fix conversions between vectors with differing element types or numbers of subparts...

[x264] / common / common.h

diff --git a/common/common.h b/common/common.h
index 1a260f9d8cea1ac13d37ae7f2df52d94e1c345cf..aa944f3a66e4a57c415a161f63664fe3f8f6bd72 100644 (file)
--- a/common/common.h
+++ b/common/common.h
@@ -99,6 +99,7 @@ char *x264_param2string( x264_param_t *p, int b_res );
 void x264_log( x264_t *h, int i_level, const char *psz_fmt, ... );
 
 void x264_reduce_fraction( int *n, int *d );
+void x264_init_vlc_tables();
 
 static inline uint8_t x264_clip_uint8( int x )
 {
@@ -213,7 +214,7 @@ typedef struct
 #define X264_SCAN8_SIZE (6*8)
 #define X264_SCAN8_0 (4+1*8)
 
-static const int x264_scan8[16+2*4] =
+static const int x264_scan8[16+2*4+3] =
 {
     /* Luma */
     4+1*8, 5+1*8, 4+2*8, 5+2*8,
@@ -228,6 +229,12 @@ static const int x264_scan8[16+2*4] =
     /* Cr */
     1+4*8, 2+4*8,
     1+5*8, 2+5*8,
+
+    /* Luma DC */
+    4+5*8,
+
+    /* Chroma DC */
+    5+5*8, 6+5*8
 };
 /*
    0 1 2 3 4 5 6 7
@@ -236,7 +243,7 @@ static const int x264_scan8[16+2*4] =
  2   B B   L L L L
  3         L L L L
  4   R R   L L L L
- 5   R R
+ 5   R R   DyDuDv
 */
 
 typedef struct x264_ratecontrol_t   x264_ratecontrol_t;
@@ -331,6 +338,7 @@ struct x264_t
         int i_max_ref1;
         int i_delay;    /* Number of frames buffered for B reordering */
         int b_have_lowres;  /* Whether 1/2 resolution luma planes are being used */
+        int b_have_sub8x8_esa;
     } frames;
 
     /* current frame being encoded */
@@ -432,7 +440,7 @@ struct x264_t
         /* current value */
         int     i_type;
         int     i_partition;
-        int     i_sub_partition[4];
+        DECLARE_ALIGNED_4( uint8_t i_sub_partition[4] );
         int     b_transform_8x8;
 
         int     i_cbp_luma;
@@ -463,6 +471,10 @@ struct x264_t
             DECLARE_ALIGNED_16( uint8_t i8x8_fdec_buf[16*16] );
             DECLARE_ALIGNED_16( int16_t i8x8_dct_buf[3][64] );
             DECLARE_ALIGNED_16( int16_t i4x4_dct_buf[15][16] );
+            uint32_t i4x4_nnz_buf[4];
+            uint32_t i8x8_nnz_buf[4];
+            int i4x4_cbp;
+            int i8x8_cbp;
 
             /* Psy trellis DCT data */
             DECLARE_ALIGNED_16( int16_t fenc_dct8[4][64] );
@@ -517,6 +529,10 @@ struct x264_t
             /* number of neighbors (top and left) that used 8x8 dct */
             int     i_neighbour_transform_size;
             int     i_neighbour_interlaced;
+
+            /* neighbor CBPs */
+            int     i_cbp_top;
+            int     i_cbp_left;
         } cache;
 
         /* */
@@ -560,11 +576,7 @@ struct x264_t
             int i_mb_count_8x8dct[2];
             int i_mb_count_ref[2][32];
             int i_mb_partition[17];
-            /* Estimated (SATD) cost as Intra/Predicted frame */
-            /* XXX: both omit the cost of MBs coded as P_SKIP */
-            int i_intra_cost;
-            int i_inter_cost;
-            int i_mbs_analysed;
+            int i_mb_cbp[6];
             /* Adaptive direct mv pred */
             int i_direct_score[2];
             /* Metrics */
@@ -591,17 +603,21 @@ struct x264_t
         int64_t i_mb_partition[2][17];
         int64_t i_mb_count_8x8dct[2];
         int64_t i_mb_count_ref[2][2][32];
+        int64_t i_mb_cbp[6];
         /* */
         int     i_direct_score[2];
         int     i_direct_frames[2];
 
     } stat;
 
+    void *scratch_buffer; /* for any temporary storage that doesn't want repeated malloc */
+
     /* CPU functions dependents */
     x264_predict_t      predict_16x16[4+3];
     x264_predict_t      predict_8x8c[4+3];
     x264_predict8x8_t   predict_8x8[9+3];
     x264_predict_t      predict_4x4[9+3];
+    x264_predict_8x8_filter_t predict_8x8_filter;
 
     x264_pixel_function_t pixf;
     x264_mc_functions_t   mc;