*/
/**
- * @file libavcodec/dsputil.h
+ * @file
* DSP utils.
* note, many functions in here may use MMX which trashes the FPU state, it is
* absolutely necessary to call emms_c() between dsp & float/double code
//#define DEBUG
/* dct code */
typedef short DCTELEM;
-typedef int DWTELEM;
-typedef short IDWTELEM;
void fdct_ifast (DCTELEM *data);
void fdct_ifast248 (DCTELEM *data);
void ff_h264_idct8_add4_c(uint8_t *dst, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
void ff_h264_idct_add8_c(uint8_t **dest, const int *blockoffset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]);
-void ff_vector_fmul_window_c(float *dst, const float *src0, const float *src1,
- const float *win, float add_bias, int len);
-void ff_float_to_int16_c(int16_t *dst, const float *src, long len);
-void ff_float_to_int16_interleave_c(int16_t *dst, const float **src, long len, int channels);
+void ff_h264_luma_dc_dequant_idct_c(DCTELEM *output, DCTELEM *input, int qmul);
+void ff_svq3_luma_dc_dequant_idct_c(DCTELEM *output, DCTELEM *input, int qp);
+void ff_svq3_add_idct_c(uint8_t *dst, DCTELEM *block, int stride, int qp, int dc);
/* encoding scans */
extern const uint8_t ff_alternate_horizontal_scan[64];
extern uint32_t ff_squareTbl[512];
extern uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP];
+void ff_put_pixels8x8_c(uint8_t *dst, uint8_t *src, int stride);
+void ff_avg_pixels8x8_c(uint8_t *dst, uint8_t *src, int stride);
+void ff_put_pixels16x16_c(uint8_t *dst, uint8_t *src, int stride);
+void ff_avg_pixels16x16_c(uint8_t *dst, uint8_t *src, int stride);
+
/* VP3 DSP functions */
void ff_vp3_idct_c(DCTELEM *block/* align 16*/);
void ff_vp3_idct_put_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
void ff_vp3_idct_add_c(uint8_t *dest/*align 8*/, int line_size, DCTELEM *block/*align 16*/);
+void ff_vp3_idct_dc_add_c(uint8_t *dest/*align 8*/, int line_size, const DCTELEM *block/*align 16*/);
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);
-/* VP6 DSP functions */
-void ff_vp6_filter_diag4_c(uint8_t *dst, uint8_t *src, int stride,
- const int16_t *h_weights, const int16_t *v_weights);
-
/* 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);
-/* CAVS functions */
-void ff_put_cavs_qpel8_mc00_c(uint8_t *dst, uint8_t *src, int stride);
-void ff_avg_cavs_qpel8_mc00_c(uint8_t *dst, uint8_t *src, int stride);
-void ff_put_cavs_qpel16_mc00_c(uint8_t *dst, uint8_t *src, int stride);
-void ff_avg_cavs_qpel16_mc00_c(uint8_t *dst, uint8_t *src, int stride);
-
-/* VC1 functions */
-void ff_put_vc1_mspel_mc00_c(uint8_t *dst, const uint8_t *src, int stride, int rnd);
-void ff_avg_vc1_mspel_mc00_c(uint8_t *dst, const uint8_t *src, int stride, int rnd);
-
/* EA functions */
void ff_ea_idct_put_c(uint8_t *dest, int linesize, DCTELEM *block);
/* 1/2^n downscaling functions from imgconvert.c */
+#if LIBAVCODEC_VERSION_MAJOR < 53
+/**
+ * @deprecated Use av_image_copy_plane() instead.
+ */
+attribute_deprecated
void ff_img_copy_plane(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
+#endif
+
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);
void ff_shrink88(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
typedef void (*tpel_mc_func)(uint8_t *block/*align width (8 or 16)*/, const uint8_t *pixels/*align 1*/, int line_size, int w, int h);
typedef void (*qpel_mc_func)(uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);
typedef void (*h264_chroma_mc_func)(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int srcStride, int h, int x, int y);
-typedef void (*h264_weight_func)(uint8_t *block, int stride, int log2_denom, int weight, int offset);
-typedef void (*h264_biweight_func)(uint8_t *dst, uint8_t *src, int stride, int log2_denom, int weightd, int weights, int offset);
typedef void (*op_fill_func)(uint8_t *block/*align width (8 or 16)*/, uint8_t value, int line_size, int h);
// although currently h<4 is not used as functions with width <8 are neither used nor implemented
typedef int (*me_cmp_func)(void /*MpegEncContext*/ *s, uint8_t *blk1/*align width (8 or 16)*/, uint8_t *blk2/*align 1*/, int line_size, int h)/* __attribute__ ((const))*/;
-
-// for snow slices
-typedef struct slice_buffer_s slice_buffer;
-
/**
* Scantable.
*/
void ff_init_scantable(uint8_t *, ScanTable *st, const uint8_t *src_scantable);
-void ff_emulated_edge_mc(uint8_t *buf, uint8_t *src, int linesize,
+void ff_emulated_edge_mc(uint8_t *buf, const uint8_t *src, int linesize,
int block_w, int block_h,
int src_x, int src_y, int w, int h);
void (*add_pixels8)(uint8_t *pixels, DCTELEM *block, int line_size);
void (*add_pixels4)(uint8_t *pixels, DCTELEM *block, int line_size);
int (*sum_abs_dctelem)(DCTELEM *block/*align 16*/);
+ /**
+ * Motion estimation with emulated edge values.
+ * @param buf pointer to destination buffer (unaligned)
+ * @param src pointer to pixel source (unaligned)
+ * @param linesize width (in pixels) for src/buf
+ * @param block_w number of pixels (per row) to copy to buf
+ * @param block_h nummber of pixel rows to copy to buf
+ * @param src_x offset of src to start of row - this may be negative
+ * @param src_y offset of src to top of image - this may be negative
+ * @param w width of src in pixels
+ * @param h height of src in pixels
+ */
+ void (*emulated_edge_mc)(uint8_t *buf, const uint8_t *src, int linesize,
+ int block_w, int block_h,
+ int src_x, int src_y, int w, int h);
/**
* translational global motion compensation.
*/
qpel_mc_func put_2tap_qpel_pixels_tab[4][16];
qpel_mc_func avg_2tap_qpel_pixels_tab[4][16];
- h264_weight_func weight_h264_pixels_tab[10];
- h264_biweight_func biweight_h264_pixels_tab[10];
-
- /* AVS specific */
- qpel_mc_func put_cavs_qpel_pixels_tab[2][16];
- qpel_mc_func avg_cavs_qpel_pixels_tab[2][16];
- void (*cavs_filter_lv)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
- void (*cavs_filter_lh)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
- void (*cavs_filter_cv)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
- void (*cavs_filter_ch)(uint8_t *pix, int stride, int alpha, int beta, int tc, int bs1, int bs2);
- void (*cavs_idct8_add)(uint8_t *dst, DCTELEM *block, int stride);
-
me_cmp_func pix_abs[2][4];
/* huffyuv specific */
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 (*h264_v_loop_filter_luma)(uint8_t *pix/*align 16*/, int stride, int alpha, int beta, int8_t *tc0);
- void (*h264_h_loop_filter_luma)(uint8_t *pix/*align 4 */, int stride, int alpha, int beta, int8_t *tc0);
- /* v/h_loop_filter_luma_intra: align 16 */
- void (*h264_v_loop_filter_luma_intra)(uint8_t *pix, int stride, int alpha, int beta);
- void (*h264_h_loop_filter_luma_intra)(uint8_t *pix, int stride, int alpha, int beta);
- void (*h264_v_loop_filter_chroma)(uint8_t *pix/*align 8*/, int stride, int alpha, int beta, int8_t *tc0);
- void (*h264_h_loop_filter_chroma)(uint8_t *pix/*align 4*/, int stride, int alpha, int beta, int8_t *tc0);
- void (*h264_v_loop_filter_chroma_intra)(uint8_t *pix/*align 8*/, int stride, int alpha, int beta);
- void (*h264_h_loop_filter_chroma_intra)(uint8_t *pix/*align 8*/, int stride, int alpha, int beta);
- // h264_loop_filter_strength: simd only. the C version is inlined in h264.c
- void (*h264_loop_filter_strength)(int16_t bS[2][4][4], uint8_t nnz[40], int8_t ref[2][40], int16_t mv[2][40][2],
- int bidir, int edges, int step, int mask_mv0, int mask_mv1, int field);
-
void (*h263_v_loop_filter)(uint8_t *src, int stride, int qscale);
void (*h263_h_loop_filter)(uint8_t *src, int stride, int qscale);
void (*x8_v_loop_filter)(uint8_t *src, int stride, int qscale);
void (*x8_h_loop_filter)(uint8_t *src, int stride, int qscale);
+ void (*vp3_idct_dc_add)(uint8_t *dest/*align 8*/, int line_size, const DCTELEM *block/*align 16*/);
void (*vp3_v_loop_filter)(uint8_t *src, int stride, int *bounding_values);
void (*vp3_h_loop_filter)(uint8_t *src, int stride, int *bounding_values);
- void (*vp6_filter_diag4)(uint8_t *dst, uint8_t *src, int stride,
- const int16_t *h_weights,const int16_t *v_weights);
-
/* assume len is a multiple of 4, and arrays are 16-byte aligned */
void (*vorbis_inverse_coupling)(float *mag, float *ang, int blocksize);
void (*ac3_downmix)(float (*samples)[256], float (*matrix)[2], int out_ch, int in_ch, int len);
- /* no alignment needed */
- void (*lpc_compute_autocorr)(const int32_t *data, int len, int lag, double *autoc);
/* assume len is a multiple of 8, and arrays are 16-byte aligned */
- void (*vector_fmul)(float *dst, const float *src, int len);
+ void (*vector_fmul)(float *dst, const float *src0, const float *src1, int len);
void (*vector_fmul_reverse)(float *dst, const float *src0, const float *src1, int len);
/* assume len is a multiple of 8, and src arrays are 16-byte aligned */
void (*vector_fmul_add)(float *dst, const float *src0, const float *src1, const float *src2, int len);
/* assume len is a multiple of 4, and arrays are 16-byte aligned */
- void (*vector_fmul_window)(float *dst, const float *src0, const float *src1, const float *win, float add_bias, int len);
+ void (*vector_fmul_window)(float *dst, const float *src0, const float *src1, const float *win, int len);
/* assume len is a multiple of 8, and arrays are 16-byte aligned */
- void (*int32_to_float_fmul_scalar)(float *dst, const int *src, float mul, int len);
void (*vector_clipf)(float *dst /* align 16 */, const float *src /* align 16 */, float min, float max, int len /* align 16 */);
/**
* Multiply a vector of floats by a scalar float. Source and
*/
void (*butterflies_float)(float *restrict v1, float *restrict v2, int len);
- /* C version: convert floats from the range [384.0,386.0] to ints in [-32768,32767]
- * simd versions: convert floats from [-32768.0,32767.0] without rescaling and arrays are 16byte aligned */
- void (*float_to_int16)(int16_t *dst, const float *src, long len);
- void (*float_to_int16_interleave)(int16_t *dst, const float **src, long len, int channels);
-
/* (I)DCT */
void (*fdct)(DCTELEM *block/* align 16*/);
void (*fdct248)(DCTELEM *block/* align 16*/);
void (*draw_edges)(uint8_t *buf, int wrap, int width, int height, int w);
#define EDGE_WIDTH 16
- /* h264 functions */
- /* NOTE!!! if you implement any of h264_idct8_add, h264_idct8_add4 then you must implement all of them
- NOTE!!! if you implement any of h264_idct_add, h264_idct_add16, h264_idct_add16intra, h264_idct_add8 then you must implement all of them
- The reason for above, is that no 2 out of one list may use a different permutation.
- */
- void (*h264_idct_add)(uint8_t *dst/*align 4*/, DCTELEM *block/*align 16*/, int stride);
- void (*h264_idct8_add)(uint8_t *dst/*align 8*/, DCTELEM *block/*align 16*/, int stride);
- void (*h264_idct_dc_add)(uint8_t *dst/*align 4*/, DCTELEM *block/*align 16*/, int stride);
- void (*h264_idct8_dc_add)(uint8_t *dst/*align 8*/, DCTELEM *block/*align 16*/, int stride);
- void (*h264_dct)(DCTELEM block[4][4]);
- void (*h264_idct_add16)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
- void (*h264_idct8_add4)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
- void (*h264_idct_add8)(uint8_t **dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
- void (*h264_idct_add16intra)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
-
- /* snow wavelet */
- void (*vertical_compose97i)(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, IDWTELEM *b3, IDWTELEM *b4, IDWTELEM *b5, int width);
- void (*horizontal_compose97i)(IDWTELEM *b, int width);
- void (*inner_add_yblock)(const uint8_t *obmc, const int obmc_stride, uint8_t * * block, int b_w, int b_h, int src_x, int src_y, int src_stride, slice_buffer * sb, int add, uint8_t * dst8);
-
void (*prefetch)(void *mem, int stride, int h);
void (*shrink[4])(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);
* @param len length of vectors, should be multiple of 16
* @param shift number of bits to discard from product
*/
- int32_t (*scalarproduct_int16)(int16_t *v1, int16_t *v2/*align 16*/, int len, int shift);
+ int32_t (*scalarproduct_int16)(const int16_t *v1, const int16_t *v2/*align 16*/, int len, int shift);
/* ape functions */
/**
* Calculate scalar product of v1 and v2,
* and v1[i] += v3[i] * mul
* @param len length of vectors, should be multiple of 16
*/
- int32_t (*scalarproduct_and_madd_int16)(int16_t *v1/*align 16*/, int16_t *v2, int16_t *v3, int len, int mul);
+ int32_t (*scalarproduct_and_madd_int16)(int16_t *v1/*align 16*/, const int16_t *v2, const int16_t *v3, int len, int mul);
/* rv30 functions */
qpel_mc_func put_rv30_tpel_pixels_tab[4][16];
*/
#define emms_c()
-/* should be defined by architectures supporting
- one or more MultiMedia extension */
-int mm_support(void);
-extern int mm_flags;
-
void dsputil_init_alpha(DSPContext* c, AVCodecContext *avctx);
void dsputil_init_arm(DSPContext* c, AVCodecContext *avctx);
void dsputil_init_bfin(DSPContext* c, AVCodecContext *avctx);
void dsputil_init_sh4(DSPContext* c, AVCodecContext *avctx);
void dsputil_init_vis(DSPContext* c, AVCodecContext *avctx);
-void ff_cavsdsp_init(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_vc1dsp_init(DSPContext* c, AVCodecContext *avctx);
__asm__ volatile ("emms;":::"memory");
}
-
-#define emms_c() \
-{\
- if (mm_flags & FF_MM_MMX)\
- emms();\
-}
+#define emms_c() emms()
#elif ARCH_ARM
#define STRIDE_ALIGN 16
-#else
-
-#define mm_flags 0
-#define mm_support() 0
-
#endif
#ifndef STRIDE_ALIGN
# define STRIDE_ALIGN 8
#endif
-#define LOCAL_ALIGNED(a, t, v, s, ...) \
- uint8_t la_##v[sizeof(t s __VA_ARGS__) + (a)]; \
- t (*v) __VA_ARGS__ = (void *)FFALIGN((uintptr_t)la_##v, a)
+#define LOCAL_ALIGNED_A(a, t, v, s, o, ...) \
+ uint8_t la_##v[sizeof(t s o) + (a)]; \
+ t (*v) o = (void *)FFALIGN((uintptr_t)la_##v, a)
+
+#define LOCAL_ALIGNED_D(a, t, v, s, o, ...) DECLARE_ALIGNED(a, t, v) s o
+
+#define LOCAL_ALIGNED(a, t, v, ...) LOCAL_ALIGNED_A(a, t, v, __VA_ARGS__,,)
#if HAVE_LOCAL_ALIGNED_8
-# define LOCAL_ALIGNED_8(t, v, s, ...) DECLARE_ALIGNED(8, t, v) s __VA_ARGS__
+# define LOCAL_ALIGNED_8(t, v, ...) LOCAL_ALIGNED_D(8, t, v, __VA_ARGS__,,)
#else
-# define LOCAL_ALIGNED_8(t, v, s, ...) LOCAL_ALIGNED(8, t, v, s, __VA_ARGS__)
+# define LOCAL_ALIGNED_8(t, v, ...) LOCAL_ALIGNED(8, t, v, __VA_ARGS__)
#endif
#if HAVE_LOCAL_ALIGNED_16
-# define LOCAL_ALIGNED_16(t, v, s, ...) DECLARE_ALIGNED(16, t, v) s __VA_ARGS__
+# define LOCAL_ALIGNED_16(t, v, ...) LOCAL_ALIGNED_D(16, t, v, __VA_ARGS__,,)
#else
-# define LOCAL_ALIGNED_16(t, v, s, ...) LOCAL_ALIGNED(16, t, v, s, __VA_ARGS__)
+# define LOCAL_ALIGNED_16(t, v, ...) LOCAL_ALIGNED(16, t, v, __VA_ARGS__)
#endif
/* PSNR */
projects / ffmpeg / blobdiff