[Coco] Including an SSC [no, this is not about including it into the SDC...]
Torsten Dittel
OS-9 at TRS-80.CC
Thu Oct 16 10:17:22 EDT 2014
Simon Jonassen <simon at ...> writes:
> It's *ALMOST* as good as the legendary SID (MOS 6581)
[You can name everything else but a C=64-fanboy, but the SID works
technically somewhat different and is thus a class on its own but that's
not what I wanted to comment here]
> What it *would* do in this case is provide an easy way to produce wound
on the coco withhout the overhead that we have today.....
I hope my CoCo wouldn't sustain any severe injuries... :-P
I just remembered a thread I started almost 12 years ago on the old CoCo
list: I had noticed the AY-scene (from ZX Spectrum and Atari ST) and their
zillion of song files in kind of standard formats. The files for the
Speccy contained sections of Z80 codes, so not very suitable for simple
replay on the CoCo, while the ST files would work if reformatted and
packed differently to support the CoCos smaller RAM (original files didn't
have the register data interlaced but in 14 blocks for better LZH
suppression). I came up with a file analysis and some proposals and Robert
Gault was experimenting with BASIC09 code to process and replay a file
basically dumping 14 register contents 50/60 times per second to the pack
(my proposal was to do that in ML and IRQ driven). I remember darkly that
he came to the conclusion that the 7040 PIC was accepting and/or
processing the register load commands very slow (he was updating the 14
single registers in direct mode- my proposal was to use the "register
string" command in direct mode and update only the registers which had
actually changed). It looked like the CoCo would spend a lot of time
(0.5ms which is an eternaty in CPU cycles) waiting for the SSC's status
bit to report "ready" until it could write the next command to the PIC. It
was never clear, where (and why) the bottleneck in the chain CPU->PIC-
>Sound-Chip exactly was, but it was not really what I had expected, so the
original idea of having background music without waste of CPU cylces
(beside one could have maybe used faster IRQs: FIRQ on CoCo2/3 or GIME
timer programmed IRQs on the CoCo3 for polling the SSC status bit).
With other words, the SSC seems to have a very poor hardware design and is
not really the holy grail one should spend a alot of time in. I find the
SID SW emulation (v2) on the CoCo (or the MOD player) much more impressive
if it is about demonstrating the capabilities, but still no background
music then...
(My 0.02€)
Regards, Torsten
More information about the Coco
mailing list