simple_idct.c

播放H264文件的播放器

        dest[0] = cm[dest[0] + ((a1 + b1) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a2 + b2) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a3 + b3) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a3 - b3) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a2 - b2) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a1 - b1) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a0 - b0) >> COL_SHIFT)];
}

static _inline void idctSparseCol (DCTELEM * col)
{
        int a0, a1, a2, a3, b0, b1, b2, b3;

        /* XXX: I did that only to give same values as previous code */
        a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4));
        a1 = a0;
        a2 = a0;
        a3 = a0;

        a0 +=  + W2*col[8*2];
        a1 +=  + W6*col[8*2];
        a2 +=  - W6*col[8*2];
        a3 +=  - W2*col[8*2];

        b0 = MUL16(W1, col[8*1]);
        b1 = MUL16(W3, col[8*1]);
        b2 = MUL16(W5, col[8*1]);
        b3 = MUL16(W7, col[8*1]);

        MAC16(b0, + W3, col[8*3]);
        MAC16(b1, - W7, col[8*3]);
        MAC16(b2, - W1, col[8*3]);
        MAC16(b3, - W5, col[8*3]);

        if(col[8*4]){
            a0 += + W4*col[8*4];
            a1 += - W4*col[8*4];
            a2 += - W4*col[8*4];
            a3 += + W4*col[8*4];
        }

        if (col[8*5]) {
            MAC16(b0, + W5, col[8*5]);
            MAC16(b1, - W1, col[8*5]);
            MAC16(b2, + W7, col[8*5]);
            MAC16(b3, + W3, col[8*5]);
        }

        if(col[8*6]){
            a0 += + W6*col[8*6];
            a1 += - W2*col[8*6];
            a2 += + W2*col[8*6];
            a3 += - W6*col[8*6];
        }

        if (col[8*7]) {
            MAC16(b0, + W7, col[8*7]);
            MAC16(b1, - W5, col[8*7]);
            MAC16(b2, + W3, col[8*7]);
            MAC16(b3, - W1, col[8*7]);
        }

        col[0 ] = ((a0 + b0) >> COL_SHIFT);
        col[8 ] = ((a1 + b1) >> COL_SHIFT);
        col[16] = ((a2 + b2) >> COL_SHIFT);
        col[24] = ((a3 + b3) >> COL_SHIFT);
        col[32] = ((a3 - b3) >> COL_SHIFT);
        col[40] = ((a2 - b2) >> COL_SHIFT);
        col[48] = ((a1 - b1) >> COL_SHIFT);
        col[56] = ((a0 - b0) >> COL_SHIFT);
}

void ff_simple_idct_put(uint8_t *dest, int line_size, DCTELEM *block)
{
    int i;
    for(i=0; i<8; i++)
        idctRowCondDC(block + i*8);

    for(i=0; i<8; i++)
        idctSparseColPut(dest + i, line_size, block + i);
}

void ff_simple_idct_add(uint8_t *dest, int line_size, DCTELEM *block)
{
    int i;
    for(i=0; i<8; i++)
        idctRowCondDC(block + i*8);

    for(i=0; i<8; i++)
        idctSparseColAdd(dest + i, line_size, block + i);
}

void ff_simple_idct(DCTELEM *block)
{
    int i;
    for(i=0; i<8; i++)
        idctRowCondDC(block + i*8);

    for(i=0; i<8; i++)
        idctSparseCol(block + i);
}

/* 2x4x8 idct */

#define CN_SHIFT 12
#define C_FIX(x) ((int)((x) * (1 << CN_SHIFT) + 0.5))
#define C1 C_FIX(0.6532814824)
#define C2 C_FIX(0.2705980501)

/* row idct is multiple by 16 * sqrt(2.0), col idct4 is normalized,
   and the butterfly must be multiplied by 0.5 * sqrt(2.0) */
#define C_SHIFT (4+1+12)

static _inline void idct4col_put(uint8_t *dest, int line_size, const DCTELEM *col)
{
    int c0, c1, c2, c3, a0, a1, a2, a3;
    const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;

    a0 = col[8*0];
    a1 = col[8*2];
    a2 = col[8*4];
    a3 = col[8*6];
    c0 = ((a0 + a2) << (CN_SHIFT - 1)) + (1 << (C_SHIFT - 1));
    c2 = ((a0 - a2) << (CN_SHIFT - 1)) + (1 << (C_SHIFT - 1));
    c1 = a1 * C1 + a3 * C2;
    c3 = a1 * C2 - a3 * C1;
    dest[0] = cm[(c0 + c1) >> C_SHIFT];
    dest += line_size;
    dest[0] = cm[(c2 + c3) >> C_SHIFT];
    dest += line_size;
    dest[0] = cm[(c2 - c3) >> C_SHIFT];
    dest += line_size;
    dest[0] = cm[(c0 - c1) >> C_SHIFT];
}

#define BF(k) \
{\
    int a0, a1;\
    a0 = ptr[k];\
    a1 = ptr[8 + k];\
    ptr[k] = a0 + a1;\
    ptr[8 + k] = a0 - a1;\
}

/* only used by DV codec. The input must be interlaced. 128 is added
   to the pixels before clamping to avoid systematic error
   (1024*sqrt(2)) offset would be needed otherwise. */
/* XXX: I think a 1.0/sqrt(2) normalization should be needed to
   compensate the extra butterfly stage - I don't have the full DV
   specification */
void ff_simple_idct248_put(uint8_t *dest, int line_size, DCTELEM *block)
{
    int i;
    DCTELEM *ptr;

    /* butterfly */
    ptr = block;
    for(i=0;i<4;i++) {
        BF(0);
        BF(1);
        BF(2);
        BF(3);
        BF(4);
        BF(5);
        BF(6);
        BF(7);
        ptr += 2 * 8;
    }

    /* IDCT8 on each line */
    for(i=0; i<8; i++) {
        idctRowCondDC(block + i*8);
    }

    /* IDCT4 and store */
    for(i=0;i<8;i++) {
        idct4col_put(dest + i, 2 * line_size, block + i);
        idct4col_put(dest + line_size + i, 2 * line_size, block + 8 + i);
    }
}

/* 8x4 & 4x8 WMV2 IDCT */
#undef CN_SHIFT
#undef C_SHIFT
#undef C_FIX
#undef C1
#undef C2
#define CN_SHIFT 12
#define C_FIX(x) ((int)((x) * 1.414213562 * (1 << CN_SHIFT) + 0.5))
#define C1 C_FIX(0.6532814824)
#define C2 C_FIX(0.2705980501)
#define C3 C_FIX(0.5)
#define C_SHIFT (4+1+12)
static _inline void idct4col_add(uint8_t *dest, int line_size, const DCTELEM *col)
{
    int c0, c1, c2, c3, a0, a1, a2, a3;
    const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;

    a0 = col[8*0];
    a1 = col[8*1];
    a2 = col[8*2];
    a3 = col[8*3];
    c0 = (a0 + a2)*C3 + (1 << (C_SHIFT - 1));
    c2 = (a0 - a2)*C3 + (1 << (C_SHIFT - 1));
    c1 = a1 * C1 + a3 * C2;
    c3 = a1 * C2 - a3 * C1;
    dest[0] = cm[dest[0] + ((c0 + c1) >> C_SHIFT)];
    dest += line_size;
    dest[0] = cm[dest[0] + ((c2 + c3) >> C_SHIFT)];
    dest += line_size;
    dest[0] = cm[dest[0] + ((c2 - c3) >> C_SHIFT)];
    dest += line_size;
    dest[0] = cm[dest[0] + ((c0 - c1) >> C_SHIFT)];
}

#define RN_SHIFT 15
#define R_FIX(x) ((int)((x) * 1.414213562 * (1 << RN_SHIFT) + 0.5))
#define R1 R_FIX(0.6532814824)
#define R2 R_FIX(0.2705980501)
#define R3 R_FIX(0.5)
#define R_SHIFT 11
static _inline void idct4row(DCTELEM *row)
{
    int c0, c1, c2, c3, a0, a1, a2, a3;
    //const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;

    a0 = row[0];
    a1 = row[1];
    a2 = row[2];
    a3 = row[3];
    c0 = (a0 + a2)*R3 + (1 << (R_SHIFT - 1));
    c2 = (a0 - a2)*R3 + (1 << (R_SHIFT - 1));
    c1 = a1 * R1 + a3 * R2;
    c3 = a1 * R2 - a3 * R1;
    row[0]= (c0 + c1) >> R_SHIFT;
    row[1]= (c2 + c3) >> R_SHIFT;
    row[2]= (c2 - c3) >> R_SHIFT;
    row[3]= (c0 - c1) >> R_SHIFT;
}

void ff_simple_idct84_add(uint8_t *dest, int line_size, DCTELEM *block)
{
    int i;

    /* IDCT8 on each line */
    for(i=0; i<4; i++) {
        idctRowCondDC(block + i*8);
    }

    /* IDCT4 and store */
    for(i=0;i<8;i++) {
        idct4col_add(dest + i, line_size, block + i);
    }
}

void ff_simple_idct48_add(uint8_t *dest, int line_size, DCTELEM *block)
{
    int i;

    /* IDCT4 on each line */
    for(i=0; i<8; i++) {
        idct4row(block + i*8);
    }

    /* IDCT8 and store */
    for(i=0; i<4; i++){
        idctSparseColAdd(dest + i, line_size, block + i);
    }
}

void ff_simple_idct44_add(uint8_t *dest, int line_size, DCTELEM *block)
{
    int i;

    /* IDCT4 on each line */
    for(i=0; i<4; i++) {
        idct4row(block + i*8);
    }

    /* IDCT4 and store */
    for(i=0; i<4; i++){
        idct4col_add(dest + i, line_size, block + i);
    }
}