search for: apsort

...-29,6 +29,8 @@ #include "scales.h" #include "misc.h" +#include <stdint.h> + #define NEGINF -9999.f static double stereo_threshholds[]={0.0, .5, 1.0, 1.5, 2.5, 4.5, 8.5, 16.5, 9e10}; @@ -966,9 +968,17 @@ /* this is for per-channel noise normalization */ static int apsort(const void *a, const void *b){ - float f1=fabs(**(float**)a); - float f2=fabs(**(float**)b); +#if 0 + float f1=fabsf(**(float**)a); + float f2=fabsf(**(float**)b); return (f1<f2)-(f1>f2); +#else + uint32_t u1 = **(uint32_t **)a; + uint32_t u2 = **(uint32_t **)b; + u1 <<= 1; +...

I have found a bug in libvorbis-1.0. I have found a fix, but need to know in what format and to where to sumbit the patch. It involves a quirky qsort in Solaris 5.8 that dumps core without the patch. The fix involves editing vorbis/lib/psy.c <p>--- >8 ---- List archives: http://www.xiph.org/archives/ Ogg project homepage: http://www.xiph.org/ogg/ To unsubscribe from this list, send a

...I am using vorbis-tools 1.0 on Solaris 8 x86. This is the command line: oggenc -b 128 -a "Keith, Toby" -t "Losing My Touch" -l "Unleashed" \ - -N "06" -o "Keith, Toby - 06. Losing My Touch.ogg" audio_06.wav Here is a backtrace: #0 0xdfa8e36b in apsort () from /usr/local/lib/libvorbis.so.0 #1 0xdf8eb94d in qsort () from /usr/lib/libc.so.1 #2 0xdfa8d90a in _vp_noise_normalize_sort () from /usr/local/lib/libvorbis.so.0 #3 0xdfa94ac4 in mapping0_forward () from /usr/local/lib/libvorbis.so.0 #4 0xdfa8b025 in vorbis_analysis () from /usr/local/...

...0> diff -u lib/psy.c lib/psy_new.c > libv.patch <p><p><p><p> -------------- next part -------------- --- lib/psy.c Sat Jul 13 05:18:33 2002 +++ lib/psy_new.c Fri Jul 26 14:50:10 2002 @@ -950,7 +950,10 @@ /* this is for per-channel noise normalization */ static int apsort(const void *a, const void *b){ - if(fabs(**(float **)a)>fabs(**(float **)b))return -1; + float f1=fabsf(**(float**)a); + float f2=fabsf(**(float**)b); + if(f1>f2)return -1; + if(f1==f2)return 0; return 1; }

Hi, We compiled libvorbis with vc6 enviroment. We've discovered that changing float to int conversion results in a 4% speedup on a benchmark we've come with on pentium4 3GHz. We'd like to submit the code. We haven't checked it on AMD. Yet, as you can see in the code, it shouldn't be a problem. Should we think of other machines except for AMD/Intel machines? Tal&Adir P.S.