RE: wavwrite

[Schirmacher, Rolf]

Hello Robert,

where did you get your wavwrite from?

Mine is the one distributed with the windows binary distribution of 2.9.13,
giving

octave.exe> help wavwrite
 -- Function File:  wavwrite (FILENAME, Y)
     Write Y to the canonical RIFF/WAVE sound file FILENAME. A sample
     rate of 8000 Hz and 16-bit samples are assumed. Each column of
     the data represents a separate channel.

 -- Function File:  wavwrite (FILENAME, Y, FS)
     Set the sample rate to FS Hz.

 -- Function File:  wavwrite (FILENAME, Y, FS, BITS)
     Set the sample rate to FS Hz and resolution to BITS bits.

     See also: wavread.

c:\programme\octave\share\octave\2.9.13\m\audio\wavwrite.m


Perhaps, you might contribute your wavwrite?

Regards,

Rolf

> -----Original Message-----
> From: Robert A. Macy [mailto:address@hidden
> Sent: Tuesday, October 09, 2007 5:15 PM
> To: Schirmacher, Rolf
> Cc: address@hidden
> Subject: Re: wavwrite
> 
> 
> My wavwrite is EXACTLY like Matlab.
> 
> >> wavwrite(signal,sf,bits,"filename.wav");
> 
> PLUS, I routinely have been using it to write 24 bit wav
> files and routinely using wavread to read 24 bit files,
> again, in the same order:
> >> [signal, sf, bits]=wavread("filename.wav");
> 
> What's the problem?
> 
> Robert
> 
> On Tue, 9 Oct 2007 09:39:13 +0200 
>  "Schirmacher, Rolf" <address@hidden>
> wrote:
> > 
> > Hello,
> > 
> > I just found some incompatibility in between octave an
> > matlab for wavwrite.
> > 
> > Octave wants to get the filename first, Matlab last. So
> > we have 
> > wavwrite(filename, signal, fs, bits) in Octave and
> > wavwrite(signal, fs, bits, filename) in Matlab.
> > 
> > Changing it would be straight forward (as far as I
> > see...), but it would
> > break existing Octave code :-(
> > 
> > So, how to proceed? 
> > 
> > Rolf
> > 
> > BTW, there are some more differences: 24 bit is currently
> > not supported in
> > octave and 32 bit is implemented as 32 bit integer.
> > Matlab has 24 bit
> > integer and 32 bit normalised floats. I regard this a
> > minor issue as both
> > seems to be correct / legal within the data format and
> > would not really
> > break code (except of giving an "unsupported" error for
> > octave 24 bit).
>