DIRAC Decoder stable version, MMX support removed.

Look for MMX_DISABLED to find the disabled functions.

Authors of this code are Marco Gerards <marco@gnu.org> and David Conrad <lessen42@gmail.com>
With changes from Jordi Ortiz <nenjordi@gmail.com>

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

1 files changed, 92 insertions, 1 deletions

diff --git a/libavcodec/dwt.h b/libavcodec/dwt.h
index b10e4f5596..441d5e6ebb 100644
--- a/libavcodec/dwt.h
+++ b/libavcodec/dwt.h
@@ -26,7 +26,12 @@
 typedef int DWTELEM;
 typedef short IDWTELEM;
 
+#define MAX_DWT_SUPPORT 8
+#define MAX_DECOMPOSITIONS 8
+
 typedef struct {
+    IDWTELEM *b[MAX_DWT_SUPPORT];
+
     IDWTELEM *b0;
     IDWTELEM *b1;
     IDWTELEM *b2;
@@ -45,13 +50,99 @@ typedef struct slice_buffer_s {
     IDWTELEM * base_buffer; ///< Buffer that this structure is caching.
 } slice_buffer;
 
+struct DWTContext;
+
+// Possible prototypes for vertical_compose functions
+typedef void (*vertical_compose_2tap)(IDWTELEM *b0, IDWTELEM *b1, int width);
+typedef void (*vertical_compose_3tap)(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, int width);
+typedef void (*vertical_compose_5tap)(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, IDWTELEM *b3, IDWTELEM *b4, int width);
+typedef void (*vertical_compose_9tap)(IDWTELEM *dst, IDWTELEM *b[8], int width);
+
 typedef struct DWTContext {
+    IDWTELEM *buffer;
+    IDWTELEM *temp;
+    int width;
+    int height;
+    int stride;
+    int decomposition_count;
+    int support;
+
+    void (*spatial_compose)(struct DWTContext *cs, int level, int width, int height, int stride);
+    void (*vertical_compose_l0)();
+    void (*vertical_compose_h0)();
+    void (*vertical_compose_l1)();
+    void (*vertical_compose_h1)();
+    void (*vertical_compose)();     ///< one set of lowpass and highpass combined
+    void (*horizontal_compose)(IDWTELEM *b, IDWTELEM *tmp, int width);
+
     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);
+
+    DWTCompose cs[MAX_DECOMPOSITIONS];
 } DWTContext;
 
-#define MAX_DECOMPOSITIONS 8
+enum dwt_type {
+    DWT_SNOW_DAUB9_7,
+    DWT_SNOW_LEGALL5_3,
+    DWT_DIRAC_DD9_7,
+    DWT_DIRAC_LEGALL5_3,
+    DWT_DIRAC_DD13_7,
+    DWT_DIRAC_HAAR0,
+    DWT_DIRAC_HAAR1,
+    DWT_DIRAC_FIDELITY,
+    DWT_DIRAC_DAUB9_7,
+    DWT_NUM_TYPES
+};
+
+// -1 if an error occurred, e.g. the dwt_type isn't recognized
+int ff_spatial_idwt_init2(DWTContext *d, IDWTELEM *buffer, int width, int height,
+                          int stride, enum dwt_type type, int decomposition_count,
+                          IDWTELEM *temp);
+
+int ff_spatial_idwt2(IDWTELEM *buffer, int width, int height, int stride,
+                     enum dwt_type type, int decomposition_count, IDWTELEM *temp);
+
+void ff_spatial_idwt_slice2(DWTContext *d, int y);
+
+// shared stuff for simd optimiztions
+#define COMPOSE_53iL0(b0, b1, b2)\
+    (b1 - ((b0 + b2 + 2) >> 2))
+
+#define COMPOSE_DIRAC53iH0(b0, b1, b2)\
+    (b1 + ((b0 + b2 + 1) >> 1))
+
+#define COMPOSE_DD97iH0(b0, b1, b2, b3, b4)\
+    (b2 + ((-b0 + 9*b1 + 9*b3 - b4 + 8) >> 4))
+
+#define COMPOSE_DD137iL0(b0, b1, b2, b3, b4)\
+    (b2 - ((-b0 + 9*b1 + 9*b3 - b4 + 16) >> 5))
+
+#define COMPOSE_HAARiL0(b0, b1)\
+    (b0 - ((b1 + 1) >> 1))
+
+#define COMPOSE_HAARiH0(b0, b1)\
+    (b0 + b1)
+
+#define COMPOSE_FIDELITYiL0(b0, b1, b2, b3, b4, b5, b6, b7, b8)\
+    (b4 - ((-8*(b0+b8) + 21*(b1+b7) - 46*(b2+b6) + 161*(b3+b5) + 128) >> 8))
+
+#define COMPOSE_FIDELITYiH0(b0, b1, b2, b3, b4, b5, b6, b7, b8)\
+    (b4 + ((-2*(b0+b8) + 10*(b1+b7) - 25*(b2+b6) + 81*(b3+b5) + 128) >> 8))
+
+#define COMPOSE_DAUB97iL1(b0, b1, b2)\
+    (b1 - ((1817*(b0 + b2) + 2048) >> 12))
+
+#define COMPOSE_DAUB97iH1(b0, b1, b2)\
+    (b1 - (( 113*(b0 + b2) + 64) >> 7))
+
+#define COMPOSE_DAUB97iL0(b0, b1, b2)\
+    (b1 + (( 217*(b0 + b2) + 2048) >> 12))
+
+#define COMPOSE_DAUB97iH0(b0, b1, b2)\
+    (b1 + ((6497*(b0 + b2) + 2048) >> 12))
+
+
 
 #define DWT_97 0
 #define DWT_53 1