Re: [iPAQ] underwater audio

On Fri, 22 Sep 2000, Dwayne Jacques Fontenot wrote:

>
> > On Fri, 22 Sep 2000, Dwayne Jacques Fontenot wrote:
> >
> > >
> > > Nicolas,
> > >
> > > Are you saying that I am actually hearing decoding errors and not a problem
> > > with the sound driver?
> >
> > Testing your mp3 with madplay would tell.
>
> same version (?) 0.59r of mpg123 on my PC plays it fine.
>
> I will also try madplay, but it sounds perfect in mpg123 x86 as it is.

Understand that the SA1110 CPU doesn't have a FPU. For that reason, a MP3
player must implement fixed point computation to run real time on the
iPaq.

On a PC, mpg123 uses floating point operations because the CPU supports
them, and mpg123's floating point accuracy is rated "full compliant". On
iPaq, mpg123 uses its fixed point implementation but this time its
accuracy is rated "not compliant".

> heh, "go bubble". :-) You make it sound like a standard term. :-)

Having worked on MP3 decoding routines and audio processing for some time,
that's the term (or more exactly a translation of it) we used in my team
when the MP3 decoder failed to produce the correct output. The most
common sign of a lack of accuracy in the computed signal is the impression
of bubbles in the sound...

Nicolas
Received on Sat Sep 23 11:11:08 2000