Re: [OctDev] pwelch bugs

[dbateman <dba...@fr...>]

Dmitri A. Sergatskov wrote:
> 
> The new version of pwelch (by  Peter V. Lanspeary) seems to have problem
> parsing argument list. E.g. "demo pwelch" would fail:
> 
> pwelch example 1: failed
> error: pwelch: arg 3 (overlap) must be real from 0 to 0.950000
> 
> apparently on the line pwelch(x,[],[],[],[],0.9);
> 
> It also gets easily confused when pwelch called with two channels, e.g.:
> 
> pwelch(x, y, [], [], [], 1000, [], [], "mean", [], "trans")
> error: pwelch: can give confidence interval for x power spectrum only
> error: evaluating if command near line 507, column 3
> error: evaluating if command near line 279, column 1
> error: called from `pwelch' in file
> `/usr/share/octave/site/m/octave-forge/signal/pwelch.m'
> 
> In fact it appears that i cannot specify any parameter between
> "sampling frequency" and "result" parameters...
> 
> Sorry, I do not have a patch, the source of pwelch.m is too
> unwieldy for me.
> 
> Sincerely,
> 
> Dmitri.
> 

The demo ran through fine for me with the 20070528 release of Octave forge
and 2.9.12. Could you give another simpler example that fails?

D.

-- 
View this message in context: http://www.nabble.com/pwelch-bugs-tf3854366.html#a10920639
Sent from the octave-dev mailing list archive at Nabble.com.