M6[3]= M5[1]*17+M5[3]*7;

    for (i=0;i<2;i++)
    {
      i1=3-i;
      img->cof[i ][j][0][0]= M6[i]+M6[i1];
      img->cof[i1][j][0][0]=M6[i]-M6[i1];
    }
  }

  // vertical
  for (i=0;i<4;i++)
  {
    for (j=0;j<4;j++)
      M5[j]=img->cof[i][j][0][0];

    M6[0]=(M5[0]+M5[2])*13;
    M6[1]=(M5[0]-M5[2])*13;
    M6[2]= M5[1]*7 -M5[3]*17;
    M6[3]= M5[1]*17+M5[3]*7;

    for (j=0;j<2;j++)
    {
      j1=3-j;
      img->cof[i][j][0][0] = ((M6[j]+M6[j1])/8) *JQ1[img->qp];
      img->cof[i][j1][0][0]= ((M6[j]-M6[j1])/8) *JQ1[img->qp];
    }
  }
  for (j=0;j<4;j++)
  {
    for (i=0;i<4;i++)
    {
      img->cof[i][j][0][0] = 3 * img->cof[i][j][0][0]/256;
    }
  }
}


void itrans_sp(struct img_par *img,  //!< image parameters
               int ioff,             //!< index to 4x4 block
               int joff,             //!<
               int i0,               //!<
               int j0)               //!<
{
  int i,j,i1,j1;
  int m5[4];
  int m6[4];
  int predicted_block[BLOCK_SIZE][BLOCK_SIZE],Fq1q2,ilev;

  for (j=0; j< BLOCK_SIZE; j++)
    for (i=0; i< BLOCK_SIZE; i++)
    {
      predicted_block[i][j]=img->mpr[i+ioff][j+joff];
    }
  for (j=0; j < BLOCK_SIZE; j++)
  {
    for (i=0; i < 2; i++)
    {
      i1=3-i;
      m5[i]=predicted_block[i][j]+predicted_block[i1][j];
      m5[i1]=predicted_block[i][j]-predicted_block[i1][j];
    }
    predicted_block[0][j]=(m5[0]+m5[1])*13;
    predicted_block[2][j]=(m5[0]-m5[1])*13;
    predicted_block[1][j]=m5[3]*17+m5[2]*7;
    predicted_block[3][j]=m5[3]*7-m5[2]*17;
  }

  //  Vertival transform

  for (i=0; i < BLOCK_SIZE; i++)
  {
    for (j=0; j < 2; j++)
    {
      j1=3-j;
      m5[j]=predicted_block[i][j]+predicted_block[i][j1];
      m5[j1]=predicted_block[i][j]-predicted_block[i][j1];
    }
    predicted_block[i][0]=(m5[0]+m5[1])*13;
    predicted_block[i][2]=(m5[0]-m5[1])*13;
    predicted_block[i][1]=m5[3]*17+m5[2]*7;
    predicted_block[i][3]=m5[3]*7-m5[2]*17;
  }

  Fq1q2=(JQQ*JQ[img->qpsp]+JQ[img->qp]/2)/JQ[img->qp];

  for (j=0;j<BLOCK_SIZE;j++)
    for (i=0;i<BLOCK_SIZE;i++)
    {
      ilev=img->cof[i0][j0][i][j]/JQ1[img->qp]*Fq1q2+predicted_block[i][j]*JQ[img->qpsp];
      img->cof[i0][j0][i][j]=sign((abs(ilev)+JQQ2)/ JQQ,ilev)*JQ1[img->qpsp];
    }
  // horizontal
  for (j=0;j<BLOCK_SIZE;j++)
  {
    for (i=0;i<BLOCK_SIZE;i++)
    {
      m5[i]=img->cof[i0][j0][i][j];
    }
    m6[0]=(m5[0]+m5[2])*13;
    m6[1]=(m5[0]-m5[2])*13;
    m6[2]=m5[1]*7-m5[3]*17;
    m6[3]=m5[1]*17+m5[3]*7;

    for (i=0;i<2;i++)
    {
      i1=3-i;
      img->m7[i][j]=m6[i]+m6[i1];
      img->m7[i1][j]=m6[i]-m6[i1];
    }
  }
  // vertical
  for (i=0;i<BLOCK_SIZE;i++)
  {
    for (j=0;j<BLOCK_SIZE;j++)
      m5[j]=img->m7[i][j];

    m6[0]=(m5[0]+m5[2])*13;
    m6[1]=(m5[0]-m5[2])*13;
    m6[2]=m5[1]*7-m5[3]*17;
    m6[3]=m5[1]*17+m5[3]*7;

    for (j=0;j<2;j++)
    {
      j1=3-j;
      img->m7[i][j]=mmax(0,mmin(255,(m6[j]+m6[j1]+JQQ2)/JQQ));
      img->m7[i][j1]=mmax(0,mmin(255,(m6[j]-m6[j1]+JQQ2)/JQQ));
    }
  }

}

/*!
 ***********************************************************************
 * \brief
 *    The routine performs transform,quantization,inverse transform, adds the diff.
 *    to the prediction and writes the result to the decoded luma frame. Includes the
 *    RD constrained quantization also.
 *
 * \par Input:
 *    block_x,block_y: Block position inside a macro block (0,4,8,12).
 *
 * \par Output:
 *    nonzero: 0 if no levels are nonzero.  1 if there are nonzero levels. \n
 *    coeff_cost: Counter for nonzero coefficients, used to discard expencive levels.
 ************************************************************************
 */
void copyblock_sp(struct img_par *img,int block_x,int block_y)
{
  int sign(int a,int b);

  int i,j,i1,j1,m5[4],m6[4];

  int predicted_block[BLOCK_SIZE][BLOCK_SIZE],quant_set1;

  //  Horizontal transform
  for (j=0; j< BLOCK_SIZE; j++)
    for (i=0; i< BLOCK_SIZE; i++)
    {
      predicted_block[i][j]=img->mpr[i+block_x][j+block_y];
    }

  for (j=0; j < BLOCK_SIZE; j++)
  {
    for (i=0; i < 2; i++)
    {
      i1=3-i;
      m5[i]=predicted_block[i][j]+predicted_block[i1][j];
      m5[i1]=predicted_block[i][j]-predicted_block[i1][j];
    }
    predicted_block[0][j]=(m5[0]+m5[1])*13;
    predicted_block[2][j]=(m5[0]-m5[1])*13;
    predicted_block[1][j]=m5[3]*17+m5[2]*7;
    predicted_block[3][j]=m5[3]*7-m5[2]*17;
  }

  //  Vertical transform

  for (i=0; i < BLOCK_SIZE; i++)
  {
    for (j=0; j < 2; j++)
    {
      j1=3-j;
      m5[j]=predicted_block[i][j]+predicted_block[i][j1];
      m5[j1]=predicted_block[i][j]-predicted_block[i][j1];
    }
    predicted_block[i][0]=(m5[0]+m5[1])*13;
    predicted_block[i][2]=(m5[0]-m5[1])*13;
    predicted_block[i][1]=m5[3]*17+m5[2]*7;
    predicted_block[i][3]=m5[3]*7-m5[2]*17;
  }

  // Quant
  quant_set1=img->qpsp;
  for (j=0;j < BLOCK_SIZE; j++)
    for (i=0; i < BLOCK_SIZE; i++)
      img->m7[i][j]=sign((abs(predicted_block[i][j])*JQ[quant_set1]+JQQ2)/ JQQ,predicted_block[i][j])*JQ1[quant_set1];

  //     IDCT.
  //     horizontal

  for (j=0; j < BLOCK_SIZE; j++)
  {
    for (i=0; i < BLOCK_SIZE; i++)
    {
      m5[i]=img->m7[i][j];
    }
    m6[0]=(m5[0]+m5[2])*13;
    m6[1]=(m5[0]-m5[2])*13;
    m6[2]=m5[1]*7-m5[3]*17;
    m6[3]=m5[1]*17+m5[3]*7;

    for (i=0; i < 2; i++)
    {
      i1=3-i;
      img->m7[i][j]=m6[i]+m6[i1];
      img->m7[i1][j]=m6[i]-m6[i1];
    }
  }

  //  vertical

  for (i=0; i < BLOCK_SIZE; i++)
  {
    for (j=0; j < BLOCK_SIZE; j++)
    {
      m5[j]=img->m7[i][j];
    }
    m6[0]=(m5[0]+m5[2])*13;
    m6[1]=(m5[0]-m5[2])*13;
    m6[2]=m5[1]*7-m5[3]*17;
    m6[3]=m5[1]*17+m5[3]*7;

    for (j=0; j < 2; j++)
    {
      j1=3-j;
      img->m7[i][j] =min(255,max(0,(m6[j]+m6[j1]+JQQ2)/JQQ));
      img->m7[i][j1]=min(255,max(0,(m6[j]-m6[j1]+JQQ2)/JQQ));
    }
  }

  //  Decoded block moved to frame memory

  for (j=0; j < BLOCK_SIZE; j++)
    for (i=0; i < BLOCK_SIZE; i++)
      imgY[img->pix_y+block_y+j][img->pix_x+block_x+i]=img->m7[i][j];

}

void itrans_sp_chroma(struct img_par *img,int ll)
{
  int i,j,i1,j2,ilev,n2,n1,j1,mb_y,qp_const,quant_set;
  int m5[BLOCK_SIZE];
  int predicted_chroma_block[MB_BLOCK_SIZE/2][MB_BLOCK_SIZE/2],Fq1q2,mp1[BLOCK_SIZE],quant_set1;

  qp_const=JQQ4;

  for (j=0; j < MB_BLOCK_SIZE/2; j++)
    for (i=0; i < MB_BLOCK_SIZE/2; i++)
    {
      predicted_chroma_block[i][j]=img->mpr[i][j];
      img->mpr[i][j]=0;
    }

  for (n2=0; n2 <= BLOCK_SIZE; n2 += BLOCK_SIZE)
  {
    for (n1=0; n1 <= BLOCK_SIZE; n1 += BLOCK_SIZE)
    {

      //  Horizontal transform.
      for (j=0; j < BLOCK_SIZE; j++)
      {
        mb_y=n2+j;
        for (i=0; i < 2; i++)
        {
          i1=3-i;
          m5[i]=predicted_chroma_block[i+n1][mb_y]+predicted_chroma_block[i1+n1][mb_y];
          m5[i1]=predicted_chroma_block[i+n1][mb_y]-predicted_chroma_block[i1+n1][mb_y];
        }
        predicted_chroma_block[n1][mb_y]=(m5[0]+m5[1])*13;
        predicted_chroma_block[n1+2][mb_y]=(m5[0]-m5[1])*13;
        predicted_chroma_block[n1+1][mb_y]=m5[3]*17+m5[2]*7;
        predicted_chroma_block[n1+3][mb_y]=m5[3]*7-m5[2]*17;
      }

      //  Vertical transform.

      for (i=0; i < BLOCK_SIZE; i++)
      {
        j1=n1+i;
        for (j=0; j < 2; j++)
        {
          j2=3-j;
          m5[j]=predicted_chroma_block[j1][n2+j]+predicted_chroma_block[j1][n2+j2];
          m5[j2]=predicted_chroma_block[j1][n2+j]-predicted_chroma_block[j1][n2+j2];
        }
        predicted_chroma_block[j1][n2+0]=(m5[0]+m5[1])*13;
        predicted_chroma_block[j1][n2+2]=(m5[0]-m5[1])*13;
        predicted_chroma_block[j1][n2+1]=m5[3]*17+m5[2]*7;
        predicted_chroma_block[j1][n2+3]=m5[3]*7-m5[2]*17;
      }
    }
  }

  //     2X2 transform of DC coeffs.
  mp1[0]=(predicted_chroma_block[0][0]+predicted_chroma_block[4][0]+predicted_chroma_block[0][4]+predicted_chroma_block[4][4])/2;
  mp1[1]=(predicted_chroma_block[0][0]-predicted_chroma_block[4][0]+predicted_chroma_block[0][4]-predicted_chroma_block[4][4])/2;
  mp1[2]=(predicted_chroma_block[0][0]+predicted_chroma_block[4][0]-predicted_chroma_block[0][4]-predicted_chroma_block[4][4])/2;
  mp1[3]=(predicted_chroma_block[0][0]-predicted_chroma_block[4][0]-predicted_chroma_block[0][4]+predicted_chroma_block[4][4])/2;

  quant_set=QP_SCALE_CR[img->qp];
  quant_set1=QP_SCALE_CR[img->qpsp];
  Fq1q2=(JQQ*JQ[quant_set1]+JQ[quant_set]/2)/JQ[quant_set];

  ilev=img->cof[0+ll][4][0][0]*Fq1q2+mp1[0]*JQ[quant_set1];
  mp1[0]=sign((abs(ilev)+JQQ2)/ JQQ,ilev)*JQ1[quant_set1];
  ilev=img->cof[1+ll][4][0][0]*Fq1q2+mp1[1]*JQ[quant_set1];
  mp1[1]=sign((abs(ilev)+JQQ2)/ JQQ,ilev)*JQ1[quant_set1];
  ilev=img->cof[0+ll][5][0][0]*Fq1q2+mp1[2]*JQ[quant_set1];
  mp1[2]=sign((abs(ilev)+JQQ2)/ JQQ,ilev)*JQ1[quant_set1];
  ilev=img->cof[1+ll][5][0][0]*Fq1q2+mp1[3]*JQ[quant_set1];
  mp1[3]=sign((abs(ilev)+JQQ2)/ JQQ,ilev)*JQ1[quant_set1];

  img->cof[0+ll][4][0][0]=(mp1[0]+mp1[1]+mp1[2]+mp1[3])/2;
  img->cof[1+ll][4][0][0]=(mp1[0]-mp1[1]+mp1[2]-mp1[3])/2;
  img->cof[0+ll][5][0][0]=(mp1[0]+mp1[1]-mp1[2]-mp1[3])/2;
  img->cof[1+ll][5][0][0]=(mp1[0]-mp1[1]-mp1[2]+mp1[3])/2;

  for (n2=0; n2 <= BLOCK_SIZE; n2 += BLOCK_SIZE)
    for (n1=0; n1 <= BLOCK_SIZE; n1 += BLOCK_SIZE)
      for (i=0;i< BLOCK_SIZE; i++)
        for (j=0;j< BLOCK_SIZE; j++)
        if ((i!=0) || (j!=0))
        {
          ilev=img->cof[n1/BLOCK_SIZE+ll][4+n2/BLOCK_SIZE][i][j]/JQ1[QP_SCALE_CR[img->qp]]*Fq1q2+predicted_chroma_block[n1+i][n2+j]*JQ[quant_set1];
          img->cof[n1/BLOCK_SIZE+ll][4+n2/BLOCK_SIZE][i][j]=sign((abs(ilev)+JQQ2)/ JQQ,ilev)*JQ1[quant_set1];
        }
}

int sign(int a , int b)
{
  int x;

  x=abs(a);
  if (b>0)
    return(x);
  else return(-x);
}