hi again..
ok.. now i also had to do my version of it..... :-) you'll find it in the attachment.
a few notes on the collection: the one that i saw from uğur güney [ciprit] works only as expected at a fixed samplingrate (44100). steffens proposal is behaves exactly as the original [dirac~], as far as i can see it, whereas frank's approach [vimpulse~] does the time-to-sample conversion automatically, which is, i think, probably rather the behaviour, that someone usually wants. the helpfile states, that [vimpulse~] is subsample accurate. as far as i can see, this is wrong (it is only sample accurate) and also not a behaviour that someone would want. if it would be truely subsample accurate, then most of the time the dirac impulse would need to be interpolated and composed of two subsequent samples (which would it turn unusuable for many applications). my attached version behaves similar to franks version, but instead of writing first the signal into a table, it uses directly [vline~] to create the impulse.
On Sun, 2007-11-18 at 00:38 +0100, Steffen Leve Poulsen wrote:
Andy Farnell skrev:
The [dirac~] object generates a unit impulse (at the sample position given by its input)
I'm trying to think of simple ways to do it in plain vanilla. Frank gave a one way before, differentiating a step impulse with [rzero~], can anyone think of other ways (I'm sure there's more than one)
i missed this one and i also didn't quite get how it works. can someone send me this one?
roman