I noticed that the same problem is presente even in the plucked string model. Probably it
depends by the fact that the delay of the filter  is still non-fractional. For example,
when no filter is present in the string model, the pitch of high notes is OK. I suppose
(but I'm not sure) that also filter coefficients should change when changing pitch.

Gab

Josep M Comajuncosas wrote:
> 
> Hi all,
> I´m having serious problems with a waveguide Physical Model of a
> clarinet I´m building. In fact it is really *out of tune* (though the
> delay line uses interpolation) and I´d like to know the most probable
> causes. I thought they could be the lowpass in the feedback loop (which
> induces some freq. dependant delay, but it doesn´t seem to be so big)
> and the waveshaping function also inside the loop. What do you think?
> I´d like to know if there is a way to apply waveshaping without inducing
> 
> phase delays, or a way to calculate it to compensate for the delay in
> the loop.
> Thancks in advance!
> 
> Josep M Comajuncosas

-- 
Gabriel Maldonado