diff options
author | Michael Niedermayer <michaelni@gmx.at> | 2011-10-23 01:49:32 +0200 |
---|---|---|
committer | Michael Niedermayer <michaelni@gmx.at> | 2011-10-23 05:13:56 +0200 |
commit | f97faf67510d92c2a18180a6ec80435d5dd1da0b (patch) | |
tree | b63cb9b8a5fc5bafa6963c07c712dbbdf419a175 /libswscale/ppc/swscale_altivec.c | |
parent | 9e8dff90efa3faae5ac07fa5fae7e154367e24fd (diff) | |
parent | e71ebb19722bd2b46831d0ec311b757a56340617 (diff) | |
download | ffmpeg-f97faf67510d92c2a18180a6ec80435d5dd1da0b.tar.gz |
Merge remote-tracking branch 'qatar/master'
* qatar/master:
id3v2: fix doxy comment - 'machine byte order' makes no sense on char arrays
VC1: restore mistakenly removed code
twinvq: check output buffer size before decoding
twinvq: return an error when the packet size is too small
lavf: export some forgotten symbols with non-av prefixes.
swscale: update altivec yuv2planeX asm to new per-plane API.
swscale: make yuv2yuvX_10_sse2/avx 8/9/16-bits aware.
yuv2planeX10 SIMD
swscale: decide whether to use yuv2plane1/X on a per-plane basis.
swscale: reintroduce full precision in 16-bit output.
Split up yuv2yuvX functions
Split out yuv2yuv1 luma and chroma in order to make them generic DSP functions
lavc: replace references to deprecated AVCodecContext.error_recognition to use AVCodecContext.err_recognition
lavc: translate non-flag-based er options into flag-based ef options at codec open
add -err_filter AVOptions to access flag-based error recognition
h264_weight: initialize "height" function argument properly.
presets: spelling error in libvpx 1080p50_60
avplay: fix fullscreen behaviour with SDL 1.2.14 on Mac OS X
Conflicts:
ffplay.c
libavformat/libavformat.v
libswscale/swscale.c
libswscale/x86/swscale_template.c
tests/ref/lavfi/pixfmts_scale
Merged-by: Michael Niedermayer <michaelni@gmx.at>
Diffstat (limited to 'libswscale/ppc/swscale_altivec.c')
-rw-r--r-- | libswscale/ppc/swscale_altivec.c | 94 |
1 files changed, 13 insertions, 81 deletions
diff --git a/libswscale/ppc/swscale_altivec.c b/libswscale/ppc/swscale_altivec.c index 076929cdaa..f2a8600925 100644 --- a/libswscale/ppc/swscale_altivec.c +++ b/libswscale/ppc/swscale_altivec.c @@ -94,34 +94,29 @@ altivec_packIntArrayToCharArray(int *val, uint8_t* dest, int dstW) //FIXME remove the usage of scratch buffers. static void -yuv2yuvX_altivec_real(SwsContext *c, - const int16_t *lumFilter, const int16_t **lumSrc, - int lumFilterSize, const int16_t *chrFilter, - const int16_t **chrUSrc, const int16_t **chrVSrc, - int chrFilterSize, const int16_t **alpSrc, - uint8_t *dest[4], int dstW, int chrDstW) +yuv2planeX_altivec(const int16_t *filter, int filterSize, + const int16_t **src, uint8_t *dest, int dstW, + const uint8_t *dither, int offset) { - uint8_t *yDest = dest[0], *uDest = dest[1], *vDest = dest[2]; - const uint8_t *lumDither = c->lumDither8, *chrDither = c->chrDither8; register int i, j; { DECLARE_ALIGNED(16, int, val)[dstW]; for (i=0; i<dstW; i++) - val[i] = lumDither[i & 7] << 12; + val[i] = dither[(i + offset) & 7] << 12; - for (j = 0; j < lumFilterSize; j++) { - vector signed short l1, vLumFilter = vec_ld(j << 1, lumFilter); - vector unsigned char perm, perm0 = vec_lvsl(j << 1, lumFilter); + for (j = 0; j < filterSize; j++) { + vector signed short l1, vLumFilter = vec_ld(j << 1, filter); + vector unsigned char perm, perm0 = vec_lvsl(j << 1, filter); vLumFilter = vec_perm(vLumFilter, vLumFilter, perm0); vLumFilter = vec_splat(vLumFilter, 0); // lumFilter[j] is loaded 8 times in vLumFilter - perm = vec_lvsl(0, lumSrc[j]); - l1 = vec_ld(0, lumSrc[j]); + perm = vec_lvsl(0, src[j]); + l1 = vec_ld(0, src[j]); for (i = 0; i < (dstW - 7); i+=8) { int offset = i << 2; - vector signed short l2 = vec_ld((i << 1) + 16, lumSrc[j]); + vector signed short l2 = vec_ld((i << 1) + 16, src[j]); vector signed int v1 = vec_ld(offset, val); vector signed int v2 = vec_ld(offset + 16, val); @@ -143,73 +138,10 @@ yuv2yuvX_altivec_real(SwsContext *c, l1 = l2; } for ( ; i < dstW; i++) { - val[i] += lumSrc[j][i] * lumFilter[j]; + val[i] += src[j][i] * filter[j]; } } - altivec_packIntArrayToCharArray(val, yDest, dstW); - } - if (uDest != 0) { - DECLARE_ALIGNED(16, int, u)[chrDstW]; - DECLARE_ALIGNED(16, int, v)[chrDstW]; - - for (i=0; i<chrDstW; i++) { - u[i] = chrDither[i & 7] << 12; - v[i] = chrDither[(i + 3) & 7] << 12; - } - - for (j = 0; j < chrFilterSize; j++) { - vector signed short l1, l1_V, vChrFilter = vec_ld(j << 1, chrFilter); - vector unsigned char perm, perm0 = vec_lvsl(j << 1, chrFilter); - vChrFilter = vec_perm(vChrFilter, vChrFilter, perm0); - vChrFilter = vec_splat(vChrFilter, 0); // chrFilter[j] is loaded 8 times in vChrFilter - - perm = vec_lvsl(0, chrUSrc[j]); - l1 = vec_ld(0, chrUSrc[j]); - l1_V = vec_ld(0, chrVSrc[j]); - - for (i = 0; i < (chrDstW - 7); i+=8) { - int offset = i << 2; - vector signed short l2 = vec_ld((i << 1) + 16, chrUSrc[j]); - vector signed short l2_V = vec_ld((i << 1) + 16, chrVSrc[j]); - - vector signed int v1 = vec_ld(offset, u); - vector signed int v2 = vec_ld(offset + 16, u); - vector signed int v1_V = vec_ld(offset, v); - vector signed int v2_V = vec_ld(offset + 16, v); - - vector signed short ls = vec_perm(l1, l2, perm); // chrUSrc[j][i] ... chrUSrc[j][i+7] - vector signed short ls_V = vec_perm(l1_V, l2_V, perm); // chrVSrc[j][i] ... chrVSrc[j][i] - - vector signed int i1 = vec_mule(vChrFilter, ls); - vector signed int i2 = vec_mulo(vChrFilter, ls); - vector signed int i1_V = vec_mule(vChrFilter, ls_V); - vector signed int i2_V = vec_mulo(vChrFilter, ls_V); - - vector signed int vf1 = vec_mergeh(i1, i2); - vector signed int vf2 = vec_mergel(i1, i2); // chrUSrc[j][i] * chrFilter[j] ... chrUSrc[j][i+7] * chrFilter[j] - vector signed int vf1_V = vec_mergeh(i1_V, i2_V); - vector signed int vf2_V = vec_mergel(i1_V, i2_V); // chrVSrc[j][i] * chrFilter[j] ... chrVSrc[j][i+7] * chrFilter[j] - - vector signed int vo1 = vec_add(v1, vf1); - vector signed int vo2 = vec_add(v2, vf2); - vector signed int vo1_V = vec_add(v1_V, vf1_V); - vector signed int vo2_V = vec_add(v2_V, vf2_V); - - vec_st(vo1, offset, u); - vec_st(vo2, offset + 16, u); - vec_st(vo1_V, offset, v); - vec_st(vo2_V, offset + 16, v); - - l1 = l2; - l1_V = l2_V; - } - for ( ; i < chrDstW; i++) { - u[i] += chrUSrc[j][i] * chrFilter[j]; - v[i] += chrVSrc[j][i] * chrFilter[j]; - } - } - altivec_packIntArrayToCharArray(u, uDest, chrDstW); - altivec_packIntArrayToCharArray(v, vDest, chrDstW); + altivec_packIntArrayToCharArray(val, dest, dstW); } } @@ -405,7 +337,7 @@ void ff_sws_init_swScale_altivec(SwsContext *c) if (!is16BPS(dstFormat) && !is9_OR_10BPS(dstFormat) && dstFormat != PIX_FMT_NV12 && dstFormat != PIX_FMT_NV21 && !c->alpPixBuf) { - c->yuv2yuvX = yuv2yuvX_altivec_real; + c->yuv2planeX = yuv2planeX_altivec; } /* The following list of supported dstFormat values should |