huffyuv: switch to encode2().

[ffmpeg] / libavcodec / dsputil.h

diff --git a/libavcodec/dsputil.h b/libavcodec/dsputil.h
index 73830f819012459372ca5ec0a9f8f04e30a65a5e..0a6165685eecf363c75b67e71e975e4cec5b144e 100644 (file)
--- a/libavcodec/dsputil.h
+++ b/libavcodec/dsputil.h
@@ -38,17 +38,17 @@
 /* dct code */
 typedef short DCTELEM;
 
-void fdct_ifast (DCTELEM *data);
-void fdct_ifast248 (DCTELEM *data);
+void ff_fdct_ifast (DCTELEM *data);
+void ff_fdct_ifast248 (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);
-void j_rev_dct2 (DCTELEM *data);
-void j_rev_dct1 (DCTELEM *data);
+void ff_j_rev_dct (DCTELEM *data);
+void ff_j_rev_dct4 (DCTELEM *data);
+void ff_j_rev_dct2 (DCTELEM *data);
+void ff_j_rev_dct1 (DCTELEM *data);
 void ff_wmv2_idct_c(DCTELEM *data);
 
 void ff_fdct_mmx(DCTELEM *block);
@@ -63,8 +63,10 @@ void ff_h264_idct_dc_add_ ## depth ## _c(uint8_t *dst, DCTELEM *block, int strid
 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)
@@ -202,6 +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,\
@@ -374,7 +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
@@ -384,8 +387,6 @@ typedef struct DSPContext {
     void (*add_hfyu_median_prediction)(uint8_t *dst, const uint8_t *top, const uint8_t *diff, int w, int *left, int *left_top);
     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);
 
@@ -449,6 +450,23 @@ typedef struct DSPContext {
      */
     void (*butterflies_float)(float *restrict v1, float *restrict v2, int len);
 
+    /**
+     * Calculate the sum and difference of two vectors of floats and interleave
+     * results into a separate output vector of floats, with each sum
+     * positioned before the corresponding difference.
+     *
+     * @param dst  output vector
+     *             constraints: 16-byte aligned
+     * @param src0 first input vector
+     *             constraints: 32-byte aligned
+     * @param src1 second input vector
+     *             constraints: 32-byte aligned
+     * @param len  number of elements in the input
+     *             constraints: multiple of 8
+     */
+    void (*butterflies_float_interleave)(float *dst, const float *src0,
+                                         const float *src1, int len);
+
     /* (I)DCT */
     void (*fdct)(DCTELEM *block/* align 16*/);
     void (*fdct248)(DCTELEM *block/* align 16*/);
@@ -477,8 +495,8 @@ typedef struct DSPContext {
      * with the zigzag/alternate scan<br>
      * an example to avoid confusion:
      * - (->decode coeffs -> zigzag reorder -> dequant -> reference idct ->...)
-     * - (x -> referece dct -> reference idct -> x)
-     * - (x -> referece dct -> simple_mmx_perm = idct_permutation -> simple_idct_mmx -> x)
+     * - (x -> reference dct -> reference idct -> x)
+     * - (x -> reference dct -> simple_mmx_perm = idct_permutation -> simple_idct_mmx -> x)
      * - (->decode coeffs -> zigzag reorder -> simple_mmx_perm -> dequant -> simple_idct_mmx ->...)
      */
     uint8_t idct_permutation[64];
@@ -562,8 +580,8 @@ typedef struct DSPContext {
     op_fill_func fill_block_tab[2];
 } DSPContext;
 
-void dsputil_static_init(void);
-void dsputil_init(DSPContext* p, AVCodecContext *avctx);
+void ff_dsputil_static_init(void);
+void ff_dsputil_init(DSPContext* p, AVCodecContext *avctx);
 
 int ff_check_alignment(void);
 
@@ -622,38 +640,23 @@ static inline int get_penalty_factor(int lambda, int lambda2, int type){
     }
 }
 
-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_mlib(DSPContext* c, AVCodecContext *avctx);
-void dsputil_init_mmi(DSPContext* c, AVCodecContext *avctx);
-void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx);
-void dsputil_init_ppc(DSPContext* c, AVCodecContext *avctx);
-void dsputil_init_sh4(DSPContext* c, AVCodecContext *avctx);
-void dsputil_init_vis(DSPContext* c, AVCodecContext *avctx);
+void ff_dsputil_init_alpha(DSPContext* c, AVCodecContext *avctx);
+void ff_dsputil_init_arm(DSPContext* c, AVCodecContext *avctx);
+void ff_dsputil_init_bfin(DSPContext* c, AVCodecContext *avctx);
+void ff_dsputil_init_mmi(DSPContext* c, AVCodecContext *avctx);
+void ff_dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx);
+void ff_dsputil_init_ppc(DSPContext* c, AVCodecContext *avctx);
+void ff_dsputil_init_sh4(DSPContext* c, AVCodecContext *avctx);
+void ff_dsputil_init_vis(DSPContext* c, AVCodecContext *avctx);
 
 void ff_dsputil_init_dwt(DSPContext *c);
 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
+#if (ARCH_ARM && HAVE_NEON) || ARCH_PPC || HAVE_MMI || HAVE_MMX
 #   define STRIDE_ALIGN 16
-#endif
-
-#elif ARCH_PPC
-
-#define STRIDE_ALIGN 16
-
-#elif HAVE_MMI
-
-#define STRIDE_ALIGN 16
-
-#endif
-
-#ifndef STRIDE_ALIGN
+#else
 #   define STRIDE_ALIGN 8
 #endif