[PD] band-limited square wave

[Martin Peach]

matthew jones wrote:
>>Do you hear them?
>>
> yes.  erm, wow I have to admit I'm stumped....goes against all my previous
> learning.

I tried this and got a zero at 2205 Hz, which is exactly 1/20 of 44100, 
so it is definitely aliasing with the sample frequency. I assume the 
[>~] object is processed at the sample level and not block level, so 
then the noise is coming from the desired step transitions occurring 
'in-between' samples. This could be overcome by interpolating the two 
samples around each transition point, but definitely there is a need for 
a sqosc~ object that would do that internally.

Martin