comparison vp3.c @ 10647:0139a62827ce libavcodec

Various VP3 optimizations. Faster checks in reverse_dc_prediction. Simplified deblocking checks. Check transform==15 first, since it's more common than 13.

10646:bb14c1586891

 * the frame. Much of this function is adapted directly from the original
 * VP3 source code.
 */
#define COMPATIBLE_FRAME(x) \
  (compatible_frame[s->all_fragments[x].coding_method] == current_frame_type)
#define FRAME_CODED(x) (s->all_fragments[x].coding_method != MODE_COPY)
#define DC_COEFF(u) (s->coeffs[u].index ? 0 : s->coeffs[u].coeff) //FIXME do somethin to simplify this

static void reverse_dc_prediction(Vp3DecodeContext *s,
                                  int first_fragment,
                                  int fragment_width,

     * prediction. For example, INTRA is the only mode in this table to
     * have a frame number of 0. That means INTRA blocks can only predict
     * from other INTRA blocks. There are 2 golden frame coding types;
     * blocks encoding in these modes can only predict from other blocks
     * that were encoded with these 1 of these 2 modes. */
    static const unsigned char compatible_frame[8] = {
        1,    /* MODE_INTER_NO_MV */
        0,    /* MODE_INTRA */
        1,    /* MODE_INTER_PLUS_MV */
        1,    /* MODE_INTER_LAST_MV */
        1,    /* MODE_INTER_PRIOR_MV */
        2,    /* MODE_USING_GOLDEN */
        2,    /* MODE_GOLDEN_MV */
        1     /* MODE_INTER_FOUR_MV */
    };
    int current_frame_type;

    /* there is a last DC predictor for each of the 3 frame types */
    short last_dc[3];

                transform= 0;
                if(x){
                    l= i-1;
                    vl = DC_COEFF(l);
                    if(FRAME_CODED(l) && COMPATIBLE_FRAME(l))
                        transform |= PL;
                }
                if(y){
                    u= i-fragment_width;
                    vu = DC_COEFF(u);
                    if(FRAME_CODED(u) && COMPATIBLE_FRAME(u))
                        transform |= PU;
                    if(x){
                        ul= i-fragment_width-1;
                        vul = DC_COEFF(ul);
                        if(FRAME_CODED(ul) && COMPATIBLE_FRAME(ul))
                            transform |= PUL;
                    }
                    if(x + 1 < fragment_width){
                        ur= i-fragment_width+1;
                        vur = DC_COEFF(ur);
                        if(FRAME_CODED(ur) && COMPATIBLE_FRAME(ur))
                            transform |= PUR;
                    }
                }

                if (transform == 0) {

                    predicted_dc /= 128;

                    /* check for outranging on the [ul u l] and
                     * [ul u ur l] predictors */
                    if ((transform == 13) || (transform == 15)) {
                        if (FFABS(predicted_dc - vu) > 128)
                            predicted_dc = vu;
                        else if (FFABS(predicted_dc - vl) > 128)
                            predicted_dc = vl;
                        else if (FFABS(predicted_dc - vul) > 128)

        if (!s->flipped_image) stride = -stride;

        for (y = 0; y < height; y++) {

            for (x = 0; x < width; x++) {
                /* do not perform left edge filter for left columns frags */
                if ((x > 0) &&
                    (s->all_fragments[fragment].coding_method != MODE_COPY)) {
                    s->dsp.vp3_h_loop_filter(
                        plane_data + s->all_fragments[fragment].first_pixel,
                        stride, bounding_values);
                }

                /* do not perform top edge filter for top row fragments */
                if ((y > 0) &&
                    (s->all_fragments[fragment].coding_method != MODE_COPY)) {
                    s->dsp.vp3_v_loop_filter(
                        plane_data + s->all_fragments[fragment].first_pixel,
                        stride, bounding_values);
                }

                /* do not perform right edge filter for right column
                 * fragments or if right fragment neighbor is also coded
                 * in this frame (it will be filtered in next iteration) */
                if ((x < width - 1) &&
                    (s->all_fragments[fragment].coding_method != MODE_COPY) &&
                    (s->all_fragments[fragment + 1].coding_method == MODE_COPY)) {
                    s->dsp.vp3_h_loop_filter(
                        plane_data + s->all_fragments[fragment + 1].first_pixel,
                        stride, bounding_values);
                }

                /* do not perform bottom edge filter for bottom row
                 * fragments or if bottom fragment neighbor is also coded
                 * in this frame (it will be filtered in the next row) */
                if ((y < height - 1) &&
                    (s->all_fragments[fragment].coding_method != MODE_COPY) &&
                    (s->all_fragments[fragment + width].coding_method == MODE_COPY)) {
                    s->dsp.vp3_v_loop_filter(
                        plane_data + s->all_fragments[fragment + width].first_pixel,
                        stride, bounding_values);
                }

                fragment++;
            }
        }

10647:0139a62827ce

 * the frame. Much of this function is adapted directly from the original
 * VP3 source code.
 */
#define COMPATIBLE_FRAME(x) \
  (compatible_frame[s->all_fragments[x].coding_method] == current_frame_type)
#define DC_COEFF(u) (s->coeffs[u].index ? 0 : s->coeffs[u].coeff) //FIXME do somethin to simplify this

static void reverse_dc_prediction(Vp3DecodeContext *s,
                                  int first_fragment,
                                  int fragment_width,

     * prediction. For example, INTRA is the only mode in this table to
     * have a frame number of 0. That means INTRA blocks can only predict
     * from other INTRA blocks. There are 2 golden frame coding types;
     * blocks encoding in these modes can only predict from other blocks
     * that were encoded with these 1 of these 2 modes. */
    static const unsigned char compatible_frame[9] = {
        1,    /* MODE_INTER_NO_MV */
        0,    /* MODE_INTRA */
        1,    /* MODE_INTER_PLUS_MV */
        1,    /* MODE_INTER_LAST_MV */
        1,    /* MODE_INTER_PRIOR_MV */
        2,    /* MODE_USING_GOLDEN */
        2,    /* MODE_GOLDEN_MV */
        1,    /* MODE_INTER_FOUR_MV */
        3     /* MODE_COPY */
    };
    int current_frame_type;

    /* there is a last DC predictor for each of the 3 frame types */
    short last_dc[3];

                transform= 0;
                if(x){
                    l= i-1;
                    vl = DC_COEFF(l);
                    if(COMPATIBLE_FRAME(l))
                        transform |= PL;
                }
                if(y){
                    u= i-fragment_width;
                    vu = DC_COEFF(u);
                    if(COMPATIBLE_FRAME(u))
                        transform |= PU;
                    if(x){
                        ul= i-fragment_width-1;
                        vul = DC_COEFF(ul);
                        if(COMPATIBLE_FRAME(ul))
                            transform |= PUL;
                    }
                    if(x + 1 < fragment_width){
                        ur= i-fragment_width+1;
                        vur = DC_COEFF(ur);
                        if(COMPATIBLE_FRAME(ur))
                            transform |= PUR;
                    }
                }

                if (transform == 0) {

                    predicted_dc /= 128;

                    /* check for outranging on the [ul u l] and
                     * [ul u ur l] predictors */
                    if ((transform == 15) || (transform == 13)) {
                        if (FFABS(predicted_dc - vu) > 128)
                            predicted_dc = vu;
                        else if (FFABS(predicted_dc - vl) > 128)
                            predicted_dc = vl;
                        else if (FFABS(predicted_dc - vul) > 128)

        if (!s->flipped_image) stride = -stride;

        for (y = 0; y < height; y++) {

            for (x = 0; x < width; x++) {
                /* This code basically just deblocks on the edges of coded blocks.
                 * However, it has to be much more complicated because of the
                 * braindamaged deblock ordering used in VP3/Theora. Order matters
                 * because some pixels get filtered twice. */
                if( s->all_fragments[fragment].coding_method != MODE_COPY )
                {
                    /* do not perform left edge filter for left columns frags */
                    if (x > 0) {
                        s->dsp.vp3_h_loop_filter(
                            plane_data + s->all_fragments[fragment].first_pixel,
                            stride, bounding_values);
                    }

                    /* do not perform top edge filter for top row fragments */
                    if (y > 0) {
                        s->dsp.vp3_v_loop_filter(
                            plane_data + s->all_fragments[fragment].first_pixel,
                            stride, bounding_values);
                    }

                    /* do not perform right edge filter for right column
                     * fragments or if right fragment neighbor is also coded
                     * in this frame (it will be filtered in next iteration) */
                    if ((x < width - 1) &&
                        (s->all_fragments[fragment + 1].coding_method == MODE_COPY)) {
                        s->dsp.vp3_h_loop_filter(
                            plane_data + s->all_fragments[fragment + 1].first_pixel,
                            stride, bounding_values);
                    }

                    /* do not perform bottom edge filter for bottom row
                     * fragments or if bottom fragment neighbor is also coded
                     * in this frame (it will be filtered in the next row) */
                    if ((y < height - 1) &&
                        (s->all_fragments[fragment + width].coding_method == MODE_COPY)) {
                        s->dsp.vp3_v_loop_filter(
                            plane_data + s->all_fragments[fragment + width].first_pixel,
                            stride, bounding_values);
                    }
                }

                fragment++;
            }
        }