umc_mpeg1_dec.cpp

这是在PCA下的基于IPP库示例代码例子,在网上下了IPP的库之后,设置相关参数就可以编译该代码.

/*//////////////////////////////////////////////////////////////////////////////
//
//                  INTEL CORPORATION PROPRIETARY INFORMATION
//     This software is supplied under the terms of a license agreement or
//     nondisclosure agreement with Intel Corporation and may not be copied
//     or disclosed except in accordance with the terms of that agreement.
//          Copyright(c) 2003-2005 Intel Corporation. All Rights Reserved.
//
*/

#include "umc_mpeg2_dec_base.h"

#pragma warning(disable: 4244)

using namespace UMC;

static short zero_memory[6*64] = {0};

Status MPEG2VideoDecoderBase::skipped_macroblockP_mpeg1(int threadID)
{
    IppVideoContext *video  = &Video[threadID];
    sMacroblock     *pMacro = &Video[threadID].macroblock;
    sSlice          *pSlice = &Video[threadID].slice;

    int      y_plain_index;
    int      u_plain_index;
    int      pitch_y, pitch_uv;
    IppiSize roiy, roiuv;
    int      nmb;
    int      toskip = pSlice->mb_address_increment-1;

    pMacro->row_l = pSlice->mb_row_prev << 4;
    pMacro->col_l = pSlice->mb_col_prev << 4;
    pitch_y       = video->frame_buffer.Y_comp_pitch;
    pitch_uv      = video->frame_buffer.U_comp_pitch;
    nmb           = min(toskip, sequenceHeader.mb_width - pSlice->mb_col_prev);
    roiy.height   = 16;
    roiuv.height  = 8;

    while(toskip>0){
      pMacro->row_c = pMacro->row_l >> 1;
      pMacro->col_c = pMacro->col_l >> 1;
      roiy.width    = 16*nmb;
      roiuv.width   = 8*nmb;

      y_plain_index = pMacro->row_l * pitch_y + pMacro->col_l;
      u_plain_index = pMacro->row_c * pitch_uv + pMacro->col_c;

      ippiCopy_8u_C1R(&video->frame_buffer.frame_p_c_n[video->frame_buffer.prev_index].Y_comp_data[y_plain_index],
                      pitch_y,
                      &video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].Y_comp_data[y_plain_index],
                      pitch_y,
                      roiy);
      ippiCopy_8u_C1R(&video->frame_buffer.frame_p_c_n[video->frame_buffer.prev_index].U_comp_data[u_plain_index],
                      pitch_uv,
                      &video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].U_comp_data[u_plain_index],
                      pitch_uv,
                      roiuv);
      ippiCopy_8u_C1R(&video->frame_buffer.frame_p_c_n[video->frame_buffer.prev_index].V_comp_data[u_plain_index],
                      pitch_uv,
                      &video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].V_comp_data[u_plain_index],
                      pitch_uv,
                      roiuv);
      toskip -= nmb;
      nmb = min(toskip, sequenceHeader.mb_width);
      pMacro->col_l = 0;
      pMacro->row_l += 16;
      if(pMacro->row_l >= sequenceHeader.mb_height*16)
        return UMC_BAD_STREAM;
    }

    return UMC_OK;
}


Status MPEG2VideoDecoderBase::skipped_macroblockB_mpeg1(int threadID)
{
    int  i;
    unsigned int width;
    unsigned int width2;
    int  offset_l, offset_c;
    int  right_fl, down_fl, right_bl, down_bl, right_fc, down_fc, right_bc, down_bc;
    int  flag1, flag2, flag1c, flag2c;
    int  right_half;
    int  down_half;
    IppVideoContext *video  = &Video[threadID];
    sMacroblock     *pMacro = &Video[threadID].macroblock;
    sSlice          *pSlice = &Video[threadID].slice;

    width  = video->frame_buffer.Y_comp_pitch;
    width2 = video->frame_buffer.U_comp_pitch;

    pMacro->row_l  = pSlice->mb_row_prev << 4;
    pMacro->col_l  = pSlice->mb_col_prev << 4;

    pMacro->row_c  = pSlice->mb_row_prev << 3;
    pMacro->col_c  = pSlice->mb_col_prev << 3;

    offset_l = pMacro->row_l * width + pMacro->col_l;
    offset_c = pMacro->row_c * width2 + pMacro->col_c;

    // All pMacro->recon* was converted into half pel mode in the previous MB.
    if(pMacro->macroblock_motion_forward && pMacro->macroblock_motion_backward)
    {
        //half mode
        down_fl     = pMacro->vector[1] >> 1;
        right_fl   = pMacro->vector[0] >> 1;
        flag1      = ((pMacro->vector[0] & 1) << 3) | ((pMacro->vector[1]  & 1) << 2);

        right_half = pMacro->vector[0] / 2;
        down_half  = pMacro->vector[1] / 2;

        down_fc    = down_half >> 1;
        right_fc   = right_half >> 1;
        flag1c     = ((right_half & 1) << 3) | ((down_half & 1) << 2);

        down_bl    = pMacro->vector[3] >> 1;
        right_bl   = pMacro->vector[2] >> 1;
        flag2      = ((pMacro->vector[2] & 1) << 3) | ((pMacro->vector[3]  & 1) << 2);

        down_half  = pMacro->vector[3] / 2;
        right_half = pMacro->vector[2] / 2;
        down_bc    = down_half >> 1;
        right_bc   = right_half >> 1;
        flag2c     = ((right_half & 1) << 3) | ((down_half & 1) << 2);

        int off_fl = down_fl * width  + right_fl;
        int off_fc = down_fc * width2 + right_fc;
        int off_bl = down_bl * width  + right_bl;
        int off_bc = down_bc * width2 + right_bc;

        for(i=0; i<pSlice->mb_address_increment-1; i++) {
          ippiMC16x16B_8u_C1(video->frame_buffer.frame_p_c_n[video->frame_buffer.prev_index].Y_comp_data + offset_l + off_fl,
                               width,
                               flag1,
                               video->frame_buffer.frame_p_c_n[video->frame_buffer.next_index].Y_comp_data + offset_l + off_bl,
                               width,
                               flag2,
                               zero_memory,
                               32,
                               video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].Y_comp_data + offset_l,
                               width,
                               0);

          ippiMC8x8B_8u_C1(video->frame_buffer.frame_p_c_n[video->frame_buffer.prev_index].U_comp_data + offset_c + off_fc,
                             width2,
                             flag1c,
                             video->frame_buffer.frame_p_c_n[video->frame_buffer.next_index].U_comp_data + offset_c + off_bc,
                             width2,
                             flag2c,
                             zero_memory,
                             16,
                             video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].U_comp_data + offset_c,
                             width2,
                             0);

          ippiMC8x8B_8u_C1(video->frame_buffer.frame_p_c_n[video->frame_buffer.prev_index].V_comp_data + offset_c + off_fc,
                             width2,
                             flag1c,
                             video->frame_buffer.frame_p_c_n[video->frame_buffer.next_index].V_comp_data + offset_c + off_bc,
                             width2,
                             flag2c,
                             zero_memory,
                             16,
                             video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].V_comp_data + offset_c,
                             width2,
                             0);

          pMacro->col_l += 16;
          if(pMacro->col_l >= sequenceHeader.mb_width * 16) {
            pMacro->row_l += 16;
            pMacro->row_c += 8;
            if(pMacro->row_l >= sequenceHeader.mb_height * 16)
              return UMC_BAD_STREAM;
            offset_l = pMacro->row_l * width;
            offset_c = pMacro->row_c * width2;
            pMacro->col_l = 0;
          } else {
            offset_l += 16;
            offset_c += 8;
          }
        }
    }
    else if(pMacro->macroblock_motion_forward)
    {
        //half mode
        down_fl    = pMacro->vector[1] >> 1;
        right_fl   = pMacro->vector[0] >> 1;
        flag1      = ((pMacro->vector[0] & 1) << 3) | ((pMacro->vector[1]  & 1) << 2);

        right_half = pMacro->vector[0] / 2;
        down_half  = pMacro->vector[1] / 2;

        down_fc    = down_half >> 1;
        right_fc   = right_half >> 1;
        flag1c     = ((right_half & 1) << 3) | ((down_half & 1) << 2);

        int off_fl = down_fl * width  + right_fl;
        int off_fc = down_fc * width2 + right_fc;

        for(i=0; i<pSlice->mb_address_increment-1; i++) {
          ippiMC16x16_8u_C1(video->frame_buffer.frame_p_c_n[video->frame_buffer.prev_index].Y_comp_data + offset_l + off_fl,
                               width,
                               zero_memory,
                               32,
                               video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].Y_comp_data  + offset_l,
                               width,
                               flag1,
                               0);

          ippiMC8x8_8u_C1(video->frame_buffer.frame_p_c_n[video->frame_buffer.prev_index].U_comp_data + offset_c + off_fc,
                             width2,
                             zero_memory,
                             16,
                             video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].U_comp_data  + offset_c,
                             width2,
                             flag1c,
                             0);

          ippiMC8x8_8u_C1(video->frame_buffer.frame_p_c_n[video->frame_buffer.prev_index].V_comp_data + offset_c + off_fc,
                             width2,
                             zero_memory,
                             16,
                             video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].V_comp_data  + offset_c,
                             width2,
                             flag1c,
                             0);
          pMacro->col_l += 16;
          if(pMacro->col_l >= sequenceHeader.mb_width * 16) {
            pMacro->row_l += 16;
            pMacro->row_c += 8;
            if(pMacro->row_l >= sequenceHeader.mb_height * 16)
              return UMC_BAD_STREAM;
            offset_l = pMacro->row_l * width;
            offset_c = pMacro->row_c * width2;
            pMacro->col_l = 0;
          } else {
            offset_l += 16;
            offset_c += 8;
          }
        }
    }
    else
    {
        down_bl    = pMacro->vector[3] >> 1;
        right_bl   = pMacro->vector[2] >> 1;
        flag2      =  ((pMacro->vector[2] & 1) << 3) | ((pMacro->vector[3]  & 1) << 2);

        down_half  = pMacro->vector[3] / 2;
        right_half = pMacro->vector[2] / 2;
        down_bc    = down_half >> 1;
        right_bc   = right_half >> 1;
        flag2c     = ((right_half & 1) << 3) | ((down_half & 1) << 2);

        int off_bl = down_bl * width  + right_bl;
        int off_bc = down_bc * width2 + right_bc;

        for(i=0; i<pSlice->mb_address_increment-1; i++) {
          ippiMC16x16_8u_C1( video->frame_buffer.frame_p_c_n[video->frame_buffer.next_index].Y_comp_data + offset_l + off_bl,
                               width,
                               zero_memory,
                               32,
                               video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].Y_comp_data  + offset_l,
                               width,
                               flag2,
                               0);

          ippiMC8x8_8u_C1( video->frame_buffer.frame_p_c_n[video->frame_buffer.next_index].U_comp_data + offset_c + off_bc,
                             width2,
                             zero_memory,
                             16,
                             video->frame_buffer.frame_p_c_n[video->frame_buffer.curr_index].U_comp_data  + offset_c,