Csound Csound-dev Csound-tekno Search About

hetro/adsyn

Date1999-03-20 20:16
FromBlake Markle
Subjecthetro/adsyn
Good day fellow Csounders,

We're still getting bad quality resynthesis using Hetro/adsyn.  We are
analyzing a sustained piano note in hetro, and resynthesizing it with
adsyn.  Even with many (50+) harmonics included and many (1024)
amplitude breakpoints checked, we get bad transient response (attack)
and distortion.  We have an "owl" sounding hoot in the attack and lots
of high frequency noise distortion.

We are using the latest csound edition (csound_con.zip) with winsound
for the adsyn process.  The file format is wav.  We would welcome any
wisdom on the subject.  If it is possible to create a reasonably
realistic analysis/resynthesis, please let us know.

Thanks,

Greg and Blake

Date1999-03-21 20:27
FromThomas Neuhaus
SubjectRe: hetro/adsyn
On Sat, 20 Mar 1999, Blake Markle wrote:

> Good day fellow Csounders,
> 
> We're still getting bad quality resynthesis using Hetro/adsyn.  We are
> analyzing a sustained piano note in hetro, and resynthesizing it with
> adsyn.  Even with many (50+) harmonics included and many (1024)
> amplitude breakpoints checked, we get bad transient response (attack)
> and distortion.  We have an "owl" sounding hoot in the attack and lots
> of high frequency noise distortion.
> 
> We are using the latest csound edition (csound_con.zip) with winsound
> for the adsyn process.  The file format is wav.  We would welcome any
> wisdom on the subject.  If it is possible to create a reasonably
> realistic analysis/resynthesis, please let us know.
> 


The effect you experience is normal due to the nature of hetrodyning
analysis and additive resynthesis. First the attack portion contains a
lot of noise which cannot be reproduced by only 50+ partials, second the
partials of a piano sound are not exactly harmonic overtones. A high
quality resynthesis should usa a dual approach such as additive synthesis
for the steady part and subtractive (filtered noise) for the residual
part.

For a good discussion of the subject see

Musical sound modelling using sinusoids plus noise
in
Musical Signal Processing edited by C. Roads et al
Swets and Zeitlinger Publishers
ISBN 90-265-1483-2


Hope that helps

Thomas


 --
Thomas Neuhaus(neuhaus@folkwang.uni-essen.de) Phone (49)-201-4903-333
ICEM Institut fuer Computermusik und elektronische Medien
Folkwang-Hochschule Essen, Klemensborn 39, D-49239 Essen  
-------------------------------------------------------------------------------
UNIX _is_ user-friendly. It just knows who its friends are.