Vocoder Idea

[Martin Czech]

> I've just started looking at vocoder schematics and flowcharts and had
> an idea for a slightly different arrangment:  instead of using a
> separate bandbass filter for each channel of the "voice" signal, use a
> single VC-bandpass that goes into the normal rectification and LP.  This
> output feeds a separate VCA for each filtered channel of the input
> signal.  The trick would be to have the values of each VCA controlled by
> a sample-and-hold circuit.  The SH for a given channel would be
> triggered when the VC-bandpass's freq corresponds to the freq of the
> given input channel.  So the VC-bandpass will perform a frequency sweep
> of the "voice" signal, which is sampled by each VCA at the time
> corresponding to the freq of that VCA matching the freq of the bandpass.
> 
> One problem might the rate at which you could perform this sweep -- will
> it be limited by the cutoff of the LP envelope filter?  Maybe the LP
> should be voltage-controlled, too.  Then the sweep could move at a
> constant number of octaves/second.  However, what might be really cool
> is a slow sweep or a triggered single-pass sweep.  Speech might not be
> understandable, but you could effectively store a filter based on
> "voice" patterns over time.  Or maybe even random triggering?
> 

I have the impression that this channel saving trick won't work.
The analysis filter needs some time to measure the incomming
spectrum correctly, and this time is inverse proportional to the
filter frequency (Fourier Theory). This would mean that the filter
has to stay very long at the lowest frequencys in order to run in,
so the overall analysis response will be very slow.
Say fmin=40Hz ->Tcycle = 25ms, if we need 3 cycles to run in this makes
75 ms just for the lowest channel. Adding time for the other channels
will lead to s&h update intervals of > 0.2s, this seems
to me as a very slow response of the overall system.

Sweeping bandpass filters are used (or better fixed bandpass and
sweeping modulation) in spectrum analysers, but these machines are designed
for giving the average spectrum of a more or less steady state process.

m.c.

------------------------------------------------------------------------
m.c. (alias mad scientist) has made it finally: 3 CDs à 72 minimum ready.
Visit my homepage at http://www.geocities.com/SoHo/Museum/4459/index.html