolution waveforms, higher-precision frequency selection, ring modulation, and analogue high, low and bandpass filters. The APU is much more of its era, limited mostly to fixed digital waveforms; it has a single channel which can be programmed with a very small waveform (e.g. used for drums in Super Mario 3) but is not otherwise particularly tonally complex.

Improve this answer
5
  • 5
    I would concur but mention that the variety of sounds the SID can produce is potentially higher (due to its filter features) than the APU. Considering that the SID's features probably aren't fully explored fully even today you could, however, argue "the APU was good enough" :)
    – tofro
    Aug 7 at 12:39
  • 1
    Indeed; check out examples like Dynamic Range for sounds you wouldn't expect any other contemporary system to replicate.
    – Tommy
    Aug 7 at 14:31
  • 2
    @tofro: One feature/quirk of the SID is that the filter uses analog circuitry that isn't particularly well calibrated, and varied substantially in different manufacturing runs. Thus, a piece of music may sound really great on one machine, and somewhat "meh" on another.
    – supercat
    Aug 8 at 22:17
  • 4
    @supercat That dwelves a bit into the differences in silicon itself: The APU was a pure digital chip, while the SID was mixed-mode digital-analog technology which was actually pretty uncommon for sound chips and is much closer to the way how analogue synthesizers work.
    – tofro
    Aug 9 at 5:49
  • @tofro: I can't think of any other "fully integrated" sound chips that combined analog electronics with digital control the way the SID does. Analog synths used a high degree of integration, but I think the C64's approach was unique. One thing I'm curious about with the SID design is whether it could have been done much more cheaply by multiplexing one set of circuitry between voices, and another set of circuitry between digital FIR or IIR filter stages.
    – supercat
    Aug 9 at 15:01
10

The NES had two voices that were designed to produced pulse-wave tones with adjustable volume, one fixed-volume triangle wave output, an adjustable-volume noise channel, and a digital audio output channel that was generally only used when nothing else was happening.

The C64 has three voices which can be independently configured for triangle, pulse wave, sawtooth, or noise generation, and offer independently adjustable volume. The chip can also feed any desired combination through a programmable low/high/band/notch filter. The filter makes it possible for the C64 to produce a range of sounds that were impossible to match on many machines, and difficult to match even on machines that used sampled audio, but the fact that there's only one of them makes it somewhat difficult to use in some kinds of games.

Improve this answer

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .