From 19a0729b4cfacfd90b8ee84ab0c093ff7e397e65 Mon Sep 17 00:00:00 2001 From: Oskar Arvidsson <oskar@irock.se> Date: Tue, 29 Mar 2011 17:48:59 +0200 Subject: Adds 8-, 9- and 10-bit versions of some of the functions used by the h264 decoder. This patch lets e.g. dsputil_init chose dsp functions with respect to the bit depth to decode. The naming scheme of bit depth dependent functions is <base name>_<bit depth>[_<prefix>] (i.e. the old clear_blocks_c is now named clear_blocks_8_c). Note: Some of the functions for high bit depth is not dependent on the bit depth, but only on the pixel size. This leaves some room for optimizing binary size. Preparatory patch for high bit depth h264 decoding support. Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com> --- libavcodec/ppc/dsputil_altivec.c | 5 +++++ libavcodec/ppc/dsputil_ppc.c | 4 ++++ libavcodec/ppc/h264_altivec.c | 7 ++++++- 3 files changed, 15 insertions(+), 1 deletion(-) (limited to 'libavcodec/ppc') diff --git a/libavcodec/ppc/dsputil_altivec.c b/libavcodec/ppc/dsputil_altivec.c index 9111d036a5..adce61b277 100644 --- a/libavcodec/ppc/dsputil_altivec.c +++ b/libavcodec/ppc/dsputil_altivec.c @@ -1384,6 +1384,8 @@ static void avg_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels, int l void dsputil_init_altivec(DSPContext* c, AVCodecContext *avctx) { + const int high_bit_depth = avctx->codec_id == CODEC_ID_H264 && avctx->bits_per_raw_sample > 8; + c->pix_abs[0][1] = sad16_x2_altivec; c->pix_abs[0][2] = sad16_y2_altivec; c->pix_abs[0][3] = sad16_xy2_altivec; @@ -1397,8 +1399,10 @@ void dsputil_init_altivec(DSPContext* c, AVCodecContext *avctx) c->pix_sum = pix_sum_altivec; c->diff_pixels = diff_pixels_altivec; c->get_pixels = get_pixels_altivec; + if (!high_bit_depth) c->clear_block = clear_block_altivec; c->add_bytes= add_bytes_altivec; + if (!high_bit_depth) { c->put_pixels_tab[0][0] = put_pixels16_altivec; /* the two functions do the same thing, so use the same code */ c->put_no_rnd_pixels_tab[0][0] = put_pixels16_altivec; @@ -1409,6 +1413,7 @@ void dsputil_init_altivec(DSPContext* c, AVCodecContext *avctx) c->put_no_rnd_pixels_tab[1][3] = put_no_rnd_pixels8_xy2_altivec; c->put_pixels_tab[0][3] = put_pixels16_xy2_altivec; c->put_no_rnd_pixels_tab[0][3] = put_no_rnd_pixels16_xy2_altivec; + } c->hadamard8_diff[0] = hadamard8_diff16_altivec; c->hadamard8_diff[1] = hadamard8_diff8x8_altivec; diff --git a/libavcodec/ppc/dsputil_ppc.c b/libavcodec/ppc/dsputil_ppc.c index c52ea61805..5f131f3196 100644 --- a/libavcodec/ppc/dsputil_ppc.c +++ b/libavcodec/ppc/dsputil_ppc.c @@ -153,8 +153,11 @@ static void prefetch_ppc(void *mem, int stride, int h) void dsputil_init_ppc(DSPContext* c, AVCodecContext *avctx) { + const int high_bit_depth = avctx->codec_id == CODEC_ID_H264 && avctx->bits_per_raw_sample > 8; + // Common optimizations whether AltiVec is available or not c->prefetch = prefetch_ppc; + if (!high_bit_depth) { switch (check_dcbzl_effect()) { case 32: c->clear_blocks = clear_blocks_dcbz32_ppc; @@ -165,6 +168,7 @@ void dsputil_init_ppc(DSPContext* c, AVCodecContext *avctx) default: break; } + } #if HAVE_ALTIVEC if(CONFIG_H264_DECODER) dsputil_h264_init_ppc(c, avctx); diff --git a/libavcodec/ppc/h264_altivec.c b/libavcodec/ppc/h264_altivec.c index 02699beef4..fae0674720 100644 --- a/libavcodec/ppc/h264_altivec.c +++ b/libavcodec/ppc/h264_altivec.c @@ -965,8 +965,10 @@ H264_WEIGHT( 8, 8) H264_WEIGHT( 8, 4) void dsputil_h264_init_ppc(DSPContext* c, AVCodecContext *avctx) { + const int high_bit_depth = avctx->codec_id == CODEC_ID_H264 && avctx->bits_per_raw_sample > 8; if (av_get_cpu_flags() & AV_CPU_FLAG_ALTIVEC) { + if (!high_bit_depth) { c->put_h264_chroma_pixels_tab[0] = put_h264_chroma_mc8_altivec; c->avg_h264_chroma_pixels_tab[0] = avg_h264_chroma_mc8_altivec; @@ -992,11 +994,13 @@ void dsputil_h264_init_ppc(DSPContext* c, AVCodecContext *avctx) { dspfunc(avg_h264_qpel, 0, 16); #undef dspfunc } + } } -void ff_h264dsp_init_ppc(H264DSPContext *c) +void ff_h264dsp_init_ppc(H264DSPContext *c, const int bit_depth) { if (av_get_cpu_flags() & AV_CPU_FLAG_ALTIVEC) { + if (bit_depth == 8) { c->h264_idct_add = ff_h264_idct_add_altivec; c->h264_idct_add8 = ff_h264_idct_add8_altivec; c->h264_idct_add16 = ff_h264_idct_add16_altivec; @@ -1019,4 +1023,5 @@ void ff_h264dsp_init_ppc(H264DSPContext *c) c->biweight_h264_pixels_tab[3] = ff_biweight_h264_pixels8x8_altivec; c->biweight_h264_pixels_tab[4] = ff_biweight_h264_pixels8x4_altivec; } + } } -- cgit v1.2.3