[sdiy] Roland TB-303 filter responses graphed

[Richie Burnett]

> kudos for doing this! :)

No worries.  They're old results that I decided to share :-)

> I don't know the 303 in detail, how does the output level
> of the VCO respond to the frequency change? Does it stay constant?

The sawtooth output from the VCO is a textbook perfect sawtooth over the 
full musical range, but the "square" waveform of the TB-303 is anything but. 
It's pulse width and shape vary greatly from the low 10's of Hz up to nearly 
a kHz.  It's amplitude is also much lower than the saw.  It is this lower 
amplitude of the square wave that drives the filter less hard, and results 
in more resonance from the filter when the square wave is selected. 
Conversely the sawtooth drives the filter harder and reduces the resonance.

> chirp as chirp can.
> which also would explain the pronunciation of the higher harmonics
> of the VCOs waveforms.

Yes, there is plenty of high-pass action in the VCF and the VCA that 
follows.  This results in the "weedy bass response" that Adam mentioned.  To 
me the TB-303 filter sounds distinctly "spitty" and "nasal" as a result of 
it's resonance increasing with frequency and the actual low-frequency bass 
energy being rolled off.  (You can clearly see the LF rolloff caused by the 
high-pass action on the left side of the graph I posted.)

> What about lower filter frequencies? … or is that as low as the 303 goes 
> for cutoff?

It goes a little bit lower.  Just for you... ;-)

http://www.richieburnett.co.uk/temp/303vcf2.gif

-Richie,