Ha! That would be me! ;-)
Ah! The man behind the mask! ;-)
I didn't realise his contact was both a DSP _and_ Apple ][ guru!
I thought it was admirable that he was willing to work to
get it right.
Yeah, Colin did a great job. He explained the basics of the process,
much like you just did. Eventually I bought a DSP book and read it. I
found it very difficult to apply that knowledge to realtime DSP on a
general purpose CPU (i.e. no DSP hardware) and 1MHz sample rate!
Machines are plenty fast enough now to do this easily, but a decade
ago, it would have been an issue.
Indeed. My first emulator was written over 10 years ago in assembly for
an 8MHz ARM 3, and crude resampling was the best I could manage, though
it did process cycle accurate 1MHz speaker data. In later C++ versions
I wanted to improve emulation in other areas ...
Anyway, do you think Colin's approach was pretty much optimal?
If there's no better way, then I'll just use what you guys cooked up.
(My aim is to write a reasonably accurate 'reference implementation' of
an emulator that will be open source, and - get this - in Java. I grow
weary of the plethora of emulators that are closed source, single
platform or just plain hard to use for other purposes. Exactly what I
mean by "other purposes" will have to wait ...)