Updated FFmpeg

git-svn-id: https://mpc-hc.svn.sourceforge.net/svnroot/mpc-hc/trunk@1792 10f7b99b-c216-0410-bff0-8a66a9350fd8

1 files changed, 95 insertions, 252 deletions

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/dsputil.h b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/dsputil.h
index 3786102dd..22c91827f 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/dsputil.h
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/dsputil.h
@@ -37,8 +37,6 @@
 //#define DEBUG

 /* dct code */

 typedef short DCTELEM;

-typedef int DWTELEM;

-typedef short IDWTELEM;

 

 void fdct_ifast (DCTELEM *data);

 void fdct_ifast248 (DCTELEM *data);

@@ -88,6 +86,7 @@ extern uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP];
 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);

@@ -96,6 +95,16 @@ void ff_vp3_h_loop_filter_c(uint8_t *src, int stride, int *bounding_values);
 void ff_vp6_filter_diag4_c(uint8_t *dst, uint8_t *src, int stride,

                            const int16_t *h_weights, const int16_t *v_weights);

 

+/* 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);

+

 /* 1/2^n downscaling functions from imgconvert.c */

 void ff_img_copy_plane(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);

 void ff_shrink22(uint8_t *dst, int dst_wrap, const uint8_t *src, int src_wrap, int width, int height);

@@ -133,8 +142,8 @@ typedef void (*op_pixels_func)(uint8_t *block/*align width (8 or 16)*/, const ui
 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);

 

 #define DEF_OLD_QPEL(name)\

 void ff_put_        ## name (uint8_t *dst/*align width (8 or 16)*/, uint8_t *src/*align 1*/, int stride);\

@@ -165,10 +174,6 @@ static void a(uint8_t *block, const uint8_t *pixels, 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.

  */

@@ -193,6 +198,7 @@ typedef struct DSPContext {
     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);

@@ -321,9 +327,6 @@ typedef struct DSPContext {
     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];

@@ -351,19 +354,6 @@ typedef struct DSPContext {
     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);

 

@@ -372,6 +362,7 @@ typedef struct DSPContext {
     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);

 

@@ -381,6 +372,8 @@ typedef struct DSPContext {
     /* 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_reverse)(float *dst, const float *src0, const float *src1, int len);

@@ -390,6 +383,8 @@ typedef struct DSPContext {
     void (*vector_fmul_window)(float *dst, const float *src0, const float *src1, const float *win, float add_bias, 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 */);

+#if CONFIG_AAC_DECODER

     /**

      * Multiply a vector of floats by a scalar float.  Source and

      * destination vectors must overlap exactly or not at all.

@@ -433,6 +428,7 @@ typedef struct DSPContext {
      * @param len length of vectors, multiple of 4

      */

     float (*scalarproduct_float)(const float *v1, const float *v2, int len);

+#endif

     /**

      * Calculate the sum and difference of two vectors of floats.

      * @param v1  first input vector, sum output, 16-byte aligned

@@ -495,21 +491,6 @@ typedef struct DSPContext {
     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]);

-

     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);

@@ -548,6 +529,20 @@ typedef struct DSPContext {
     void (*x8_setup_spatial_compensation)(uint8_t *src, uint8_t *dst, int linesize,

            int * range, int * sum,  int edges);

 

+    /**

+     * Calculate scalar product of two vectors.

+     * @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);

+    /* 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);

+

     /* rv30 functions */

     qpel_mc_func put_rv30_tpel_pixels_tab[4][16];

     qpel_mc_func avg_rv30_tpel_pixels_tab[4][16];

@@ -622,8 +617,14 @@ extern int mm_flags;
 

 void dsputil_init_mmx(DSPContext* c, AVCodecContext *avctx);

 

-#define DECLARE_ALIGNED_16(t, v, ...) DECLARE_ALIGNED(16, t, v)

-#define DECLARE_ALIGNED_8(t, v, ...)  DECLARE_ALIGNED(8, t, v)

+void ff_dsputil_init_dwt(DSPContext *c);

+void ff_cavsdsp_init(DSPContext* c, AVCodecContext *avctx);

+void ff_rv30dsp_init(DSPContext* c, AVCodecContext *avctx);

+void ff_rv40dsp_init(DSPContext* c, AVCodecContext *avctx);

+void ff_vc1dsp_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 HAVE_MMX

 

@@ -656,222 +657,26 @@ static inline void emms(void)
 #   define STRIDE_ALIGN 8

 #endif

 

-/* PSNR */

-void get_psnr(uint8_t *orig_image[3], uint8_t *coded_image[3],

-              int orig_linesize[3], int coded_linesize,

-              AVCodecContext *avctx);

+#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)

 

-/* FFT computation */

-

-/* NOTE: soon integer code will be added, so you must use the

-   FFTSample type */

-typedef float FFTSample;

-

-typedef struct FFTComplex {

-    FFTSample re, im;

-} FFTComplex;

-

-typedef struct FFTContext {

-    int nbits;

-    int inverse;

-    uint16_t *revtab;

-    FFTComplex *exptab;

-    FFTComplex *exptab1; /* only used by SSE code */

-    FFTComplex *tmp_buf;

-    int mdct_size; /* size of MDCT (i.e. number of input data * 2) */

-    int mdct_bits; /* n = 2^nbits */

-    /* pre/post rotation tables */

-    FFTSample *tcos;

-    FFTSample *tsin;

-    void (*fft_permute)(struct FFTContext *s, FFTComplex *z);

-    void (*fft_calc)(struct FFTContext *s, FFTComplex *z);

-    void (*imdct_calc)(struct FFTContext *s, FFTSample *output, const FFTSample *input);

-    void (*imdct_half)(struct FFTContext *s, FFTSample *output, const FFTSample *input);

-    void (*mdct_calc)(struct FFTContext *s, FFTSample *output, const FFTSample *input);

-    int split_radix;

-    int permutation;

-#define FF_MDCT_PERM_NONE       0

-#define FF_MDCT_PERM_INTERLEAVE 1

-} FFTContext;

-

-#if CONFIG_HARDCODED_TABLES

-#define COSTABLE_CONST const

-#define SINTABLE_CONST const

-#define SINETABLE_CONST const

+#if HAVE_LOCAL_ALIGNED_8

+#   define LOCAL_ALIGNED_8(t, v, s, ...) DECLARE_ALIGNED(8, t, v) s __VA_ARGS__

 #else

-#define COSTABLE_CONST

-#define SINTABLE_CONST

-#define SINETABLE_CONST

+#   define LOCAL_ALIGNED_8(t, v, s, ...) LOCAL_ALIGNED(8, t, v, s, __VA_ARGS__)

 #endif

 

-#define COSTABLE(size) \

-    COSTABLE_CONST DECLARE_ALIGNED_16(FFTSample, ff_cos_##size)[size/2]

-#define SINTABLE(size) \

-    SINTABLE_CONST DECLARE_ALIGNED_16(FFTSample, ff_sin_##size)[size/2]

-#define SINETABLE(size) \

-    SINETABLE_CONST DECLARE_ALIGNED_16(float, ff_sine_##size)[size]

-extern COSTABLE(16);

-extern COSTABLE(32);

-extern COSTABLE(64);

-extern COSTABLE(128);

-extern COSTABLE(256);

-extern COSTABLE(512);

-extern COSTABLE(1024);

-extern COSTABLE(2048);

-extern COSTABLE(4096);

-extern COSTABLE(8192);

-extern COSTABLE(16384);

-extern COSTABLE(32768);

-extern COSTABLE(65536);

-extern COSTABLE_CONST FFTSample* const ff_cos_tabs[17];

-

-/**

- * Initializes the cosine table in ff_cos_tabs[index]

- * \param index index in ff_cos_tabs array of the table to initialize

- */

-void ff_init_ff_cos_tabs(int index);

-

-extern SINTABLE(16);

-extern SINTABLE(32);

-extern SINTABLE(64);

-extern SINTABLE(128);

-extern SINTABLE(256);

-extern SINTABLE(512);

-extern SINTABLE(1024);

-extern SINTABLE(2048);

-extern SINTABLE(4096);

-extern SINTABLE(8192);

-extern SINTABLE(16384);

-extern SINTABLE(32768);

-extern SINTABLE(65536);

-

-/**

- * Sets up a complex FFT.

- * @param nbits           log2 of the length of the input array

- * @param inverse         if 0 perform the forward transform, if 1 perform the inverse

- */

-int ff_fft_init(FFTContext *s, int nbits, int inverse);

-void ff_fft_permute_c(FFTContext *s, FFTComplex *z);

-void ff_fft_calc_c(FFTContext *s, FFTComplex *z);

-

-void ff_fft_init_mmx(FFTContext *s);

-

-/**

- * Do the permutation needed BEFORE calling ff_fft_calc().

- */

-static inline void ff_fft_permute(FFTContext *s, FFTComplex *z)

-{

-    s->fft_permute(s, z);

-}

-/**

- * Do a complex FFT with the parameters defined in ff_fft_init(). The

- * input data must be permuted before. No 1.0/sqrt(n) normalization is done.

- */

-static inline void ff_fft_calc(FFTContext *s, FFTComplex *z)

-{

-    s->fft_calc(s, z);

-}

-void ff_fft_end(FFTContext *s);

-

-/* MDCT computation */

-

-static inline void ff_imdct_calc(FFTContext *s, FFTSample *output, const FFTSample *input)

-{

-    s->imdct_calc(s, output, input);

-}

-static inline void ff_imdct_half(FFTContext *s, FFTSample *output, const FFTSample *input)

-{

-    s->imdct_half(s, output, input);

-}

-

-static inline void ff_mdct_calc(FFTContext *s, FFTSample *output,

-                                const FFTSample *input)

-{

-    s->mdct_calc(s, output, input);

-}

-

-/**

- * Generate a Kaiser-Bessel Derived Window.

- * @param   window  pointer to half window

- * @param   alpha   determines window shape

- * @param   n       size of half window

- */

-void ff_kbd_window_init(float *window, float alpha, int n);

-

-/**

- * Generate a sine window.

- * @param   window  pointer to half window

- * @param   n       size of half window

- */

-void ff_sine_window_init(float *window, int n);

-/**

- * initialize the specified entry of ff_sine_windows

- */

-void ff_init_ff_sine_windows(int index);

-extern SINETABLE(  32);

-extern SINETABLE(  64);

-extern SINETABLE( 128);

-extern SINETABLE( 256);

-extern SINETABLE( 512);

-extern SINETABLE(1024);

-extern SINETABLE(2048);

-extern SINETABLE(4096);

-extern SINETABLE_CONST float * const ff_sine_windows[13];

-

-int ff_mdct_init(FFTContext *s, int nbits, int inverse, double scale);

-void ff_imdct_calc_c(FFTContext *s, FFTSample *output, const FFTSample *input);

-void ff_imdct_half_c(FFTContext *s, FFTSample *output, const FFTSample *input);

-void ff_mdct_calc_c(FFTContext *s, FFTSample *output, const FFTSample *input);

-void ff_mdct_end(FFTContext *s);

-

-/* Real Discrete Fourier Transform */

-

-enum RDFTransformType {

-    RDFT,

-    IRDFT,

-    RIDFT,

-    IRIDFT,

-};

-

-typedef struct {

-    int nbits;

-    int inverse;

-    int sign_convention;

-

-    /* pre/post rotation tables */

-    const FFTSample *tcos;

-    SINTABLE_CONST FFTSample *tsin;

-    FFTContext fft;

-} RDFTContext;

-

-/**

- * Sets up a real FFT.

- * @param nbits           log2 of the length of the input array

- * @param trans           the type of transform

- */

-int ff_rdft_init(RDFTContext *s, int nbits, enum RDFTransformType trans);

-void ff_rdft_calc(RDFTContext *s, FFTSample *data);

-void ff_rdft_end(RDFTContext *s);

-

-/* Discrete Cosine Transform */

-

-typedef struct {

-    int nbits;

-    int inverse;

-    FFTSample *data;

-    RDFTContext rdft;

-    const float *costab;

-    FFTSample *csc2;

-} DCTContext;

+#if HAVE_LOCAL_ALIGNED_16

+#   define LOCAL_ALIGNED_16(t, v, s, ...) DECLARE_ALIGNED(16, t, v) s __VA_ARGS__

+#else

+#   define LOCAL_ALIGNED_16(t, v, s, ...) LOCAL_ALIGNED(16, t, v, s, __VA_ARGS__)

+#endif

 

-/**

- * Sets up (Inverse)DCT.

- * @param nbits           log2 of the length of the input array

- * @param inverse         >0 forward transform, <0 inverse transform

- */

-int  ff_dct_init(DCTContext *s, int nbits, int inverse);

-void ff_dct_calc(DCTContext *s, FFTSample *data);

-void ff_dct_end (DCTContext *s);

+/* PSNR */

+void get_psnr(uint8_t *orig_image[3], uint8_t *coded_image[3],

+              int orig_linesize[3], int coded_linesize,

+              AVCodecContext *avctx);

 

 #define WRAPPER8_16(name8, name16)\

 static int name16(void /*MpegEncContext*/ *s, uint8_t *dst, uint8_t *src, int stride, int h){\

@@ -972,4 +777,42 @@ static inline void copy_block17(uint8_t *dst, const uint8_t *src, int dstStride,
 

 const char* avcodec_get_current_idct_mmx(AVCodecContext *avctx,DSPContext *c);

 

+/* ffdshow custom code begin */

+#ifndef D3DCOLOR_DEFINED

+typedef uint32_t D3DCOLOR;

+#define D3DCOLOR_DEFINED

+#endif

+

+#define D3DCOLOR_ARGB(a,r,g,b) \

+    ((D3DCOLOR)((((a)&0xff)<<24)|(((r)&0xff)<<16)|(((g)&0xff)<<8)|((b)&0xff)))

+

+static const double Rec601_Kr = 0.299;

+static const double Rec601_Kb = 0.114;

+static const double Rec601_Kg = 0.587;

+static const double Rec709_Kr = 0.2125;

+static const double Rec709_Kb = 0.0721;

+static const double Rec709_Kg = 0.7154;

+

+static inline uint32_t YCrCbToRGB_Rec709(uint8_t A, uint8_t Y, uint8_t Cr, uint8_t Cb)

+{

+

+  double rp = Y + 2*(Cr-128)*(1.0-Rec709_Kr);

+  double gp = Y - 2*(Cb-128)*(1.0-Rec709_Kb)*Rec709_Kb/Rec709_Kg - 2*(Cr-128)*(1.0-Rec709_Kr)*Rec709_Kr/Rec709_Kg;

+  double bp = Y + 2*(Cb-128)*(1.0-Rec709_Kb);

+

+  return D3DCOLOR_ARGB(A, (uint8_t)fabs(rp), (uint8_t)fabs(gp), (uint8_t)fabs(bp));

+}

+

+static inline uint32_t YCrCbToRGB_Rec601(uint8_t A, uint8_t Y, uint8_t Cr, uint8_t Cb)

+{

+

+  double rp = Y + 2*(Cr-128)*(1.0-Rec601_Kr);

+  double gp = Y - 2*(Cb-128)*(1.0-Rec601_Kb)*Rec601_Kb/Rec601_Kg - 2*(Cr-128)*(1.0-Rec601_Kr)*Rec601_Kr/Rec601_Kg;

+  double bp = Y + 2*(Cb-128)*(1.0-Rec601_Kb);

+

+  return D3DCOLOR_ARGB(A, (uint8_t)fabs(rp), (uint8_t)fabs(gp), (uint8_t)fabs(bp));

+}

+

+/* ffdshow custom code end */

+

 #endif /* AVCODEC_DSPUTIL_H */