[sdiy] 40-band vocorder

[Aaron Lanterman]

On Mar 31, 2009, at 11:44 PM, Kyle Stephens wrote:

> Meh, still learning these things, and with much help from your  
> videos Aaron

Thanks for the kind words!

BTW, I plan to (assuming all goes as planned) teach the course again  
this coming Spring.

> - I'm gonna cut your department a check sometime soon :)

And please do:

http://users.ece.gatech.edu/~lanterma/ece4803/donations.html

> But back on topic, vocoders and number of bands - is this a case of  
> more being better or diminishing returns? Juergen Haible mentioned a  
> vocoder project some time ago and said he'd keep the stage count  
> down to keep the analog character.

I'm guessing to a large extent this just has to do with the expense  
and difficulty of building and calibrating a bunch of tight and well- 
matched filters. If you look at something like the 10-band Buchla  
filter, or the 296 Programmable Spectral Processor, there are massive  
amounts of trimpots all over the place. I can't imagine what a PITA it  
must be to adjust all those trimmers.

In terms of "analog character" - for the aforementioned reasons,  
analog vocoders don't have a whole lot of channels, and that gives a  
particular sound.

- Aaron