[sdiy] karplus-strong drum synthesis

[Tom Wiltshire]

On 21 Aug 2011, at 09:22, Gordon JC Pearce wrote:

> On Sat, 20 Aug 2011 13:02:29 -0700
> Joel B <onephatcat at earthlink.net> wrote:
> 
>> Assuming i'm even a little bit right about this, the question is, if I start at sample 0, there is no sample - 1 to use in the equation, so how is that the condition handled? Do you wrap to the last sample at t the end of the wavetable?
> 
> A Karplus-Strong synthesizer is basically a delay line with feedback, full of some noise, with a low-pass filter in the feedback loop.  As long as you stick a few values in there to begin with then what happens first time round the loop isn't important.
> 
> Implement it as a simple ringbuffer with fractional delay and you're set.

What techniques do people use for fractional delay?

By which I mean, "What techniques have you people used successfully for fractional delay?" rather than "what approaches are there?"

It seems to me there are various equivalent ways of looking at it; filtering, interpolation, oversampling. Maybe these would lead you to different solutions, or maybe it's all the same.

Thanks,
Tom