vocoder's frequencies values?

Kasper's icon

hi

I am working on a more/less vocoder idea - for doing so I had a look at the "classic-vocoder" max exemple patch, but also at different other vocoders and have one question :

_How is made the choice of the values (frequencies) of the bands used in a vocoder? is there a rule here or those (the 15 ones in the exemple) are used because it sounds good?

thanks

kasper

Roman Thilenius's icon

third-octave or harmonic series work well, but what you also should try
is to _limit the filterbank to a certain range, noone needs frequencies below
50 Hz for speech-to-strings.

why dont you make it variable?

Kasper's icon

by third-octave you mean +/- 3 notes by octave, right ? but should I repeat the same notes, or rather twist it (kind of C E G# C E G# etc or C F B E A etc)
harmonic series make sense

of course i am making it variable - just wondered what was the reasons for the choices made on the hardware stuff

many thanks

kasper

Roman Thilenius's icon

not notes, just a third of an octave, like in the common "31-band equalizer"...
63, 80, 100, 125, 160, 200, 250, ...

the fixed filterbanks i know are all linear, but i dont know much vocoder hardware
when i think about it.