Speex dev - Sep 2005 - Unifying fixed point and floating point

I've been browsing around the speex-1.10 source tree.

Firstly I see 'float' is still used in many places. Can't we change
them all to the appropriate type (spx_wordxx_t or similar) defined in arch.h ?

Secondly I see there is a lot of duplication regarding real-valued constants.
One way would be to say:
#ifndef FIXED_POINT
#define CONSTANT(x) (x)
#else
#define CONSTANT(x) ((int)((x)*16384.0 + 0.5))
#endif
and then use the macro appropriately.

This would assume that all the compilers we intent to support has the ability to
evaluate constant expressions at compile time. In my experience this assumption
is reasonable.

Regards
-n
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/speex-dev/attachments/20050918/02c7471e/attachment.htm

Hi Nic,

Things are a bit more complicated than that. I can't change all the
floats to speex_wordxx_t without first finding out what fixed-point
representation I want to use for each value and then choose the MULxx
operator to use (and making sure there's no possible overflow or loss of
precision).

	Jean-Marc

Le dimanche 18 septembre 2005 ? 00:05 +0200, Nic Roets a ?crit
:
> I've been browsing around the speex-1.10 source tree.
>  
> Firstly I see 'float' is still used in many places. Can't we
change
> them all to the appropriate type (spx_wordxx_t or similar) defined in
> arch.h ?
>  
> Secondly I see there is a lot of duplication regarding real-valued
> constants. One way would be to say:
> #ifndef FIXED_POINT
> #define CONSTANT(x) (x)
> #else
> #define CONSTANT(x) ((int)((x)*16384.0 + 0.5))
> #endif
> and then use the macro appropriately.
>  
> This would assume that all the compilers we intent to support has the
> ability to evaluate constant expressions at compile time. In my
> experience this assumption is reasonable.
>  
> Regards
> -n
>  
> _______________________________________________
> Speex-dev mailing list
> Speex-dev@xiph.org
> http://lists.xiph.org/mailman/listinfo/speex-dev
-- 
Jean-Marc Valin <Jean-Marc.Valin@USherbrooke.ca>
Universit? de Sherbrooke