blackman and flattopwin cannot be used with pwelch

[Francesco Potorti`]

>>>> pwelch.m as distributed in the signal package of octave-forge does not
>>>> accept a window with negative values.  However, blackman(4096) and
>>>> flattopwin(4096) generate negative values.
>>>>
>>>> I suspect that as far as blackman is concerned, it is just a rounding
>>>> error, but flattopwin definitely generates negative values.
>>>>       
>>> Could you propose a fix?

Here it is.  By the way, I corrected a couple of places where the name
of the window (Hann, Hamming) was used where the name of the function
(hanning, hamming) should have.

--- pwelch.m.orig	2008-01-18 13:38:54.000000000 +0100
+++ pwelch.m	2008-02-05 14:13:04.000000000 +0100
@@ -62,7 +62,7 @@
 %% y           %% [non-empty vector] system-output time-series data
 %%
-%% window      %% [real vector] of window-function values between 0 and 1; the
+%% window      %% [real vector] of window-function real values; the
 %%             %%       data segment has the same length as the window.
-%%             %%       Default window shape is Hamming.
+%%             %%       Default window shape is hamming.
 %%             %% [integer scalar] length of each data segment.  The default
 %%             %%       value is window=sqrt(length(x)) rounded up to the
@@ -187,5 +187,5 @@
 %%    20dB/decade.  You can inspect the FFT of a Hann window by plotting
 %%    "abs(fft(postpad(hanning(256),4096,0)))".
-%%    The default window is Hamming.
+%%    The default window is hamming.
 %% 4) ZERO PADDING:
 %%    Zero-padding reduces the frequency step in the
@@ -229,5 +229,5 @@
 %%          %% units of overlap are "number of samples"
 %%          %% defaults: Nfft=min(length(x),256), Fs=2*pi, length(window)=Nfft,
-%%          %%           window=Hanning, do not detrend,
+%%          %%           window=hanning, do not detrend,
 %%          %% N.B.  "Sloppy" is not available.
 %%
@@ -236,5 +236,5 @@
 %%          [Pxx,f]=pwelch(x,window,overlap,nfft,Fs,...);
 %%          %% units of overlap are "number of samples"
-%%          %% defaults: length(window)==length(x)/8, window=Hamming,
+%%          %% defaults: length(window)==length(x)/8, window=hamming,
 %%          %%           Nfft=max(256,NextPow2), Fs=2*pi, do not detrend
 %%          %% NextPow2 is the next power of 2 greater than or equal to the
@@ -487,6 +487,6 @@ else
       elseif ( is_win==1 && ( ~isreal(arg) || fix(arg)~=arg || arg<=3 ) )
         error( 'pwelch: arg %d (window) must be integer >3', iarg+1 );
-      elseif ( is_win>1 && ( ~isreal(arg) || any(arg<0) ) )
-        error( 'pwelch: arg %d (window) vector must be real and >=0',iarg+1);
+      elseif ( is_win>1 && ( ~isreal(arg) ) )
+        error( 'pwelch: arg %d (window) vector must be real',iarg+1);
       end
       window = arg;

-- 
Francesco Potortì (ricercatore)        Voice: +39 050 315 3058 (op.2111)
ISTI - Area della ricerca CNR          Fax:   +39 050 315 2040
via G. Moruzzi 1, I-56124 Pisa         Email: Potorti at isti.cnr.it
Web: http://fly.isti.cnr.it/           Key:   fly.isti.cnr.it/public.key