Re: [stella] video music

Subject: Re: [stella] video music
From: Mark De Smet <desmet@xxxxxxxxxxxx>
Date: Sat, 24 Oct 1998 15:57:23 -0500 (CDT)
> >The audio input port of the Supercharger only returns a single bit (whether
> >the voltage is above or below a threshold. Good the reading an audio tape
> >but horrible for audio digitization.

> That's why I was going to go with using the pot lines for an A/D converter.
> But alas, my friend who has the magazine with the schematics to build
> this device is now living in Japan (and won't return for 3 years).

This would not be a hard circuit to construct.  You would probably want a
simple op-amp to buffer(amplify, and shift) your audio input from the
console. You can then run the opamp output through a resistor into the
dumped input.  You would use the dumped input in the same manner as when
reading the paddles.  The only programming difference would be possibly
changing the time constants. You would configure the circuit simply by
adjusting the op-amps gain, and the value of the resistor to change the
time constant of how long it takes to charge the 0.068uF capacitor in the

Here's the catch: the input audio voltage changes vary rapidly, so that
you would actually be measuring the average(not actually the average, but
close anyway) voltage input over your sampling interval(you are not
actually measuring the voltage, but instead the capacitive time constant).
So if you manage to read from the port(dump the input, and wait for a 1)
1000 times/second, a 1khz sine wave would read the same as no sound.  The
same would apply for most sound above 1khz.  You would be able to do some
reasonable interpretation for frequencies below 500hz.  I am not
completely sure of these numbers because I don't know that much about
sampling, and how many samples are required for reasonable representation,
but I think these will serve as adequate approximations.  This method
requires extrememly fast reading of the dumped input ports.  I've never
done anything with them, so I don't know if it is possible to read that

Is the input latched?  What I mean is when the input port goes high, does
it stay high if the capacitor is discharged extermally, or will it go low
when ever the voltage on the capacitor is below the input's threshold?  If
it is latched, you may be able to do some sort of peak sampling.

An alternative possibility.  Use a series of comparators so that each one
will go high with different audio voltage levels, and run these into any
of the joystick input ports.  This would allow detection of up to 8
different audio levels(one per port).  You would still have to read the
ports fast, but I would imagine the read itself would be much faster.  
The circuit for this would require a lot of hardware, and still may not
function vary well.  This effectivly would provide a set of inputs the
same as the supercharger idea would, but at several different levels.

If you add a set at the outputs of the comparators, you would be able to
do the peak sampling, but you would have to somehow clear the latches.

Did you consider using a 4 bit ADC and running it off of the i/o ports?

Mark De Smet

Archives (includes files) at
Unsub & more at

Current Thread