Re: 32bit pcm and float support in .wav Input plugin

> This patch was, however, not 64 bit clean.  In particular you cannot
> depend on 'long' beeing 32 bits.  One should use one of the glib types
> with specific size, i.e. guint32 when you depend on a variable beeing of a
> specific size.

I'll have a look at the CVS version again tomorrow to recheck it.

> Did you test 12 bits files?  They seem to be broken to me.

Yes I did a test with 12 bit files. The wav standard requires them each 
sample to be contained in a 16bit word in the upper bits (the 4 lower 
bits are 0). So there's no difference in playback compared to a 16bit 
wav. It's just that the fmt chunk contains the 12 so a program reading 
the wav may truncate the lower 4 bits of each sample.

-- 
---> doj / cubic
----> http://cubic.org/~doj
-----> http://llg.cubic.org