Speex dev - Apr 2014 - [PATCH 2/3] Use fabsf() instead of fabs() since we have floats, not double

On Tue, Apr 15, 2014 at 12:33 PM, Peter Meerwald <pmeerw at pmeerw.net>
wrote:
> Signed-off-by: Peter Meerwald <pmeerw at pmeerw.net>
> ---
>  libspeexdsp/resample.c | 14 +++++++-------
>  1 file changed, 7 insertions(+), 7 deletions(-)
>
> diff --git a/libspeexdsp/resample.c b/libspeexdsp/resample.c
> index e32ca45..a19b997 100644
> --- a/libspeexdsp/resample.c
> +++ b/libspeexdsp/resample.c
> @@ -85,7 +85,7 @@ static void speex_free (void *ptr) {free(ptr);}
>  #ifdef FIXED_POINT
>  #define WORD2INT(x) ((x) < -32767 ? -32768 : ((x) > 32766 ? 32767 :
(x)))
>  #else
> -#define WORD2INT(x) ((x) < -32767.5f ? -32768 : ((x) > 32766.5f ?
32767 :
> floor(.5+(x))))
> +#define WORD2INT(x) ((x) < -32767.5f ? -32768 : ((x) > 32766.5f ?
32767 :
> floorf(.5+(x))))
>  #endif
>
>  #define IMAX(a,b) ((a) > (b) ? (a) : (b))
> @@ -273,12 +273,12 @@ static spx_word16_t sinc(float cutoff, float x, int
> N, struct FuncDef *window_fu
>  {
>     /*fprintf (stderr, "%f ", x);*/
>     float xx = x * cutoff;
> -   if (fabs(x)<1e-6f)
> +   if (fabsf(x)<1e-6f)
>        return WORD2INT(32768.*cutoff);
> -   else if (fabs(x) > .5f*N)
> +   else if (fabsf(x) > .5f*N)
>        return 0;
>     /*FIXME: Can it really be any slower than this? */
> -   return WORD2INT(32768.*cutoff*sin(M_PI*xx)/(M_PI*xx) *
> compute_func(fabs(2.*x/N), window_func));
> +   return WORD2INT(32768.*cutoff*sin(M_PI*xx)/(M_PI*xx) *
> compute_func(fabsf(2.*x/N), window_func));
>  }
>  #else
>  /* The slow way of computing a sinc for the table. Should improve that
> some day */
> @@ -286,12 +286,12 @@ static spx_word16_t sinc(float cutoff, float x, int
> N, struct FuncDef *window_fu
>  {
>     /*fprintf (stderr, "%f ", x);*/
>     float xx = x * cutoff;
> -   if (fabs(x)<1e-6)
> +   if (fabsf(x)<1e-6f)
>        return cutoff;
> -   else if (fabs(x) > .5*N)
> +   else if (fabsf(x) > .5f*N)
>        return 0;
>     /*FIXME: Can it really be any slower than this? */
> -   return cutoff*sin(M_PI*xx)/(M_PI*xx) * compute_func(fabs(2.*x/N),
> window_func);
> +   return cutoff*sin(M_PI*xx)/(M_PI*xx) * compute_func(fabsf(2.*x/N),
> window_func);
>  }
>  #endif
>
Seems fine, fabsf is supported everywhere fabs is?

Best,
Tristan
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://lists.xiph.org/pipermail/speex-dev/attachments/20140415/14518979/attachment.htm

On 15/04/14 03:45 PM, Tristan Matthews wrote:
> Seems fine, fabsf is supported everywhere fabs is?
No. AFAIK, fabsf() is C99-only, which would be a problem.

	Jean-Marc

> > Seems fine, fabsf is supported everywhere fabs is?
> 
> No. AFAIK, fabsf() is C99-only, which would be a problem.
right, fabsf() is C99, fabs() is C89 (according to fabs(3) man)

I wasn't aware of this constraint; it might be a good idea to enforce/test 
C89 compliance with "gcc -std=c89" then

p.

-- 

Peter Meerwald
+43-664-2444418 (mobile)