search for: bark_noise_median

Could someone please try this version of bark_noise_median() on a x86 box? On my PPC box, this runs about 1.85x as fast as the current version, making the encoding speed about 6.5% faster (at least on my test case). The basic idea is to (a) avoid needless float/int conversions and (b) use multiplies instead of divides. I'm assuming (after having d...

...p->noiseoffset=_ogg_malloc(n*sizeof(float *)); p->tonecurves[i]=_ogg_malloc(P_LEVELS*sizeof(float *)); new (corrected) lines: p->noisemedian=_ogg_malloc(n*sizeof(float)); p->noiseoffset=_ogg_malloc(n*sizeof(float)); p->tonecurves[i]=_ogg_malloc(P_LEVELS*sizeof(float)); 3. modified bark_noise_median(): 3/a: floating point based (new) style: #define BIN(x) ((int)((x)*negFour)) #define BINdB(x) ((x)*negQuarter) #define BINCOUNT (200*4) #define LASTBIN (BINCOUNT-1) tatic void bark_noise_median(long n,float *b, float *f,float *noise,...

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, While trying to build the current vorbis libraries and examples on MacOS with CodeWarrior 6, I stumbled on some problems, for which I've made some patches. The first problem was the static function bark_noise_median() in vorbis/lib/psy.c (r1.49), which has as a second parameter a "const long *". It is only called once in _vp_compute_mask() in psy.c, but with a "unsigned long *" parameter (it's the bark member of the vorbis_look_psy structure). This causes CW6 to complain with: Error...