diff options
Diffstat (limited to 'libavcodec/dsputil.h')
-rw-r--r-- | libavcodec/dsputil.h | 88 |
1 files changed, 33 insertions, 55 deletions
diff --git a/libavcodec/dsputil.h b/libavcodec/dsputil.h index ef2956eecb..22c51a0962 100644 --- a/libavcodec/dsputil.h +++ b/libavcodec/dsputil.h @@ -3,20 +3,20 @@ * Copyright (c) 2000, 2001, 2002 Fabrice Bellard * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at> * - * This file is part of Libav. + * This file is part of FFmpeg. * - * Libav is free software; you can redistribute it and/or + * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * - * Libav is distributed in the hope that it will be useful, + * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public - * License along with Libav; if not, write to the Free Software + * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA */ @@ -40,8 +40,10 @@ typedef short DCTELEM; void fdct_ifast (DCTELEM *data); void fdct_ifast248 (DCTELEM *data); -void ff_jpeg_fdct_islow (DCTELEM *data); -void ff_fdct248_islow (DCTELEM *data); +void ff_jpeg_fdct_islow_8(DCTELEM *data); +void ff_jpeg_fdct_islow_10(DCTELEM *data); +void ff_fdct248_islow_8(DCTELEM *data); +void ff_fdct248_islow_10(DCTELEM *data); void j_rev_dct (DCTELEM *data); void j_rev_dct4 (DCTELEM *data); @@ -58,13 +60,13 @@ void ff_h264_idct8_add_ ## depth ## _c(uint8_t *dst, DCTELEM *block, int stride) void ff_h264_idct_add_ ## depth ## _c(uint8_t *dst, DCTELEM *block, int stride);\ void ff_h264_idct8_dc_add_ ## depth ## _c(uint8_t *dst, DCTELEM *block, int stride);\ void ff_h264_idct_dc_add_ ## depth ## _c(uint8_t *dst, DCTELEM *block, int stride);\ -void ff_h264_lowres_idct_add_ ## depth ## _c(uint8_t *dst, int stride, DCTELEM *block);\ -void ff_h264_lowres_idct_put_ ## depth ## _c(uint8_t *dst, int stride, DCTELEM *block);\ void ff_h264_idct_add16_ ## depth ## _c(uint8_t *dst, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);\ void ff_h264_idct_add16intra_ ## depth ## _c(uint8_t *dst, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);\ void ff_h264_idct8_add4_ ## depth ## _c(uint8_t *dst, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);\ +void ff_h264_idct_add8_422_ ## depth ## _c(uint8_t **dest, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);\ void ff_h264_idct_add8_ ## depth ## _c(uint8_t **dest, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);\ void ff_h264_luma_dc_dequant_idct_ ## depth ## _c(DCTELEM *output, DCTELEM *input, int qmul);\ +void ff_h264_chroma422_dc_dequant_idct_ ## depth ## _c(DCTELEM *block, int qmul);\ void ff_h264_chroma_dc_dequant_idct_ ## depth ## _c(DCTELEM *block, int qmul); H264_IDCT( 8) @@ -111,14 +113,15 @@ void ff_vp3_idct_dc_add_c(uint8_t *dest/*align 8*/, int line_size, const DCTELEM void ff_vp3_v_loop_filter_c(uint8_t *src, int stride, int *bounding_values); void ff_vp3_h_loop_filter_c(uint8_t *src, int stride, int *bounding_values); -/* Bink functions */ -void ff_bink_idct_c (DCTELEM *block); -void ff_bink_idct_add_c(uint8_t *dest, int linesize, DCTELEM *block); -void ff_bink_idct_put_c(uint8_t *dest, int linesize, DCTELEM *block); - /* EA functions */ void ff_ea_idct_put_c(uint8_t *dest, int linesize, DCTELEM *block); +/* RV40 functions */ +void ff_put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride); +void ff_avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride); +void ff_put_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride); +void ff_avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride); + /* 1/2^n downscaling functions from imgconvert.c */ void ff_shrink22(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height); void ff_shrink44(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height); @@ -130,7 +133,7 @@ void ff_gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy, /* minimum alignment rules ;) If you notice errors in the align stuff, need more alignment for some ASM code for some CPU or need to use a function with less aligned data then send a mail -to the libav-devel mailing list, ... +to the ffmpeg-devel mailing list, ... !warning These alignments might not match reality, (missing attribute((align)) stuff somewhere possible). @@ -201,6 +204,8 @@ typedef struct ScanTable{ } ScanTable; void ff_init_scantable(uint8_t *, ScanTable *st, const uint8_t *src_scantable); +void ff_init_scantable_permutation(uint8_t *idct_permutation, + int idct_permutation_type); #define EMULATED_EDGE(depth) \ void ff_emulated_edge_mc_ ## depth (uint8_t *buf, const uint8_t *src, int linesize,\ @@ -211,8 +216,6 @@ EMULATED_EDGE(8) EMULATED_EDGE(9) EMULATED_EDGE(10) -#define ff_emulated_edge_mc ff_emulated_edge_mc_8 - void ff_add_pixels_clamped_c(const DCTELEM *block, uint8_t *dest, int linesize); void ff_put_pixels_clamped_c(const DCTELEM *block, uint8_t *dest, int linesize); void ff_put_signed_pixels_clamped_c(const DCTELEM *block, uint8_t *dest, int linesize); @@ -221,12 +224,16 @@ void ff_put_signed_pixels_clamped_c(const DCTELEM *block, uint8_t *dest, int lin * DSPContext. */ typedef struct DSPContext { + /** + * Size of DCT coefficients. + */ + int dct_bits; + /* pixel ops : interface with DCT */ void (*get_pixels)(DCTELEM *block/*align 16*/, const uint8_t *pixels/*align 8*/, int line_size); void (*diff_pixels)(DCTELEM *block/*align 16*/, const uint8_t *s1/*align 8*/, const uint8_t *s2/*align 8*/, int stride); void (*put_pixels_clamped)(const DCTELEM *block/*align 16*/, uint8_t *pixels/*align 8*/, int line_size); void (*put_signed_pixels_clamped)(const DCTELEM *block/*align 16*/, uint8_t *pixels/*align 8*/, int line_size); - void (*put_pixels_nonclamped)(const DCTELEM *block/*align 16*/, uint8_t *pixels/*align 8*/, int line_size); void (*add_pixels_clamped)(const DCTELEM *block/*align 16*/, uint8_t *pixels/*align 8*/, int line_size); void (*add_pixels8)(uint8_t *pixels, DCTELEM *block, int line_size); void (*add_pixels4)(uint8_t *pixels, DCTELEM *block, int line_size); @@ -371,7 +378,6 @@ typedef struct DSPContext { /* huffyuv specific */ void (*add_bytes)(uint8_t *dst/*align 16*/, uint8_t *src/*align 16*/, int w); - void (*add_bytes_l2)(uint8_t *dst/*align 16*/, uint8_t *src1/*align 16*/, uint8_t *src2/*align 16*/, int w); void (*diff_bytes)(uint8_t *dst/*align 16*/, uint8_t *src1/*align 16*/, uint8_t *src2/*align 1*/,int w); /** * subtract huffyuv's variant of median prediction @@ -382,7 +388,6 @@ typedef struct DSPContext { int (*add_hfyu_left_prediction)(uint8_t *dst, const uint8_t *src, int w, int left); void (*add_hfyu_left_prediction_bgr32)(uint8_t *dst, const uint8_t *src, int w, int *red, int *green, int *blue, int *alpha); /* this might write to dst[w] */ - void (*add_png_paeth_prediction)(uint8_t *dst, uint8_t *src, uint8_t *top, int w, int bpp); void (*bswap_buf)(uint32_t *dst, const uint32_t *src, int w); void (*bswap16_buf)(uint16_t *dst, const uint16_t *src, int len); @@ -421,31 +426,16 @@ typedef struct DSPContext { void (*vector_fmul_scalar)(float *dst, const float *src, float mul, int len); /** - * Multiply a vector of floats by concatenated short vectors of - * floats and by a scalar float. Source and destination vectors - * must overlap exactly or not at all. - * [0]: short vectors of length 2, 8-byte aligned - * [1]: short vectors of length 4, 16-byte aligned - * @param dst output vector, 16-byte aligned + * Multiply a vector of floats by a scalar float and add to + * destination vector. Source and destination vectors must + * overlap exactly or not at all. + * @param dst result vector, 16-byte aligned * @param src input vector, 16-byte aligned - * @param sv array of pointers to short vectors - * @param mul scalar value - * @param len number of elements in src and dst, multiple of 4 - */ - void (*vector_fmul_sv_scalar[2])(float *dst, const float *src, - const float **sv, float mul, int len); - /** - * Multiply short vectors of floats by a scalar float, store - * concatenated result. - * [0]: short vectors of length 2, 8-byte aligned - * [1]: short vectors of length 4, 16-byte aligned - * @param dst output vector, 16-byte aligned - * @param sv array of pointers to short vectors * @param mul scalar value - * @param len number of output elements, multiple of 4 + * @param len length of vector, multiple of 4 */ - void (*sv_fmul_scalar[2])(float *dst, const float **sv, - float mul, int len); + void (*vector_fmac_scalar)(float *dst, const float *src, float mul, + int len); /** * Calculate the scalar product of two vectors of floats. * @param v1 first vector, 16-byte aligned @@ -571,19 +561,7 @@ typedef struct DSPContext { void (*vector_clip_int32)(int32_t *dst, const int32_t *src, int32_t min, int32_t max, unsigned int len); - /* rv30 functions */ - qpel_mc_func put_rv30_tpel_pixels_tab[4][16]; - qpel_mc_func avg_rv30_tpel_pixels_tab[4][16]; - - /* rv40 functions */ - qpel_mc_func put_rv40_qpel_pixels_tab[4][16]; - qpel_mc_func avg_rv40_qpel_pixels_tab[4][16]; - h264_chroma_mc_func put_rv40_chroma_pixels_tab[3]; - h264_chroma_mc_func avg_rv40_chroma_pixels_tab[3]; - - /* bink functions */ op_fill_func fill_block_tab[2]; - void (*scale_block)(const uint8_t src[64]/*align 8*/, uint8_t *dst/*align 8*/, int linesize); } DSPContext; void dsputil_static_init(void); @@ -657,14 +635,14 @@ void dsputil_init_sh4(DSPContext* c, AVCodecContext *avctx); void dsputil_init_vis(DSPContext* c, AVCodecContext *avctx); void ff_dsputil_init_dwt(DSPContext *c); -void ff_rv30dsp_init(DSPContext* c, AVCodecContext *avctx); -void ff_rv40dsp_init(DSPContext* c, AVCodecContext *avctx); void ff_intrax8dsp_init(DSPContext* c, AVCodecContext *avctx); void ff_mlp_init(DSPContext* c, AVCodecContext *avctx); void ff_mlp_init_x86(DSPContext* c, AVCodecContext *avctx); + #if ARCH_ARM + #if HAVE_NEON # define STRIDE_ALIGN 16 #endif |