| 1 |
/* |
| 2 |
INT123_dct64.c: DCT64, the plain C version |
| 3 |
|
| 4 |
copyright ?-2006 by the mpg123 project - free software under the terms of the LGPL 2.1 |
| 5 |
see COPYING and AUTHORS files in distribution or http://mpg123.org |
| 6 |
initially written by Michael Hipp |
| 7 |
*/ |
| 8 |
|
| 9 |
/* |
| 10 |
* Discrete Cosine Tansform (DCT) for subband synthesis |
| 11 |
* |
| 12 |
* -funroll-loops (for gcc) will remove the loops for better performance |
| 13 |
* using loops in the source-code enhances readabillity |
| 14 |
* |
| 15 |
* |
| 16 |
* TODO: write an optimized version for the down-sampling modes |
| 17 |
* (in these modes the bands 16-31 (2:1) or 8-31 (4:1) are zero |
| 18 |
*/ |
| 19 |
|
| 20 |
#include "mpg123lib_intern.h" |
| 21 |
|
| 22 |
void INT123_dct64(real *out0,real *out1,real *samples) |
| 23 |
{ |
| 24 |
real bufs[64]; |
| 25 |
|
| 26 |
{ |
| 27 |
register int i,j; |
| 28 |
register real *b1,*b2,*bs; |
| 29 |
register const real *costab; |
| 30 |
|
| 31 |
b1 = samples; |
| 32 |
bs = bufs; |
| 33 |
costab = INT123_pnts[0]+16; |
| 34 |
b2 = b1 + 32; |
| 35 |
|
| 36 |
for(i=15;i>=0;i--) |
| 37 |
*bs++ = (*b1++ + *--b2); |
| 38 |
for(i=15;i>=0;i--) |
| 39 |
*bs++ = REAL_MUL((*--b2 - *b1++), *--costab); |
| 40 |
|
| 41 |
b1 = bufs; |
| 42 |
costab = INT123_pnts[1]+8; |
| 43 |
b2 = b1 + 16; |
| 44 |
|
| 45 |
{ |
| 46 |
for(i=7;i>=0;i--) |
| 47 |
*bs++ = (*b1++ + *--b2); |
| 48 |
for(i=7;i>=0;i--) |
| 49 |
*bs++ = REAL_MUL((*--b2 - *b1++), *--costab); |
| 50 |
b2 += 32; |
| 51 |
costab += 8; |
| 52 |
for(i=7;i>=0;i--) |
| 53 |
*bs++ = (*b1++ + *--b2); |
| 54 |
for(i=7;i>=0;i--) |
| 55 |
*bs++ = REAL_MUL((*b1++ - *--b2), *--costab); |
| 56 |
b2 += 32; |
| 57 |
} |
| 58 |
|
| 59 |
bs = bufs; |
| 60 |
costab = INT123_pnts[2]; |
| 61 |
b2 = b1 + 8; |
| 62 |
|
| 63 |
for(j=2;j;j--) |
| 64 |
{ |
| 65 |
for(i=3;i>=0;i--) |
| 66 |
*bs++ = (*b1++ + *--b2); |
| 67 |
for(i=3;i>=0;i--) |
| 68 |
*bs++ = REAL_MUL((*--b2 - *b1++), costab[i]); |
| 69 |
b2 += 16; |
| 70 |
for(i=3;i>=0;i--) |
| 71 |
*bs++ = (*b1++ + *--b2); |
| 72 |
for(i=3;i>=0;i--) |
| 73 |
*bs++ = REAL_MUL((*b1++ - *--b2), costab[i]); |
| 74 |
b2 += 16; |
| 75 |
} |
| 76 |
|
| 77 |
b1 = bufs; |
| 78 |
costab = INT123_pnts[3]; |
| 79 |
b2 = b1 + 4; |
| 80 |
|
| 81 |
for(j=4;j;j--) |
| 82 |
{ |
| 83 |
*bs++ = (*b1++ + *--b2); |
| 84 |
*bs++ = (*b1++ + *--b2); |
| 85 |
*bs++ = REAL_MUL((*--b2 - *b1++), costab[1]); |
| 86 |
*bs++ = REAL_MUL((*--b2 - *b1++), costab[0]); |
| 87 |
b2 += 8; |
| 88 |
*bs++ = (*b1++ + *--b2); |
| 89 |
*bs++ = (*b1++ + *--b2); |
| 90 |
*bs++ = REAL_MUL((*b1++ - *--b2), costab[1]); |
| 91 |
*bs++ = REAL_MUL((*b1++ - *--b2), costab[0]); |
| 92 |
b2 += 8; |
| 93 |
} |
| 94 |
bs = bufs; |
| 95 |
costab = INT123_pnts[4]; |
| 96 |
|
| 97 |
for(j=8;j;j--) |
| 98 |
{ |
| 99 |
real v0,v1; |
| 100 |
v0=*b1++; v1 = *b1++; |
| 101 |
*bs++ = (v0 + v1); |
| 102 |
*bs++ = REAL_MUL((v0 - v1), (*costab)); |
| 103 |
v0=*b1++; v1 = *b1++; |
| 104 |
*bs++ = (v0 + v1); |
| 105 |
*bs++ = REAL_MUL((v1 - v0), (*costab)); |
| 106 |
} |
| 107 |
|
| 108 |
} |
| 109 |
|
| 110 |
|
| 111 |
{ |
| 112 |
register real *b1; |
| 113 |
register int i; |
| 114 |
|
| 115 |
for(b1=bufs,i=8;i;i--,b1+=4) |
| 116 |
b1[2] += b1[3]; |
| 117 |
|
| 118 |
for(b1=bufs,i=4;i;i--,b1+=8) |
| 119 |
{ |
| 120 |
b1[4] += b1[6]; |
| 121 |
b1[6] += b1[5]; |
| 122 |
b1[5] += b1[7]; |
| 123 |
} |
| 124 |
|
| 125 |
for(b1=bufs,i=2;i;i--,b1+=16) |
| 126 |
{ |
| 127 |
b1[8] += b1[12]; |
| 128 |
b1[12] += b1[10]; |
| 129 |
b1[10] += b1[14]; |
| 130 |
b1[14] += b1[9]; |
| 131 |
b1[9] += b1[13]; |
| 132 |
b1[13] += b1[11]; |
| 133 |
b1[11] += b1[15]; |
| 134 |
} |
| 135 |
} |
| 136 |
|
| 137 |
|
| 138 |
out0[0x10*16] = REAL_SCALE_DCT64(bufs[0]); |
| 139 |
out0[0x10*15] = REAL_SCALE_DCT64(bufs[16+0] + bufs[16+8]); |
| 140 |
out0[0x10*14] = REAL_SCALE_DCT64(bufs[8]); |
| 141 |
out0[0x10*13] = REAL_SCALE_DCT64(bufs[16+8] + bufs[16+4]); |
| 142 |
out0[0x10*12] = REAL_SCALE_DCT64(bufs[4]); |
| 143 |
out0[0x10*11] = REAL_SCALE_DCT64(bufs[16+4] + bufs[16+12]); |
| 144 |
out0[0x10*10] = REAL_SCALE_DCT64(bufs[12]); |
| 145 |
out0[0x10* 9] = REAL_SCALE_DCT64(bufs[16+12] + bufs[16+2]); |
| 146 |
out0[0x10* 8] = REAL_SCALE_DCT64(bufs[2]); |
| 147 |
out0[0x10* 7] = REAL_SCALE_DCT64(bufs[16+2] + bufs[16+10]); |
| 148 |
out0[0x10* 6] = REAL_SCALE_DCT64(bufs[10]); |
| 149 |
out0[0x10* 5] = REAL_SCALE_DCT64(bufs[16+10] + bufs[16+6]); |
| 150 |
out0[0x10* 4] = REAL_SCALE_DCT64(bufs[6]); |
| 151 |
out0[0x10* 3] = REAL_SCALE_DCT64(bufs[16+6] + bufs[16+14]); |
| 152 |
out0[0x10* 2] = REAL_SCALE_DCT64(bufs[14]); |
| 153 |
out0[0x10* 1] = REAL_SCALE_DCT64(bufs[16+14] + bufs[16+1]); |
| 154 |
out0[0x10* 0] = REAL_SCALE_DCT64(bufs[1]); |
| 155 |
|
| 156 |
out1[0x10* 0] = REAL_SCALE_DCT64(bufs[1]); |
| 157 |
out1[0x10* 1] = REAL_SCALE_DCT64(bufs[16+1] + bufs[16+9]); |
| 158 |
out1[0x10* 2] = REAL_SCALE_DCT64(bufs[9]); |
| 159 |
out1[0x10* 3] = REAL_SCALE_DCT64(bufs[16+9] + bufs[16+5]); |
| 160 |
out1[0x10* 4] = REAL_SCALE_DCT64(bufs[5]); |
| 161 |
out1[0x10* 5] = REAL_SCALE_DCT64(bufs[16+5] + bufs[16+13]); |
| 162 |
out1[0x10* 6] = REAL_SCALE_DCT64(bufs[13]); |
| 163 |
out1[0x10* 7] = REAL_SCALE_DCT64(bufs[16+13] + bufs[16+3]); |
| 164 |
out1[0x10* 8] = REAL_SCALE_DCT64(bufs[3]); |
| 165 |
out1[0x10* 9] = REAL_SCALE_DCT64(bufs[16+3] + bufs[16+11]); |
| 166 |
out1[0x10*10] = REAL_SCALE_DCT64(bufs[11]); |
| 167 |
out1[0x10*11] = REAL_SCALE_DCT64(bufs[16+11] + bufs[16+7]); |
| 168 |
out1[0x10*12] = REAL_SCALE_DCT64(bufs[7]); |
| 169 |
out1[0x10*13] = REAL_SCALE_DCT64(bufs[16+7] + bufs[16+15]); |
| 170 |
out1[0x10*14] = REAL_SCALE_DCT64(bufs[15]); |
| 171 |
out1[0x10*15] = REAL_SCALE_DCT64(bufs[16+15]); |
| 172 |
|
| 173 |
} |
| 174 |
|
| 175 |
|