[Coco] Re: Banking ROMS (was-6309 microprocessor project:11-01-2003)

[Roger Taylor]

At 02:21 PM 11/2/2003 +0100, you wrote:
> > Now, dern it... how did you create the "We Gotcha!" vocals in your Mega-Bug
> > game?!  :)  I hacked that game when I was about 17 years old and discovered
> > a sort of compressed format, where the values were delays between speaker
> > clicks, I think.  Anyway, I've always wanted to know that.
>
>I guess it's kind of 1-Bit PCM sampled using the CoCo's "Zero Crossing
>Detector" (cassette port). Recording is done saving the (loop-) times
>between "bit changes" of the cassette port. Playback is done outputting
>the "bit changes" to the CoCo's "One Bit Sound" and waiting the recorded
>(loop-) times again (kind of sequencer). Steve must have been an expert
>of this stuff, remember his great "Spectrum Audio Analizer" Pak...

My guess back then was the same.  Through the cassette port, I would think 
the zero-crossing signal is the only way.  Counting delays between signals 
would definately sample some frequencies.

I guess the loudest frequency gets the priority since everything else is 
riding ontop of it, so it crosses the zero-point more.  I hope I said that 
right.

And I have written spectrum analyzers that really weren't calibrated or 
used any scale, but did show the "power" of each "frequency" seen by the 
cassette audio in line.  When there was silence or noise, the high end 
always blasted through the roof usually on the 2 or 3 count.  That is, how 
many times a delay of 2 was seen during a certain timeframe of samples, vs. 
how many times 1 or 3 was seen during the same timeframe.