[CLAM] about CLAM's demo output and another issue

Xavier Amatriain xavier at create.ucsb.edu
Wed Jul 19 23:38:51 PDT 2006


Neediss Gu wrote:
> Any transformation such as large or middle range pitch modification, 
> but the output should sounds less artifact.
> Can anyone help me to generate a demo output based 
> on  http://neediss.googlepages.com/female.wav.
You can find some samples of rather extreme PitchShift at 
http://www.create.ucsb.edu/~xavier/temp/female.zip

I did not bother setting the analysis params and simply  did the 
transformation with the defaults. That means that the results could be 
much better
than this with some work.

> runtime error R6025
> - pure virtual function call
>
> **
It looks like this is a Windows error on some versions (google dixit). I 
recommend you install the newest service packs (or switch to Linux :)

> if(mag[i]==0) mag[i]=TData(0.0001);  -> ? Maybe it should be 
> if(mag[i]==0) mag[i]=TData(-200.0); 
> In "void Spectrum::ToDB()", it is -200.0. And -200 may be better 
> meanful than 0.001.
No. This is correct. mag[i] is in linear scale so 0.0001 is actually 
more or less equivalent to -200 in dB.





More information about the clam-users mailing list