home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!noc.near.net!hri.com!ukma!cs.widener.edu!dsinc!ub!toz!cyberman
- From: cyberman@toz.buffalo.ny.us (Cyberman)
- Newsgroups: sci.electronics
- Subject: Re: Color Computer 3 inte
- Message-ID: <gate.38NuVB1w165w@toz.buffalo.ny.us>
- Date: 16 Dec 92 00:51:13 GMT
- Lines: 56
- X-Maildoor: WaflineMail 1.00r
-
-
- Sir Richardson spoke forth saying:
- "Re: Color Computer 3 inte" toward
- on 12-14-92 17:12
- MR> On a similar vein, about 3 months ago I tried to get a digitized
- MR> sound driver working under level II, using the FIRQ. The idea was to
- MR> use the empty space in $FExx and vector the FIRQ there, bypassing the
- MR> OS9 stuff (which was sloooow, and would crash if used!). I got to the
- MR> point where I had a simple square wave generator running in ~$10
- MR> bytes, and which worked.... for a bit. It would always crash after a
- MR> while, tho, as it seemed the IRQ was interfereing somehow. So I
- MR> finally dropped it. Too bad, a background sound driver would have
- MR> come in real handy.
-
- Where does MicroWare hide the kernal? If I remember there is
- a task switch map. (T0/T1). this allows the computer to
- rapidly switch to the kernal and then return to the program.
- The question is how do they implement it?
-
- SYSTEM MEMORY MAP:
- RAM 00000 - 7FFFF (512K bytes)
- ROM 78000 - 7FEFF when enabled
- I/O XFF00 - XFFFF I/O space and GIME regs
-
- 64K PROCESS MAP:
- RAM 0000 - FEFF (possible vector page FEXX)
- I/O FF00 - FFFF (appears in all pages)
-
- now acording to this FExx is NOT empty.
-
- Note: the Vector Page RAM at 7FE00 - 7FEFF (when enabled), will appear instead
- of the RAM or ROM at XFE00 - XFEFF. (see FF90 Bit 3)
-
- And this states the Vector page RAM is there when enabled.
- That may have more to do with your problem. I was considering
- constructing a MOD player for the COCO (that would be cool).
- of course the big limitation is memory (as ussual). Another is
- adjusting the interupt rate (60hz on the CC][ the CC]I[ has a
- variable one).
-
- To solve the memory problem I figure I can chop the lower 4
- bits on each sample byte. Then one can have 2x the sample space.
- This wouldn't make a difference since the D/A has only 6 bits of
- resolution anyway! (add 4 4 bit samples the result is 1 6 bit
- value). It may not sound great but it might be a first for the
- Coco (who cares?). It might be able to only work on the CC3
- because it may be the only coco FAST enough to do the digital
- mixing.
-
- Stephen Cyberman@Toz.Buffalo.NY.US
- Mangled on Tue 12-15-1992 at 17:51:19
-
- ... Click ... click ... click ... damn, out of taglines!
- ---
- * Blue Wave/QWK v2.11 *
-
-