1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
|
/*********************************************************************************
** ITU-T G.722.1 (2005-05) - Fixed point implementation for main body and Annex C
** > Software Release 2.1 (2008-06)
** (Simple repackaging; no change from 2005-05 Release 2.0 code)
**
** © 2004 Polycom, Inc.
**
** All rights reserved.
**
*********************************************************************************/
/*********************************************************************************
* Filename: dct_type_iv_a.c
*
* Purpose: Discrete Cosine Transform, Type IV used for MLT
*
* The basis functions are
*
* cos(PI*(t+0.5)*(k+0.5)/block_length)
*
* for time t and basis function number k. Due to the symmetry of the expression
* in t and k, it is clear that the forward and inverse transforms are the same.
*
*********************************************************************************/
/*********************************************************************************
Include files
*********************************************************************************/
#include "defs.h"
#include "count.h"
#include "dct4_a.h"
/*********************************************************************************
External variable declarations
*********************************************************************************/
extern Word16 anal_bias[DCT_LENGTH];
extern Word16 dct_core_a[DCT_LENGTH_DIV_32][DCT_LENGTH_DIV_32];
extern cos_msin_t a_cos_msin_2 [DCT_LENGTH_DIV_32];
extern cos_msin_t a_cos_msin_4 [DCT_LENGTH_DIV_16];
extern cos_msin_t a_cos_msin_8 [DCT_LENGTH_DIV_8];
extern cos_msin_t a_cos_msin_16[DCT_LENGTH_DIV_4];
extern cos_msin_t a_cos_msin_32[DCT_LENGTH_DIV_2];
extern cos_msin_t a_cos_msin_64[DCT_LENGTH];
extern cos_msin_t *a_cos_msin_table[];
/*********************************************************************************
Function: dct_type_iv_a
Syntax: void dct_type_iv_a (input, output, dct_length)
Word16 input[], output[], dct_length;
Description: Discrete Cosine Transform, Type IV used for MLT
Design Notes:
WMOPS: | 24kbit | 32kbit
-------|--------------|----------------
AVG | 1.14 | 1.14
-------|--------------|----------------
MAX | 1.14 | 1.14
-------|--------------|----------------
14kHz | 24kbit | 32kbit | 48kbit
-------|--------------|----------------|----------------
AVG | 2.57 | 2.57 | 2.57
-------|--------------|----------------|----------------
MAX | 2.57 | 2.57 | 2.57
-------|--------------|----------------|----------------
*********************************************************************************/
void dct_type_iv_a (Word16 *input,Word16 *output,Word16 dct_length)
{
Word16 buffer_a[MAX_DCT_LENGTH], buffer_b[MAX_DCT_LENGTH], buffer_c[MAX_DCT_LENGTH];
Word16 *in_ptr, *in_ptr_low, *in_ptr_high, *next_in_base;
Word16 *out_ptr_low, *out_ptr_high, *next_out_base;
Word16 *out_buffer, *in_buffer, *buffer_swap;
Word16 in_val_low, in_val_high;
Word16 out_val_low, out_val_high;
Word16 in_low_even, in_low_odd;
Word16 in_high_even, in_high_odd;
Word16 out_low_even, out_low_odd;
Word16 out_high_even, out_high_odd;
Word16 *pair_ptr;
Word16 cos_even, cos_odd, msin_even, msin_odd;
Word16 neg_cos_odd;
Word16 neg_msin_even;
Word32 sum;
Word16 set_span, set_count, set_count_log, pairs_left, sets_left;
Word16 i,k;
Word16 index;
cos_msin_t **table_ptr_ptr, *cos_msin_ptr;
Word16 temp;
Word32 acca;
Word16 dct_length_log;
/*++++++++++++++++++++++++++++++++++++++++++++++++++++++*/
/* Do the sum/difference butterflies, the first part of */
/* converting one N-point transform into N/2 two-point */
/* transforms, where N = 1 << DCT_LENGTH_LOG. = 64/128 */
/*++++++++++++++++++++++++++++++++++++++++++++++++++++++*/
test();
if (dct_length==DCT_LENGTH)
{
dct_length_log = DCT_LENGTH_LOG;
/* Add bias offsets */
for (i=0;i<dct_length;i++)
{
input[i] = add(input[i],anal_bias[i]);
move16();
}
}
else
dct_length_log = MAX_DCT_LENGTH_LOG;
index = 0L;
move16();
in_buffer = input;
move16();
out_buffer = buffer_a;
move16();
temp = sub(dct_length_log,2);
for (set_count_log=0;set_count_log<=temp;set_count_log++)
{
/*===========================================================*/
/* Initialization for the loop over sets at the current size */
/*===========================================================*/
/* set_span = 1 << (DCT_LENGTH_LOG - set_count_log); */
set_span = shr(dct_length,set_count_log);
set_count = shl(1,set_count_log);
in_ptr = in_buffer;
move16();
next_out_base = out_buffer;
move16();
/*=====================================*/
/* Loop over all the sets of this size */
/*=====================================*/
for (sets_left=set_count;sets_left>0;sets_left--)
{
/*||||||||||||||||||||||||||||||||||||||||||||*/
/* Set up output pointers for the current set */
/*||||||||||||||||||||||||||||||||||||||||||||*/
out_ptr_low = next_out_base;
next_out_base = next_out_base + set_span;
out_ptr_high = next_out_base;
/*||||||||||||||||||||||||||||||||||||||||||||||||||*/
/* Loop over all the butterflies in the current set */
/*||||||||||||||||||||||||||||||||||||||||||||||||||*/
do
{
in_val_low = *in_ptr++;
in_val_high = *in_ptr++;
acca = L_add(in_val_low,in_val_high);
acca = L_shr(acca,1);
out_val_low = extract_l(acca);
acca = L_sub(in_val_low,in_val_high);
acca = L_shr(acca,1);
out_val_high = extract_l(acca);
*out_ptr_low++ = out_val_low;
*--out_ptr_high = out_val_high;
test();
} while (out_ptr_low < out_ptr_high);
} /* End of loop over sets of the current size */
/*============================================================*/
/* Decide which buffers to use as input and output next time. */
/* Except for the first time (when the input buffer is the */
/* subroutine input) we just alternate the local buffers. */
/*============================================================*/
in_buffer = out_buffer;
move16();
if (out_buffer == buffer_a)
out_buffer = buffer_b;
else
out_buffer = buffer_a;
index = add(index,1);
} /* End of loop over set sizes */
/*++++++++++++++++++++++++++++++++*/
/* Do N/2 two-point transforms, */
/* where N = 1 << DCT_LENGTH_LOG */
/*++++++++++++++++++++++++++++++++*/
pair_ptr = in_buffer;
move16();
buffer_swap = buffer_c;
move16();
temp = sub(dct_length_log,1);
temp = shl(1,temp);
for (pairs_left=temp; pairs_left > 0; pairs_left--)
{
for ( k=0; k<CORE_SIZE; k++ )
{
sum=0L;
move32();
for ( i=0; i<CORE_SIZE; i++ )
{
sum = L_mac(sum, pair_ptr[i],dct_core_a[i][k]);
}
buffer_swap[k] = round(sum);
}
/* address arithmetic */
pair_ptr += CORE_SIZE;
buffer_swap += CORE_SIZE;
}
for (i=0;i<dct_length;i++)
{
in_buffer[i] = buffer_c[i];
move16();
}
table_ptr_ptr = a_cos_msin_table;
/*++++++++++++++++++++++++++++++*/
/* Perform rotation butterflies */
/*++++++++++++++++++++++++++++++*/
temp = sub(dct_length_log,2);
for (set_count_log = temp; set_count_log >= 0; set_count_log--)
{
/*===========================================================*/
/* Initialization for the loop over sets at the current size */
/*===========================================================*/
/* set_span = 1 << (DCT_LENGTH_LOG - set_count_log); */
set_span = shr(dct_length,set_count_log);
set_count = shl(1,set_count_log);
next_in_base = in_buffer;
move16();
test();
if (set_count_log == 0)
{
next_out_base = output;
}
else
{
next_out_base = out_buffer;
}
/*=====================================*/
/* Loop over all the sets of this size */
/*=====================================*/
for (sets_left = set_count; sets_left > 0;sets_left--)
{
/*|||||||||||||||||||||||||||||||||||||||||*/
/* Set up the pointers for the current set */
/*|||||||||||||||||||||||||||||||||||||||||*/
in_ptr_low = next_in_base;
move16();
temp = shr(set_span,1);
/* address arithmetic */
in_ptr_high = in_ptr_low + temp;
next_in_base += set_span;
out_ptr_low = next_out_base;
next_out_base += set_span;
out_ptr_high = next_out_base;
cos_msin_ptr = *table_ptr_ptr;
/*||||||||||||||||||||||||||||||||||||||||||||||||||||||*/
/* Loop over all the butterfly pairs in the current set */
/*||||||||||||||||||||||||||||||||||||||||||||||||||||||*/
do
{
/* address arithmetic */
in_low_even = *in_ptr_low++;
in_low_odd = *in_ptr_low++;
in_high_even = *in_ptr_high++;
in_high_odd = *in_ptr_high++;
cos_even = cos_msin_ptr[0].cosine;
move16();
msin_even = cos_msin_ptr[0].minus_sine;
move16();
cos_odd = cos_msin_ptr[1].cosine;
move16();
msin_odd = cos_msin_ptr[1].minus_sine;
move16();
cos_msin_ptr += 2;
sum = 0L;
sum=L_mac(sum,cos_even,in_low_even);
neg_msin_even = negate(msin_even);
sum=L_mac(sum,neg_msin_even,in_high_even);
out_low_even = round(sum);
sum = 0L;
sum=L_mac(sum,msin_even,in_low_even);
sum=L_mac(sum,cos_even,in_high_even);
out_high_even= round(sum);
sum = 0L;
sum=L_mac(sum,cos_odd,in_low_odd);
sum=L_mac(sum,msin_odd,in_high_odd);
out_low_odd= round(sum);
sum = 0L;
sum=L_mac(sum,msin_odd,in_low_odd);
neg_cos_odd = negate(cos_odd);
sum=L_mac(sum,neg_cos_odd,in_high_odd);
out_high_odd= round(sum);
*out_ptr_low++ = out_low_even;
*--out_ptr_high = out_high_even;
*out_ptr_low++ = out_low_odd;
*--out_ptr_high = out_high_odd;
test();
} while (out_ptr_low < out_ptr_high);
} /* End of loop over sets of the current size */
/*=============================================*/
/* Swap input and output buffers for next time */
/*=============================================*/
buffer_swap = in_buffer;
in_buffer = out_buffer;
out_buffer = buffer_swap;
table_ptr_ptr++;
}
}
|
