[sdiy] Anyone built the ELEKTOR VOCODER?

[Theo]

Here is a idea for a different approach.
Have a 30 band spectrum analyzer on the  modulator input.
And use only 3 tunable filters on the reconstruction/carrier side.

When the reconstruction filters are tuned in real-time to the vowels that
the spectrum analyzer detects,
it should be possible to get good understandable speech form only 3 filters.

This construction should allow for some additional tricks,
like detuning/offset  the reconstruction filters e.g. to a frequency range
that best fits the carrier signal.
This would be close to patching the analyzing bands to the "wrong"
reconstruction bands,
only now the "patch" can be under CV control.

Just a wild idea really.
Theo


From: Rick Jansen <rja at euronet.nl>

>
> > > Also, can someone address this question for me: 30 channels at 1/3
> > > octave sounds very impressive, but is it necessary? I would think you
> > > could have closer spaced channels (1/3 octave) in the frequency range
of
> > > the human voice plus some harmonic multiples, and have them 1/2 octave
> > > spaced outside of that.
>
> Aha, here surfaces an old project I still intend to do, and already bought
> chips for, 6 years ago or so: National Semiconductor has (had now, I
think)
> the LMF380, a switched capacitor filter chip that you can cascade to build
> 1/3 octave filterbanks, like for spectrum analyzers. Ech chip has 3 1/3
octave
> filters.
>
> My plan was (is) to at least build a filter section with it and see if
> vocoder-like effects are possible, or else turn it into a spectrum
analyzer.
> (The LEDs necessary for a hardware spectrum analyzer add up to quite an
amount
> actually!)
>
> For switched capacitor filters you need to do anti-aliasing filtering if
> you plan to use the filtered signals for audio after filtering, else
you'll
> get a lot of digital noise.
>
> Cheers,
>
> Rick Jansen
> --
> Web: http://www.euronet.nl/~rja/
> email: rja at euronet.nl