jrevdct.c

32位操作系统OS/2的MPEG播放机

		    tmp3 += z1 + z4;
		} else {
		    /* d1 == 0, d3 != 0, d5 != 0, d7 != 0 */
		    z1 = d7;
		    z2 = d5 + d3;
		    z3 = d7 + d3;
		    z5 = MULTIPLY(z3 + d5, FIX(1.175875602));
		    
		    tmp0 = MULTIPLY(d7, FIX(0.298631336)); 
		    tmp1 = MULTIPLY(d5, FIX(2.053119869));
		    tmp2 = MULTIPLY(d3, FIX(3.072711026));
		    z1 = MULTIPLY(d7, - FIX(0.899976223));
		    z2 = MULTIPLY(z2, - FIX(2.562915447));
		    z3 = MULTIPLY(z3, - FIX(1.961570560));
		    z4 = MULTIPLY(d5, - FIX(0.390180644));
		    
		    z3 += z5;
		    z4 += z5;
		    
		    tmp0 += z1 + z3;
		    tmp1 += z2 + z4;
		    tmp2 += z2 + z3;
		    tmp3 = z1 + z4;
		}
	    } else {
		if (d1) {
		    /* d1 != 0, d3 == 0, d5 != 0, d7 != 0 */
		    z1 = d7 + d1;
		    z2 = d5;
		    z3 = d7;
		    z4 = d5 + d1;
		    z5 = MULTIPLY(z3 + z4, FIX(1.175875602));
		    
		    tmp0 = MULTIPLY(d7, FIX(0.298631336)); 
		    tmp1 = MULTIPLY(d5, FIX(2.053119869));
		    tmp3 = MULTIPLY(d1, FIX(1.501321110));
		    z1 = MULTIPLY(z1, - FIX(0.899976223));
		    z2 = MULTIPLY(d5, - FIX(2.562915447));
		    z3 = MULTIPLY(d7, - FIX(1.961570560));
		    z4 = MULTIPLY(z4, - FIX(0.390180644));
		    
		    z3 += z5;
		    z4 += z5;
		    
		    tmp0 += z1 + z3;
		    tmp1 += z2 + z4;
		    tmp2 = z2 + z3;
		    tmp3 += z1 + z4;
		} else {
		    /* d1 == 0, d3 == 0, d5 != 0, d7 != 0 */
		    tmp0 = MULTIPLY(d7, - FIX(0.601344887)); 
		    z1 = MULTIPLY(d7, - FIX(0.899976223));
		    z3 = MULTIPLY(d7, - FIX(1.961570560));
		    tmp1 = MULTIPLY(d5, - FIX(0.509795578));
		    z2 = MULTIPLY(d5, - FIX(2.562915447));
		    z4 = MULTIPLY(d5, - FIX(0.390180644));
		    z5 = MULTIPLY(d5 + d7, FIX(1.175875602));
		    
		    z3 += z5;
		    z4 += z5;
		    
		    tmp0 += z3;
		    tmp1 += z4;
		    tmp2 = z2 + z3;
		    tmp3 = z1 + z4;
		}
	    }
	} else {
	    if (d3) {
		if (d1) {
		    /* d1 != 0, d3 != 0, d5 == 0, d7 != 0 */
		    z1 = d7 + d1;
		    z3 = d7 + d3;
		    z5 = MULTIPLY(z3 + d1, FIX(1.175875602));
		    
		    tmp0 = MULTIPLY(d7, FIX(0.298631336)); 
		    tmp2 = MULTIPLY(d3, FIX(3.072711026));
		    tmp3 = MULTIPLY(d1, FIX(1.501321110));
		    z1 = MULTIPLY(z1, - FIX(0.899976223));
		    z2 = MULTIPLY(d3, - FIX(2.562915447));
		    z3 = MULTIPLY(z3, - FIX(1.961570560));
		    z4 = MULTIPLY(d1, - FIX(0.390180644));
		    
		    z3 += z5;
		    z4 += z5;
		    
		    tmp0 += z1 + z3;
		    tmp1 = z2 + z4;
		    tmp2 += z2 + z3;
		    tmp3 += z1 + z4;
		} else {
		    /* d1 == 0, d3 != 0, d5 == 0, d7 != 0 */
		    z3 = d7 + d3;
		    
		    tmp0 = MULTIPLY(d7, - FIX(0.601344887)); 
		    z1 = MULTIPLY(d7, - FIX(0.899976223));
		    tmp2 = MULTIPLY(d3, FIX(0.509795579));
		    z2 = MULTIPLY(d3, - FIX(2.562915447));
		    z5 = MULTIPLY(z3, FIX(1.175875602));
		    z3 = MULTIPLY(z3, - FIX(0.785694958));
		    
		    tmp0 += z3;
		    tmp1 = z2 + z5;
		    tmp2 += z3;
		    tmp3 = z1 + z5;
		}
	    } else {
		if (d1) {
		    /* d1 != 0, d3 == 0, d5 == 0, d7 != 0 */
		    z1 = d7 + d1;
		    z5 = MULTIPLY(z1, FIX(1.175875602));

		    z1 = MULTIPLY(z1, FIX(0.275899379));
		    z3 = MULTIPLY(d7, - FIX(1.961570560));
		    tmp0 = MULTIPLY(d7, - FIX(1.662939224)); 
		    z4 = MULTIPLY(d1, - FIX(0.390180644));
		    tmp3 = MULTIPLY(d1, FIX(1.111140466));

		    tmp0 += z1;
		    tmp1 = z4 + z5;
		    tmp2 = z3 + z5;
		    tmp3 += z1;
		} else {
		    /* d1 == 0, d3 == 0, d5 == 0, d7 != 0 */
		    tmp0 = MULTIPLY(d7, - FIX(1.387039845));
		    tmp1 = MULTIPLY(d7, FIX(1.175875602));
		    tmp2 = MULTIPLY(d7, - FIX(0.785694958));
		    tmp3 = MULTIPLY(d7, FIX(0.275899379));
		}
	    }
	}
    } else {
	if (d5) {
	    if (d3) {
		if (d1) {
		    /* d1 != 0, d3 != 0, d5 != 0, d7 == 0 */
		    z2 = d5 + d3;
		    z4 = d5 + d1;
		    z5 = MULTIPLY(d3 + z4, FIX(1.175875602));
		    
		    tmp1 = MULTIPLY(d5, FIX(2.053119869));
		    tmp2 = MULTIPLY(d3, FIX(3.072711026));
		    tmp3 = MULTIPLY(d1, FIX(1.501321110));
		    z1 = MULTIPLY(d1, - FIX(0.899976223));
		    z2 = MULTIPLY(z2, - FIX(2.562915447));
		    z3 = MULTIPLY(d3, - FIX(1.961570560));
		    z4 = MULTIPLY(z4, - FIX(0.390180644));
		    
		    z3 += z5;
		    z4 += z5;
		    
		    tmp0 = z1 + z3;
		    tmp1 += z2 + z4;
		    tmp2 += z2 + z3;
		    tmp3 += z1 + z4;
		} else {
		    /* d1 == 0, d3 != 0, d5 != 0, d7 == 0 */
		    z2 = d5 + d3;
		    
		    z5 = MULTIPLY(z2, FIX(1.175875602));
		    tmp1 = MULTIPLY(d5, FIX(1.662939225));
		    z4 = MULTIPLY(d5, - FIX(0.390180644));
		    z2 = MULTIPLY(z2, - FIX(1.387039845));
		    tmp2 = MULTIPLY(d3, FIX(1.111140466));
		    z3 = MULTIPLY(d3, - FIX(1.961570560));
		    
		    tmp0 = z3 + z5;
		    tmp1 += z2;
		    tmp2 += z2;
		    tmp3 = z4 + z5;
		}
	    } else {
		if (d1) {
		    /* d1 != 0, d3 == 0, d5 != 0, d7 == 0 */
		    z4 = d5 + d1;
		    
		    z5 = MULTIPLY(z4, FIX(1.175875602));
		    z1 = MULTIPLY(d1, - FIX(0.899976223));
		    tmp3 = MULTIPLY(d1, FIX(0.601344887));
		    tmp1 = MULTIPLY(d5, - FIX(0.509795578));
		    z2 = MULTIPLY(d5, - FIX(2.562915447));
		    z4 = MULTIPLY(z4, FIX(0.785694958));
		    
		    tmp0 = z1 + z5;
		    tmp1 += z4;
		    tmp2 = z2 + z5;
		    tmp3 += z4;
		} else {
		    /* d1 == 0, d3 == 0, d5 != 0, d7 == 0 */
		    tmp0 = MULTIPLY(d5, FIX(1.175875602));
		    tmp1 = MULTIPLY(d5, FIX(0.275899380));
		    tmp2 = MULTIPLY(d5, - FIX(1.387039845));
		    tmp3 = MULTIPLY(d5, FIX(0.785694958));
		}
	    }
	} else {
	    if (d3) {
		if (d1) {
		    /* d1 != 0, d3 != 0, d5 == 0, d7 == 0 */
		    z5 = d1 + d3;
		    tmp3 = MULTIPLY(d1, FIX(0.211164243));
		    tmp2 = MULTIPLY(d3, - FIX(1.451774981));
		    z1 = MULTIPLY(d1, FIX(1.061594337));
		    z2 = MULTIPLY(d3, - FIX(2.172734803));
		    z4 = MULTIPLY(z5, FIX(0.785694958));
		    z5 = MULTIPLY(z5, FIX(1.175875602));
		    
		    tmp0 = z1 - z4;
		    tmp1 = z2 + z4;
		    tmp2 += z5;
		    tmp3 += z5;
		} else {
		    /* d1 == 0, d3 != 0, d5 == 0, d7 == 0 */
		    tmp0 = MULTIPLY(d3, - FIX(0.785694958));
		    tmp1 = MULTIPLY(d3, - FIX(1.387039845));
		    tmp2 = MULTIPLY(d3, - FIX(0.275899379));
		    tmp3 = MULTIPLY(d3, FIX(1.175875602));
		}
	    } else {
		if (d1) {
		    /* d1 != 0, d3 == 0, d5 == 0, d7 == 0 */
		    tmp0 = MULTIPLY(d1, FIX(0.275899379));
		    tmp1 = MULTIPLY(d1, FIX(0.785694958));
		    tmp2 = MULTIPLY(d1, FIX(1.175875602));
		    tmp3 = MULTIPLY(d1, FIX(1.387039845));
		} else {
		    /* d1 == 0, d3 == 0, d5 == 0, d7 == 0 */
		    tmp0 = tmp1 = tmp2 = tmp3 = 0;
		}
	    }
	}
    }

    /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */

    dataptr[DCTSIZE*0] = (DCTELEM) DESCALE(tmp10 + tmp3,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp10 - tmp3,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp11 + tmp2,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*6] = (DCTELEM) DESCALE(tmp11 - tmp2,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(tmp12 + tmp1,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp12 - tmp1,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp13 + tmp0,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(tmp13 - tmp0,
					   CONST_BITS+PASS1_BITS+3);
    
    dataptr++;			/* advance pointer to next column */
  }
}

#else


void
j_rev_dct_sparse (data, pos) 
     DCTBLOCK data;
     int pos;
{
  j_rev_dct(data);
}

void
j_rev_dct (data)
  DCTBLOCK data;
{
  INT32 tmp0, tmp1, tmp2, tmp3;
  INT32 tmp10, tmp11, tmp12, tmp13;
  INT32 z1, z2, z3, z4, z5;
  register DCTELEM *dataptr;
  int rowctr;
  SHIFT_TEMPS

  /* Pass 1: process rows. */
  /* Note results are scaled up by sqrt(8) compared to a true IDCT; */
  /* furthermore, we scale the results by 2**PASS1_BITS. */

  dataptr = data;
  for (rowctr = DCTSIZE-1; rowctr >= 0; rowctr--) {
    /* Due to quantization, we will usually find that many of the input
     * coefficients are zero, especially the AC terms.  We can exploit this
     * by short-circuiting the IDCT calculation for any row in which all
     * the AC terms are zero.  In that case each output is equal to the
     * DC coefficient (with scale factor as needed).
     * With typical images and quantization tables, half or more of the
     * row DCT calculations can be simplified this way.
     */

    if ((dataptr[1] | dataptr[2] | dataptr[3] | dataptr[4] |
	 dataptr[5] | dataptr[6] | dataptr[7]) == 0) {
      /* AC terms all zero */
      DCTELEM dcval = (DCTELEM) (dataptr[0] << PASS1_BITS);
      
      dataptr[0] = dcval;
      dataptr[1] = dcval;
      dataptr[2] = dcval;
      dataptr[3] = dcval;
      dataptr[4] = dcval;
      dataptr[5] = dcval;
      dataptr[6] = dcval;
      dataptr[7] = dcval;
      
      dataptr += DCTSIZE;	/* advance pointer to next row */
      continue;
    }

    /* Even part: reverse the even part of the forward DCT. */
    /* The rotator is sqrt(2)*c(-6). */

    z2 = (INT32) dataptr[2];
    z3 = (INT32) dataptr[6];

    z1 = MULTIPLY(z2 + z3, FIX(0.541196100));
    tmp2 = z1 + MULTIPLY(z3, - FIX(1.847759065));
    tmp3 = z1 + MULTIPLY(z2, FIX(0.765366865));

    tmp0 = ((INT32) dataptr[0] + (INT32) dataptr[4]) << CONST_BITS;
    tmp1 = ((INT32) dataptr[0] - (INT32) dataptr[4]) << CONST_BITS;

    tmp10 = tmp0 + tmp3;
    tmp13 = tmp0 - tmp3;
    tmp11 = tmp1 + tmp2;
    tmp12 = tmp1 - tmp2;
    
    /* Odd part per figure 8; the matrix is unitary and hence its
     * transpose is its inverse.  i0..i3 are y7,y5,y3,y1 respectively.
     */

    tmp0 = (INT32) dataptr[7];
    tmp1 = (INT32) dataptr[5];
    tmp2 = (INT32) dataptr[3];
    tmp3 = (INT32) dataptr[1];

    z1 = tmp0 + tmp3;
    z2 = tmp1 + tmp2;
    z3 = tmp0 + tmp2;
    z4 = tmp1 + tmp3;
    z5 = MULTIPLY(z3 + z4, FIX(1.175875602)); /* sqrt(2) * c3 */
    
    tmp0 = MULTIPLY(tmp0, FIX(0.298631336)); /* sqrt(2) * (-c1+c3+c5-c7) */
    tmp1 = MULTIPLY(tmp1, FIX(2.053119869)); /* sqrt(2) * ( c1+c3-c5+c7) */
    tmp2 = MULTIPLY(tmp2, FIX(3.072711026)); /* sqrt(2) * ( c1+c3+c5-c7) */
    tmp3 = MULTIPLY(tmp3, FIX(1.501321110)); /* sqrt(2) * ( c1+c3-c5-c7) */
    z1 = MULTIPLY(z1, - FIX(0.899976223)); /* sqrt(2) * (c7-c3) */
    z2 = MULTIPLY(z2, - FIX(2.562915447)); /* sqrt(2) * (-c1-c3) */
    z3 = MULTIPLY(z3, - FIX(1.961570560)); /* sqrt(2) * (-c3-c5) */
    z4 = MULTIPLY(z4, - FIX(0.390180644)); /* sqrt(2) * (c5-c3) */
    
    z3 += z5;
    z4 += z5;
    
    tmp0 += z1 + z3;
    tmp1 += z2 + z4;
    tmp2 += z2 + z3;
    tmp3 += z1 + z4;

    /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */

    dataptr[0] = (DCTELEM) DESCALE(tmp10 + tmp3, CONST_BITS-PASS1_BITS);
    dataptr[7] = (DCTELEM) DESCALE(tmp10 - tmp3, CONST_BITS-PASS1_BITS);
    dataptr[1] = (DCTELEM) DESCALE(tmp11 + tmp2, CONST_BITS-PASS1_BITS);
    dataptr[6] = (DCTELEM) DESCALE(tmp11 - tmp2, CONST_BITS-PASS1_BITS);
    dataptr[2] = (DCTELEM) DESCALE(tmp12 + tmp1, CONST_BITS-PASS1_BITS);
    dataptr[5] = (DCTELEM) DESCALE(tmp12 - tmp1, CONST_BITS-PASS1_BITS);
    dataptr[3] = (DCTELEM) DESCALE(tmp13 + tmp0, CONST_BITS-PASS1_BITS);
    dataptr[4] = (DCTELEM) DESCALE(tmp13 - tmp0, CONST_BITS-PASS1_BITS);

    dataptr += DCTSIZE;		/* advance pointer to next row */
  }

  /* Pass 2: process columns. */
  /* Note that we must descale the results by a factor of 8 == 2**3, */
  /* and also undo the PASS1_BITS scaling. */

  dataptr = data;
  for (rowctr = DCTSIZE-1; rowctr >= 0; rowctr--) {
    /* Columns of zeroes can be exploited in the same way as we did with rows.
     * However, the row calculation has created many nonzero AC terms, so the
     * simplification applies less often (typically 5% to 10% of the time).
     * On machines with very fast multiplication, it's possible that the
     * test takes more time than it's worth.  In that case this section
     * may be commented out.
     */

#ifndef NO_ZERO_COLUMN_TEST
    if ((dataptr[DCTSIZE*1] | dataptr[DCTSIZE*2] | dataptr[DCTSIZE*3] |
	 dataptr[DCTSIZE*4] | dataptr[DCTSIZE*5] | dataptr[DCTSIZE*6] |
	 dataptr[DCTSIZE*7]) == 0) {
      /* AC terms all zero */
      DCTELEM dcval = (DCTELEM) DESCALE((INT32) dataptr[0], PASS1_BITS+3);
      
      dataptr[DCTSIZE*0] = dcval;
      dataptr[DCTSIZE*1] = dcval;
      dataptr[DCTSIZE*2] = dcval;
      dataptr[DCTSIZE*3] = dcval;
      dataptr[DCTSIZE*4] = dcval;
      dataptr[DCTSIZE*5] = dcval;
      dataptr[DCTSIZE*6] = dcval;
      dataptr[DCTSIZE*7] = dcval;
      
      dataptr++;		/* advance pointer to next column */
      continue;
    }
#endif

    /* Even part: reverse the even part of the forward DCT. */
    /* The rotator is sqrt(2)*c(-6). */

    z2 = (INT32) dataptr[DCTSIZE*2];
    z3 = (INT32) dataptr[DCTSIZE*6];

    z1 = MULTIPLY(z2 + z3, FIX(0.541196100));
    tmp2 = z1 + MULTIPLY(z3, - FIX(1.847759065));
    tmp3 = z1 + MULTIPLY(z2, FIX(0.765366865));

    tmp0 = ((INT32) dataptr[DCTSIZE*0] + (INT32) dataptr[DCTSIZE*4]) << CONST_BITS;
    tmp1 = ((INT32) dataptr[DCTSIZE*0] - (INT32) dataptr[DCTSIZE*4]) << CONST_BITS;

    tmp10 = tmp0 + tmp3;
    tmp13 = tmp0 - tmp3;
    tmp11 = tmp1 + tmp2;
    tmp12 = tmp1 - tmp2;
    
    /* Odd part per figure 8; the matrix is unitary and hence its
     * transpose is its inverse.  i0..i3 are y7,y5,y3,y1 respectively.
     */

    tmp0 = (INT32) dataptr[DCTSIZE*7];
    tmp1 = (INT32) dataptr[DCTSIZE*5];
    tmp2 = (INT32) dataptr[DCTSIZE*3];
    tmp3 = (INT32) dataptr[DCTSIZE*1];

    z1 = tmp0 + tmp3;
    z2 = tmp1 + tmp2;
    z3 = tmp0 + tmp2;
    z4 = tmp1 + tmp3;
    z5 = MULTIPLY(z3 + z4, FIX(1.175875602)); /* sqrt(2) * c3 */
    
    tmp0 = MULTIPLY(tmp0, FIX(0.298631336)); /* sqrt(2) * (-c1+c3+c5-c7) */
    tmp1 = MULTIPLY(tmp1, FIX(2.053119869)); /* sqrt(2) * ( c1+c3-c5+c7) */
    tmp2 = MULTIPLY(tmp2, FIX(3.072711026)); /* sqrt(2) * ( c1+c3+c5-c7) */
    tmp3 = MULTIPLY(tmp3, FIX(1.501321110)); /* sqrt(2) * ( c1+c3-c5-c7) */
    z1 = MULTIPLY(z1, - FIX(0.899976223)); /* sqrt(2) * (c7-c3) */
    z2 = MULTIPLY(z2, - FIX(2.562915447)); /* sqrt(2) * (-c1-c3) */
    z3 = MULTIPLY(z3, - FIX(1.961570560)); /* sqrt(2) * (-c3-c5) */
    z4 = MULTIPLY(z4, - FIX(0.390180644)); /* sqrt(2) * (c5-c3) */
    
    z3 += z5;
    z4 += z5;
    
    tmp0 += z1 + z3;
    tmp1 += z2 + z4;
    tmp2 += z2 + z3;
    tmp3 += z1 + z4;

    /* Final output stage: inputs are tmp10..tmp13, tmp0..tmp3 */

    dataptr[DCTSIZE*0] = (DCTELEM) DESCALE(tmp10 + tmp3,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*7] = (DCTELEM) DESCALE(tmp10 - tmp3,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*1] = (DCTELEM) DESCALE(tmp11 + tmp2,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*6] = (DCTELEM) DESCALE(tmp11 - tmp2,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*2] = (DCTELEM) DESCALE(tmp12 + tmp1,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*5] = (DCTELEM) DESCALE(tmp12 - tmp1,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*3] = (DCTELEM) DESCALE(tmp13 + tmp0,
					   CONST_BITS+PASS1_BITS+3);
    dataptr[DCTSIZE*4] = (DCTELEM) DESCALE(tmp13 - tmp0,
					   CONST_BITS+PASS1_BITS+3);
    
    dataptr++;			/* advance pointer to next column */
  }
}


#endif