Once upon a midnight dreary, John Foust had spoken clearly:
Yes, there are/were many different formats for recording digital
data on audio cassettes. Some used frequency, some used phase
info, and some recorded to the tape using something other than
conventional voice-style recording. Think of them as modems
at speeds of 120 to 2400 baud.
For the CoCo & Model 100/102/200, they used (I think) either a
zero-crossing circuit or band-pass filter, with (this may be backwards) a
zero bit at 1200Hz and a 1 bit at 2400Hz. (overall, they advertised a 1500
baud info speed.) With a zero-crossing circuit, you just measure
microseconds between each zero cross, and you know exactly what digit you
have. Using the "double-speed poke" for the CoCo1/2/3 (The first
"overclocking" I did - ran a 1Mhz 6809 at 1.78 Mhz with that poke!) there
were also pokes that would change the timing code to accurately save/load
at the double-speed. Using that, I could save/load to tape faster than a
stock (i.e. no disk speed-up stuff) Commie 64 could save/load from disk!
I've found a number of decoders that were developed
for the
emulator scene. Most are quite crude and unforgiving.
The tape loader for the CoCo2/3 emulator from Jeff Vavasour (sp?) worked
very well. To use it, you hooked your tape player directly to the line-in
on a Sound-Blaster compatible sound board. But this could very easily be
the exception and not the rule, and I've not used it myself much... I
invested in a disk system early in my CoCoing career...
I concluded that I wanted to store absolutely
uncompressed
digitized audio until I confirmed that any of today's various
compression methods wouldn't obliterate the encoded data.
Given the dozens of encoders and compression schemes out there,
how do you know if a particular scheme won't wipe out the
data by simplifying waveforms, fudging phase relationships, etc.?
Which they prolly would. You'd be much better off zipping the files than
even thinking of attempting any MP3, RealAudio or whatnot. Those
modem-centric compressors are what's called "lossy" compressors - which
means they throw out information that they don't think is very neccessary
or info when gone may not detract from the audible (read: to the human ear
- not the computer) quality to music.
Running a tape waveform would prolly result in 1 of 2 things, both moot:
1) The sound info would be munged beyond repair - so what's the point;
2) The wave file would not be compressed at all - so what's the point...
Storage is cheap. You don't need stereo. If the
waveforms use
audio in the range 1200 to 2400 Hz, for example, then Nyquist
tells you to oversample by 2 to 8 times, meaning even 8 Khz,
8-bit might be overkill with roughly 8 K/sec storage,
and 22 Khz sampling is certainly adequate.
My vote is mono 22Khz 16-bit (tho 8 might be fine for some types of
computers - it would prolly depend on the type of encoding it used) and
then zip the file... you could prolly swing 1.25Meg per minute of sound
zipped (assuming 2:1 compression - it might be higher) which would give you
8.6 hours of audio stored per 650Meg CD... Not bad #'s IMHO.
====
Here's my question:
What about the ADAM computer from Coleco??? It uses a digital tape that
holds (around) 256K or so... (Never set mine up yet).
Is there any way you could run that thang thru an audio player and have a
PC routine re-digitalize it, or are you stuck with read a thing and
serial-send it over to another PC?
Always Questions... ;-)
Roger "Merch" Merchberger