fre, 29 02 2008 kl. 12:03 +0100, skrev Eric Chassande-Mottin:
> hi,
> 
>  I have rewritten resample.m using a polyphase implementation.
>  See codes attached to this message.
> 
>  With the test resample_test.m, I compare this code to the current version
>  of resample.m in OctaveForge. The test consists in
>  resampling sinusoids of various frequencies and evaluate the discrepancy
>  between the result of the resampling algorithm and
>  the exact waveform one expects.
> 
>  in (a), one sees that the RMS error of
>  OctaveForge's resample.m is large and this is
>  due to a spurious phase shift in the resampled
>  time-series (see (c) red and blue).
> 
>  in (b), the rejection of OctaveForge's resample.m
>  appears to be quite poor.
> 
>  I've checked that the polyphase version satisfies the
>  specifications (rolloff width and rejection in the
>  stopband) of the antialiasing filter (blue dash lines
>  in (a) and (b)).
> 
>  I suggest to replace resample.m with my code.
Other people will have to comment the algorithm as I'm not qualified. I
do have a few comments:

1) Your implementation is not documented. I know this is the boring
part, but it really needs to be done. Otherwise people can only use the
program if they read the actual code.

2) You should add a version number to the GPL license. I would prefer if
you chose GPL version 3 (or later) as this is what Octave uses.

3) There is no need to add a semi-colon (;) after 'end*' keywords. So,
you don't have to write "endif;", you can just use "endif". The same
goes for "endwhile", "endfor", and so on.

4) I didn't read your test program, but it might be nice if this was
added as a test or demo in the actual implementation, using "%!test" or
"%!demo".

Thanks,
Søren


-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Octave-dev mailing list
Octave-dev@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/octave-dev

Reply via email to