repositories / deb_ffmpeg.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
Imported Debian version 2.4.3~trusty1
[deb_ffmpeg.git] / ffmpeg / libavcodec / g723_1.c
CommitLineData
2ba45a60
DM		 1/*
2 * G.723.1 compatible decoder
3 * Copyright (c) 2006 Benjamin Larsson
4 * Copyright (c) 2010 Mohamed Naufal Basheer
5 *
6 * This file is part of FFmpeg.
7 *
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
12 *
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 */
22
23/**
24 * @file
25 * G.723.1 compatible decoder
26 */
27
28#define BITSTREAM_READER_LE
29#include "libavutil/channel_layout.h"
30#include "libavutil/mem.h"
31#include "libavutil/opt.h"
32#include "avcodec.h"
33#include "get_bits.h"
34#include "acelp_vectors.h"
35#include "celp_filters.h"
36#include "celp_math.h"
37#include "g723_1_data.h"
38#include "internal.h"
39
40#define CNG_RANDOM_SEED 12345
41
42typedef struct g723_1_context {
43 AVClass *class;
44
45 G723_1_Subframe subframe[4];
46 enum FrameType cur_frame_type;
47 enum FrameType past_frame_type;
48 enum Rate cur_rate;
49 uint8_t lsp_index[LSP_BANDS];
50 int pitch_lag[2];
51 int erased_frames;
52
53 int16_t prev_lsp[LPC_ORDER];
54 int16_t sid_lsp[LPC_ORDER];
55 int16_t prev_excitation[PITCH_MAX];
56 int16_t excitation[PITCH_MAX + FRAME_LEN + 4];
57 int16_t synth_mem[LPC_ORDER];
58 int16_t fir_mem[LPC_ORDER];
59 int iir_mem[LPC_ORDER];
60
61 int random_seed;
62 int cng_random_seed;
63 int interp_index;
64 int interp_gain;
65 int sid_gain;
66 int cur_gain;
67 int reflection_coef;
68 int pf_gain; ///< formant postfilter
69 ///< gain scaling unit memory
70 int postfilter;
71
72 int16_t audio[FRAME_LEN + LPC_ORDER + PITCH_MAX + 4];
73 int16_t prev_data[HALF_FRAME_LEN];
74 int16_t prev_weight_sig[PITCH_MAX];
75
76
77 int16_t hpf_fir_mem; ///< highpass filter fir
78 int hpf_iir_mem; ///< and iir memories
79 int16_t perf_fir_mem[LPC_ORDER]; ///< perceptual filter fir
80 int16_t perf_iir_mem[LPC_ORDER]; ///< and iir memories
81
82 int16_t harmonic_mem[PITCH_MAX];
83} G723_1_Context;
84
85static av_cold int g723_1_decode_init(AVCodecContext *avctx)
86{
87 G723_1_Context *p = avctx->priv_data;
88
89 avctx->channel_layout = AV_CH_LAYOUT_MONO;
90 avctx->sample_fmt = AV_SAMPLE_FMT_S16;
91 avctx->channels = 1;
92 p->pf_gain = 1 << 12;
93
94 memcpy(p->prev_lsp, dc_lsp, LPC_ORDER * sizeof(*p->prev_lsp));
95 memcpy(p->sid_lsp, dc_lsp, LPC_ORDER * sizeof(*p->sid_lsp));
96
97 p->cng_random_seed = CNG_RANDOM_SEED;
98 p->past_frame_type = SID_FRAME;
99
100 return 0;
101}
102
103/**
104 * Unpack the frame into parameters.
105 *
106 * @param p the context
107 * @param buf pointer to the input buffer
108 * @param buf_size size of the input buffer
109 */
110static int unpack_bitstream(G723_1_Context *p, const uint8_t *buf,
111 int buf_size)
112{
113 GetBitContext gb;
114 int ad_cb_len;
115 int temp, info_bits, i;
116
117 init_get_bits(&gb, buf, buf_size * 8);
118
119 /* Extract frame type and rate info */
120 info_bits = get_bits(&gb, 2);
121
122 if (info_bits == 3) {
123 p->cur_frame_type = UNTRANSMITTED_FRAME;
124 return 0;
125 }
126
127 /* Extract 24 bit lsp indices, 8 bit for each band */
128 p->lsp_index[2] = get_bits(&gb, 8);
129 p->lsp_index[1] = get_bits(&gb, 8);
130 p->lsp_index[0] = get_bits(&gb, 8);
131
132 if (info_bits == 2) {
133 p->cur_frame_type = SID_FRAME;
134 p->subframe[0].amp_index = get_bits(&gb, 6);
135 return 0;
136 }
137
138 /* Extract the info common to both rates */
139 p->cur_rate = info_bits ? RATE_5300 : RATE_6300;
140 p->cur_frame_type = ACTIVE_FRAME;
141
142 p->pitch_lag[0] = get_bits(&gb, 7);
143 if (p->pitch_lag[0] > 123) /* test if forbidden code */
144 return -1;
145 p->pitch_lag[0] += PITCH_MIN;
146 p->subframe[1].ad_cb_lag = get_bits(&gb, 2);
147
148 p->pitch_lag[1] = get_bits(&gb, 7);
149 if (p->pitch_lag[1] > 123)
150 return -1;
151 p->pitch_lag[1] += PITCH_MIN;
152 p->subframe[3].ad_cb_lag = get_bits(&gb, 2);
153 p->subframe[0].ad_cb_lag = 1;
154 p->subframe[2].ad_cb_lag = 1;
155
156 for (i = 0; i < SUBFRAMES; i++) {
157 /* Extract combined gain */
158 temp = get_bits(&gb, 12);
159 ad_cb_len = 170;
160 p->subframe[i].dirac_train = 0;
161 if (p->cur_rate == RATE_6300 && p->pitch_lag[i >> 1] < SUBFRAME_LEN - 2) {
162 p->subframe[i].dirac_train = temp >> 11;
163 temp &= 0x7FF;
164 ad_cb_len = 85;
165 }
166 p->subframe[i].ad_cb_gain = FASTDIV(temp, GAIN_LEVELS);
167 if (p->subframe[i].ad_cb_gain < ad_cb_len) {
168 p->subframe[i].amp_index = temp - p->subframe[i].ad_cb_gain *
169 GAIN_LEVELS;
170 } else {
171 return -1;
172 }
173 }
174
175 p->subframe[0].grid_index = get_bits1(&gb);
176 p->subframe[1].grid_index = get_bits1(&gb);
177 p->subframe[2].grid_index = get_bits1(&gb);
178 p->subframe[3].grid_index = get_bits1(&gb);
179
180 if (p->cur_rate == RATE_6300) {
181 skip_bits1(&gb); /* skip reserved bit */
182
183 /* Compute pulse_pos index using the 13-bit combined position index */
184 temp = get_bits(&gb, 13);
185 p->subframe[0].pulse_pos = temp / 810;
186
187 temp -= p->subframe[0].pulse_pos * 810;
188 p->subframe[1].pulse_pos = FASTDIV(temp, 90);
189
190 temp -= p->subframe[1].pulse_pos * 90;
191 p->subframe[2].pulse_pos = FASTDIV(temp, 9);
192 p->subframe[3].pulse_pos = temp - p->subframe[2].pulse_pos * 9;
193
194 p->subframe[0].pulse_pos = (p->subframe[0].pulse_pos << 16) +
195 get_bits(&gb, 16);
196 p->subframe[1].pulse_pos = (p->subframe[1].pulse_pos << 14) +
197 get_bits(&gb, 14);
198 p->subframe[2].pulse_pos = (p->subframe[2].pulse_pos << 16) +
199 get_bits(&gb, 16);
200 p->subframe[3].pulse_pos = (p->subframe[3].pulse_pos << 14) +
201 get_bits(&gb, 14);
202
203 p->subframe[0].pulse_sign = get_bits(&gb, 6);
204 p->subframe[1].pulse_sign = get_bits(&gb, 5);
205 p->subframe[2].pulse_sign = get_bits(&gb, 6);
206 p->subframe[3].pulse_sign = get_bits(&gb, 5);
207 } else { /* 5300 bps */
208 p->subframe[0].pulse_pos = get_bits(&gb, 12);
209 p->subframe[1].pulse_pos = get_bits(&gb, 12);
210 p->subframe[2].pulse_pos = get_bits(&gb, 12);
211 p->subframe[3].pulse_pos = get_bits(&gb, 12);
212
213 p->subframe[0].pulse_sign = get_bits(&gb, 4);
214 p->subframe[1].pulse_sign = get_bits(&gb, 4);
215 p->subframe[2].pulse_sign = get_bits(&gb, 4);
216 p->subframe[3].pulse_sign = get_bits(&gb, 4);
217 }
218
219 return 0;
220}
221
222/**
223 * Bitexact implementation of sqrt(val/2).
224 */
225static int16_t square_root(unsigned val)
226{
227 av_assert2(!(val & 0x80000000));
228
229 return (ff_sqrt(val << 1) >> 1) & (~1);
230}
231
232/**
233 * Calculate the number of left-shifts required for normalizing the input.
234 *
235 * @param num input number
236 * @param width width of the input, 15 or 31 bits
237 */
238static int normalize_bits(int num, int width)
239{
240 return width - av_log2(num) - 1;
241}
242
243#define normalize_bits_int16(num) normalize_bits(num, 15)
244#define normalize_bits_int32(num) normalize_bits(num, 31)
245
246/**
247 * Scale vector contents based on the largest of their absolutes.
248 */
249static int scale_vector(int16_t *dst, const int16_t *vector, int length)
250{
251 int bits, max = 0;
252 int i;
253
254 for (i = 0; i < length; i++)
255 max |= FFABS(vector[i]);
256
257 bits= 14 - av_log2_16bit(max);
258 bits= FFMAX(bits, 0);
259
260 for (i = 0; i < length; i++)
261 dst[i] = vector[i] << bits >> 3;
262
263 return bits - 3;
264}
265
266/**
267 * Perform inverse quantization of LSP frequencies.
268 *
269 * @param cur_lsp the current LSP vector
270 * @param prev_lsp the previous LSP vector
271 * @param lsp_index VQ indices
272 * @param bad_frame bad frame flag
273 */
274static void inverse_quant(int16_t *cur_lsp, int16_t *prev_lsp,
275 uint8_t *lsp_index, int bad_frame)
276{
277 int min_dist, pred;
278 int i, j, temp, stable;
279
280 /* Check for frame erasure */
281 if (!bad_frame) {
282 min_dist = 0x100;
283 pred = 12288;
284 } else {
285 min_dist = 0x200;
286 pred = 23552;
287 lsp_index[0] = lsp_index[1] = lsp_index[2] = 0;
288 }
289
290 /* Get the VQ table entry corresponding to the transmitted index */
291 cur_lsp[0] = lsp_band0[lsp_index[0]][0];
292 cur_lsp[1] = lsp_band0[lsp_index[0]][1];
293 cur_lsp[2] = lsp_band0[lsp_index[0]][2];
294 cur_lsp[3] = lsp_band1[lsp_index[1]][0];
295 cur_lsp[4] = lsp_band1[lsp_index[1]][1];
296 cur_lsp[5] = lsp_band1[lsp_index[1]][2];
297 cur_lsp[6] = lsp_band2[lsp_index[2]][0];
298 cur_lsp[7] = lsp_band2[lsp_index[2]][1];
299 cur_lsp[8] = lsp_band2[lsp_index[2]][2];
300 cur_lsp[9] = lsp_band2[lsp_index[2]][3];
301
302 /* Add predicted vector & DC component to the previously quantized vector */
303 for (i = 0; i < LPC_ORDER; i++) {
304 temp = ((prev_lsp[i] - dc_lsp[i]) * pred + (1 << 14)) >> 15;
305 cur_lsp[i] += dc_lsp[i] + temp;
306 }
307
308 for (i = 0; i < LPC_ORDER; i++) {
309 cur_lsp[0] = FFMAX(cur_lsp[0], 0x180);
310 cur_lsp[LPC_ORDER - 1] = FFMIN(cur_lsp[LPC_ORDER - 1], 0x7e00);
311
312 /* Stability check */
313 for (j = 1; j < LPC_ORDER; j++) {
314 temp = min_dist + cur_lsp[j - 1] - cur_lsp[j];
315 if (temp > 0) {
316 temp >>= 1;
317 cur_lsp[j - 1] -= temp;
318 cur_lsp[j] += temp;
319 }
320 }
321 stable = 1;
322 for (j = 1; j < LPC_ORDER; j++) {
323 temp = cur_lsp[j - 1] + min_dist - cur_lsp[j] - 4;
324 if (temp > 0) {
325 stable = 0;
326 break;
327 }
328 }
329 if (stable)
330 break;
331 }
332 if (!stable)
333 memcpy(cur_lsp, prev_lsp, LPC_ORDER * sizeof(*cur_lsp));
334}
335
336/**
337 * Bitexact implementation of 2ab scaled by 1/2^16.
338 *
339 * @param a 32 bit multiplicand
340 * @param b 16 bit multiplier
341 */
342#define MULL2(a, b) \
343 MULL(a,b,15)
344
345/**
346 * Convert LSP frequencies to LPC coefficients.
347 *
348 * @param lpc buffer for LPC coefficients
349 */
350static void lsp2lpc(int16_t *lpc)
351{
352 int f1[LPC_ORDER / 2 + 1];
353 int f2[LPC_ORDER / 2 + 1];
354 int i, j;
355
356 /* Calculate negative cosine */
357 for (j = 0; j < LPC_ORDER; j++) {
358 int index = (lpc[j] >> 7) & 0x1FF;
359 int offset = lpc[j] & 0x7f;
360 int temp1 = cos_tab[index] << 16;
361 int temp2 = (cos_tab[index + 1] - cos_tab[index]) *
362 ((offset << 8) + 0x80) << 1;
363
364 lpc[j] = -(av_sat_dadd32(1 << 15, temp1 + temp2) >> 16);
365 }
366
367 /*
368 * Compute sum and difference polynomial coefficients
369 * (bitexact alternative to lsp2poly() in lsp.c)
370 */
371 /* Initialize with values in Q28 */
372 f1[0] = 1 << 28;
373 f1[1] = (lpc[0] << 14) + (lpc[2] << 14);
374 f1[2] = lpc[0] * lpc[2] + (2 << 28);
375
376 f2[0] = 1 << 28;
377 f2[1] = (lpc[1] << 14) + (lpc[3] << 14);
378 f2[2] = lpc[1] * lpc[3] + (2 << 28);
379
380 /*
381 * Calculate and scale the coefficients by 1/2 in
382 * each iteration for a final scaling factor of Q25
383 */
384 for (i = 2; i < LPC_ORDER / 2; i++) {
385 f1[i + 1] = f1[i - 1] + MULL2(f1[i], lpc[2 * i]);
386 f2[i + 1] = f2[i - 1] + MULL2(f2[i], lpc[2 * i + 1]);
387
388 for (j = i; j >= 2; j--) {
389 f1[j] = MULL2(f1[j - 1], lpc[2 * i]) +
390 (f1[j] >> 1) + (f1[j - 2] >> 1);
391 f2[j] = MULL2(f2[j - 1], lpc[2 * i + 1]) +
392 (f2[j] >> 1) + (f2[j - 2] >> 1);
393 }
394
395 f1[0] >>= 1;
396 f2[0] >>= 1;
397 f1[1] = ((lpc[2 * i] << 16 >> i) + f1[1]) >> 1;
398 f2[1] = ((lpc[2 * i + 1] << 16 >> i) + f2[1]) >> 1;
399 }
400
401 /* Convert polynomial coefficients to LPC coefficients */
402 for (i = 0; i < LPC_ORDER / 2; i++) {
403 int64_t ff1 = f1[i + 1] + f1[i];
404 int64_t ff2 = f2[i + 1] - f2[i];
405
406 lpc[i] = av_clipl_int32(((ff1 + ff2) << 3) + (1 << 15)) >> 16;
407 lpc[LPC_ORDER - i - 1] = av_clipl_int32(((ff1 - ff2) << 3) +
408 (1 << 15)) >> 16;
409 }
410}
411
412/**
413 * Quantize LSP frequencies by interpolation and convert them to
414 * the corresponding LPC coefficients.
415 *
416 * @param lpc buffer for LPC coefficients
417 * @param cur_lsp the current LSP vector
418 * @param prev_lsp the previous LSP vector
419 */
420static void lsp_interpolate(int16_t *lpc, int16_t *cur_lsp, int16_t *prev_lsp)
421{
422 int i;
423 int16_t *lpc_ptr = lpc;
424
425 /* cur_lsp * 0.25 + prev_lsp * 0.75 */
426 ff_acelp_weighted_vector_sum(lpc, cur_lsp, prev_lsp,
427 4096, 12288, 1 << 13, 14, LPC_ORDER);
428 ff_acelp_weighted_vector_sum(lpc + LPC_ORDER, cur_lsp, prev_lsp,
429 8192, 8192, 1 << 13, 14, LPC_ORDER);
430 ff_acelp_weighted_vector_sum(lpc + 2 * LPC_ORDER, cur_lsp, prev_lsp,
431 12288, 4096, 1 << 13, 14, LPC_ORDER);
432 memcpy(lpc + 3 * LPC_ORDER, cur_lsp, LPC_ORDER * sizeof(*lpc));
433
434 for (i = 0; i < SUBFRAMES; i++) {
435 lsp2lpc(lpc_ptr);
436 lpc_ptr += LPC_ORDER;
437 }
438}
439
440/**
441 * Generate a train of dirac functions with period as pitch lag.
442 */
443static void gen_dirac_train(int16_t *buf, int pitch_lag)
444{
445 int16_t vector[SUBFRAME_LEN];
446 int i, j;
447
448 memcpy(vector, buf, SUBFRAME_LEN * sizeof(*vector));
449 for (i = pitch_lag; i < SUBFRAME_LEN; i += pitch_lag) {
450 for (j = 0; j < SUBFRAME_LEN - i; j++)
451 buf[i + j] += vector[j];
452 }
453}
454
455/**
456 * Generate fixed codebook excitation vector.
457 *
458 * @param vector decoded excitation vector
459 * @param subfrm current subframe
460 * @param cur_rate current bitrate
461 * @param pitch_lag closed loop pitch lag
462 * @param index current subframe index
463 */
464static void gen_fcb_excitation(int16_t *vector, G723_1_Subframe *subfrm,
465 enum Rate cur_rate, int pitch_lag, int index)
466{
467 int temp, i, j;
468
469 memset(vector, 0, SUBFRAME_LEN * sizeof(*vector));
470
471 if (cur_rate == RATE_6300) {
472 if (subfrm->pulse_pos >= max_pos[index])
473 return;
474
475 /* Decode amplitudes and positions */
476 j = PULSE_MAX - pulses[index];
477 temp = subfrm->pulse_pos;
478 for (i = 0; i < SUBFRAME_LEN / GRID_SIZE; i++) {
479 temp -= combinatorial_table[j][i];
480 if (temp >= 0)
481 continue;
482 temp += combinatorial_table[j++][i];
483 if (subfrm->pulse_sign & (1 << (PULSE_MAX - j))) {
484 vector[subfrm->grid_index + GRID_SIZE * i] =
485 -fixed_cb_gain[subfrm->amp_index];
486 } else {
487 vector[subfrm->grid_index + GRID_SIZE * i] =
488 fixed_cb_gain[subfrm->amp_index];
489 }
490 if (j == PULSE_MAX)
491 break;
492 }
493 if (subfrm->dirac_train == 1)
494 gen_dirac_train(vector, pitch_lag);
495 } else { /* 5300 bps */
496 int cb_gain = fixed_cb_gain[subfrm->amp_index];
497 int cb_shift = subfrm->grid_index;
498 int cb_sign = subfrm->pulse_sign;
499 int cb_pos = subfrm->pulse_pos;
500 int offset, beta, lag;
501
502 for (i = 0; i < 8; i += 2) {
503 offset = ((cb_pos & 7) << 3) + cb_shift + i;
504 vector[offset] = (cb_sign & 1) ? cb_gain : -cb_gain;
505 cb_pos >>= 3;
506 cb_sign >>= 1;
507 }
508
509 /* Enhance harmonic components */
510 lag = pitch_contrib[subfrm->ad_cb_gain << 1] + pitch_lag +
511 subfrm->ad_cb_lag - 1;
512 beta = pitch_contrib[(subfrm->ad_cb_gain << 1) + 1];
513
514 if (lag < SUBFRAME_LEN - 2) {
515 for (i = lag; i < SUBFRAME_LEN; i++)
516 vector[i] += beta * vector[i - lag] >> 15;
517 }
518 }
519}
520
521/**
522 * Get delayed contribution from the previous excitation vector.
523 */
524static void get_residual(int16_t *residual, int16_t *prev_excitation, int lag)
525{
526 int offset = PITCH_MAX - PITCH_ORDER / 2 - lag;
527 int i;
528
529 residual[0] = prev_excitation[offset];
530 residual[1] = prev_excitation[offset + 1];
531
532 offset += 2;
533 for (i = 2; i < SUBFRAME_LEN + PITCH_ORDER - 1; i++)
534 residual[i] = prev_excitation[offset + (i - 2) % lag];
535}
536
537static int dot_product(const int16_t *a, const int16_t *b, int length)
538{
539 int sum = ff_dot_product(a,b,length);
540 return av_sat_add32(sum, sum);
541}
542
543/**
544 * Generate adaptive codebook excitation.
545 */
546static void gen_acb_excitation(int16_t *vector, int16_t *prev_excitation,
547 int pitch_lag, G723_1_Subframe *subfrm,
548 enum Rate cur_rate)
549{
550 int16_t residual[SUBFRAME_LEN + PITCH_ORDER - 1];
551 const int16_t *cb_ptr;
552 int lag = pitch_lag + subfrm->ad_cb_lag - 1;
553
554 int i;
555 int sum;
556
557 get_residual(residual, prev_excitation, lag);
558
559 /* Select quantization table */
560 if (cur_rate == RATE_6300 && pitch_lag < SUBFRAME_LEN - 2) {
561 cb_ptr = adaptive_cb_gain85;
562 } else
563 cb_ptr = adaptive_cb_gain170;
564
565 /* Calculate adaptive vector */
566 cb_ptr += subfrm->ad_cb_gain * 20;
567 for (i = 0; i < SUBFRAME_LEN; i++) {
568 sum = ff_dot_product(residual + i, cb_ptr, PITCH_ORDER);
569 vector[i] = av_sat_dadd32(1 << 15, av_sat_add32(sum, sum)) >> 16;
570 }
571}
572
573/**
574 * Estimate maximum auto-correlation around pitch lag.
575 *
576 * @param buf buffer with offset applied
577 * @param offset offset of the excitation vector
578 * @param ccr_max pointer to the maximum auto-correlation
579 * @param pitch_lag decoded pitch lag
580 * @param length length of autocorrelation
581 * @param dir forward lag(1) / backward lag(-1)
582 */
583static int autocorr_max(const int16_t *buf, int offset, int *ccr_max,
584 int pitch_lag, int length, int dir)
585{
586 int limit, ccr, lag = 0;
587 int i;
588
589 pitch_lag = FFMIN(PITCH_MAX - 3, pitch_lag);
590 if (dir > 0)
591 limit = FFMIN(FRAME_LEN + PITCH_MAX - offset - length, pitch_lag + 3);
592 else
593 limit = pitch_lag + 3;
594
595 for (i = pitch_lag - 3; i <= limit; i++) {
596 ccr = dot_product(buf, buf + dir * i, length);
597
598 if (ccr > *ccr_max) {
599 *ccr_max = ccr;
600 lag = i;
601 }
602 }
603 return lag;
604}
605
606/**
607 * Calculate pitch postfilter optimal and scaling gains.
608 *
609 * @param lag pitch postfilter forward/backward lag
610 * @param ppf pitch postfilter parameters
611 * @param cur_rate current bitrate
612 * @param tgt_eng target energy
613 * @param ccr cross-correlation
614 * @param res_eng residual energy
615 */
616static void comp_ppf_gains(int lag, PPFParam *ppf, enum Rate cur_rate,
617 int tgt_eng, int ccr, int res_eng)
618{
619 int pf_residual; /* square of postfiltered residual */
620 int temp1, temp2;
621
622 ppf->index = lag;
623
624 temp1 = tgt_eng * res_eng >> 1;
625 temp2 = ccr * ccr << 1;
626
627 if (temp2 > temp1) {
628 if (ccr >= res_eng) {
629 ppf->opt_gain = ppf_gain_weight[cur_rate];
630 } else {
631 ppf->opt_gain = (ccr << 15) / res_eng *
632 ppf_gain_weight[cur_rate] >> 15;
633 }
634 /* pf_res^2 = tgt_eng + 2*ccr*gain + res_eng*gain^2 */
635 temp1 = (tgt_eng << 15) + (ccr * ppf->opt_gain << 1);
636 temp2 = (ppf->opt_gain * ppf->opt_gain >> 15) * res_eng;
637 pf_residual = av_sat_add32(temp1, temp2 + (1 << 15)) >> 16;
638
639 if (tgt_eng >= pf_residual << 1) {
640 temp1 = 0x7fff;
641 } else {
642 temp1 = (tgt_eng << 14) / pf_residual;
643 }
644
645 /* scaling_gain = sqrt(tgt_eng/pf_res^2) */
646 ppf->sc_gain = square_root(temp1 << 16);
647 } else {
648 ppf->opt_gain = 0;
649 ppf->sc_gain = 0x7fff;
650 }
651
652 ppf->opt_gain = av_clip_int16(ppf->opt_gain * ppf->sc_gain >> 15);
653}
654
655/**
656 * Calculate pitch postfilter parameters.
657 *
658 * @param p the context
659 * @param offset offset of the excitation vector
660 * @param pitch_lag decoded pitch lag
661 * @param ppf pitch postfilter parameters
662 * @param cur_rate current bitrate
663 */
664static void comp_ppf_coeff(G723_1_Context *p, int offset, int pitch_lag,
665 PPFParam *ppf, enum Rate cur_rate)
666{
667
668 int16_t scale;
669 int i;
670 int temp1, temp2;
671
672 /*
673 * 0 - target energy
674 * 1 - forward cross-correlation
675 * 2 - forward residual energy
676 * 3 - backward cross-correlation
677 * 4 - backward residual energy
678 */
679 int energy[5] = {0, 0, 0, 0, 0};
680 int16_t *buf = p->audio + LPC_ORDER + offset;
681 int fwd_lag = autocorr_max(buf, offset, &energy[1], pitch_lag,
682 SUBFRAME_LEN, 1);
683 int back_lag = autocorr_max(buf, offset, &energy[3], pitch_lag,
684 SUBFRAME_LEN, -1);
685
686 ppf->index = 0;
687 ppf->opt_gain = 0;
688 ppf->sc_gain = 0x7fff;
689
690 /* Case 0, Section 3.6 */
691 if (!back_lag && !fwd_lag)
692 return;
693
694 /* Compute target energy */
695 energy[0] = dot_product(buf, buf, SUBFRAME_LEN);
696
697 /* Compute forward residual energy */
698 if (fwd_lag)
699 energy[2] = dot_product(buf + fwd_lag, buf + fwd_lag, SUBFRAME_LEN);
700
701 /* Compute backward residual energy */
702 if (back_lag)
703 energy[4] = dot_product(buf - back_lag, buf - back_lag, SUBFRAME_LEN);
704
705 /* Normalize and shorten */
706 temp1 = 0;
707 for (i = 0; i < 5; i++)
708 temp1 = FFMAX(energy[i], temp1);
709
710 scale = normalize_bits(temp1, 31);
711 for (i = 0; i < 5; i++)
712 energy[i] = (energy[i] << scale) >> 16;
713
714 if (fwd_lag && !back_lag) { /* Case 1 */
715 comp_ppf_gains(fwd_lag, ppf, cur_rate, energy[0], energy[1],
716 energy[2]);
717 } else if (!fwd_lag) { /* Case 2 */
718 comp_ppf_gains(-back_lag, ppf, cur_rate, energy[0], energy[3],
719 energy[4]);
720 } else { /* Case 3 */
721
722 /*
723 * Select the largest of energy[1]^2/energy[2]
724 * and energy[3]^2/energy[4]
725 */
726 temp1 = energy[4] * ((energy[1] * energy[1] + (1 << 14)) >> 15);
727 temp2 = energy[2] * ((energy[3] * energy[3] + (1 << 14)) >> 15);
728 if (temp1 >= temp2) {
729 comp_ppf_gains(fwd_lag, ppf, cur_rate, energy[0], energy[1],
730 energy[2]);
731 } else {
732 comp_ppf_gains(-back_lag, ppf, cur_rate, energy[0], energy[3],
733 energy[4]);
734 }
735 }
736}
737
738/**
739 * Classify frames as voiced/unvoiced.
740 *
741 * @param p the context
742 * @param pitch_lag decoded pitch_lag
743 * @param exc_eng excitation energy estimation
744 * @param scale scaling factor of exc_eng
745 *
746 * @return residual interpolation index if voiced, 0 otherwise
747 */
748static int comp_interp_index(G723_1_Context *p, int pitch_lag,
749 int *exc_eng, int *scale)
750{
751 int offset = PITCH_MAX + 2 * SUBFRAME_LEN;
752 int16_t *buf = p->audio + LPC_ORDER;
753
754 int index, ccr, tgt_eng, best_eng, temp;
755
756 *scale = scale_vector(buf, p->excitation, FRAME_LEN + PITCH_MAX);
757 buf += offset;
758
759 /* Compute maximum backward cross-correlation */
760 ccr = 0;
761 index = autocorr_max(buf, offset, &ccr, pitch_lag, SUBFRAME_LEN * 2, -1);
762 ccr = av_sat_add32(ccr, 1 << 15) >> 16;
763
764 /* Compute target energy */
765 tgt_eng = dot_product(buf, buf, SUBFRAME_LEN * 2);
766 *exc_eng = av_sat_add32(tgt_eng, 1 << 15) >> 16;
767
768 if (ccr <= 0)
769 return 0;
770
771 /* Compute best energy */
772 best_eng = dot_product(buf - index, buf - index, SUBFRAME_LEN * 2);
773 best_eng = av_sat_add32(best_eng, 1 << 15) >> 16;
774
775 temp = best_eng * *exc_eng >> 3;
776
777 if (temp < ccr * ccr) {
778 return index;
779 } else
780 return 0;
781}
782
783/**
784 * Peform residual interpolation based on frame classification.
785 *
786 * @param buf decoded excitation vector
787 * @param out output vector
788 * @param lag decoded pitch lag
789 * @param gain interpolated gain
790 * @param rseed seed for random number generator
791 */
792static void residual_interp(int16_t *buf, int16_t *out, int lag,
793 int gain, int *rseed)
794{
795 int i;
796 if (lag) { /* Voiced */
797 int16_t *vector_ptr = buf + PITCH_MAX;
798 /* Attenuate */
799 for (i = 0; i < lag; i++)
800 out[i] = vector_ptr[i - lag] * 3 >> 2;
801 av_memcpy_backptr((uint8_t*)(out + lag), lag * sizeof(*out),
802 (FRAME_LEN - lag) * sizeof(*out));
803 } else { /* Unvoiced */
804 for (i = 0; i < FRAME_LEN; i++) {
805 *rseed = *rseed * 521 + 259;
806 out[i] = gain * *rseed >> 15;
807 }
808 memset(buf, 0, (FRAME_LEN + PITCH_MAX) * sizeof(*buf));
809 }
810}
811
812/**
813 * Perform IIR filtering.
814 *
815 * @param fir_coef FIR coefficients
816 * @param iir_coef IIR coefficients
817 * @param src source vector
818 * @param dest destination vector
819 * @param width width of the output, 16 bits(0) / 32 bits(1)
820 */
821#define iir_filter(fir_coef, iir_coef, src, dest, width)\
822{\
823 int m, n;\
824 int res_shift = 16 & ~-(width);\
825 int in_shift = 16 - res_shift;\
826\
827 for (m = 0; m < SUBFRAME_LEN; m++) {\
828 int64_t filter = 0;\
829 for (n = 1; n <= LPC_ORDER; n++) {\
830 filter -= (fir_coef)[n - 1] * (src)[m - n] -\
831 (iir_coef)[n - 1] * ((dest)[m - n] >> in_shift);\
832 }\
833\
834 (dest)[m] = av_clipl_int32(((src)[m] << 16) + (filter << 3) +\
835 (1 << 15)) >> res_shift;\
836 }\
837}
838
839/**
840 * Adjust gain of postfiltered signal.
841 *
842 * @param p the context
843 * @param buf postfiltered output vector
844 * @param energy input energy coefficient
845 */
846static void gain_scale(G723_1_Context *p, int16_t * buf, int energy)
847{
848 int num, denom, gain, bits1, bits2;
849 int i;
850
851 num = energy;
852 denom = 0;
853 for (i = 0; i < SUBFRAME_LEN; i++) {
854 int temp = buf[i] >> 2;
855 temp *= temp;
856 denom = av_sat_dadd32(denom, temp);
857 }
858
859 if (num && denom) {
860 bits1 = normalize_bits(num, 31);
861 bits2 = normalize_bits(denom, 31);
862 num = num << bits1 >> 1;
863 denom <<= bits2;
864
865 bits2 = 5 + bits1 - bits2;
866 bits2 = FFMAX(0, bits2);
867
868 gain = (num >> 1) / (denom >> 16);
869 gain = square_root(gain << 16 >> bits2);
870 } else {
871 gain = 1 << 12;
872 }
873
874 for (i = 0; i < SUBFRAME_LEN; i++) {
875 p->pf_gain = (15 * p->pf_gain + gain + (1 << 3)) >> 4;
876 buf[i] = av_clip_int16((buf[i] * (p->pf_gain + (p->pf_gain >> 4)) +
877 (1 << 10)) >> 11);
878 }
879}
880
881/**
882 * Perform formant filtering.
883 *
884 * @param p the context
885 * @param lpc quantized lpc coefficients
886 * @param buf input buffer
887 * @param dst output buffer
888 */
889static void formant_postfilter(G723_1_Context *p, int16_t *lpc,
890 int16_t *buf, int16_t *dst)
891{
892 int16_t filter_coef[2][LPC_ORDER];
893 int filter_signal[LPC_ORDER + FRAME_LEN], *signal_ptr;
894 int i, j, k;
895
896 memcpy(buf, p->fir_mem, LPC_ORDER * sizeof(*buf));
897 memcpy(filter_signal, p->iir_mem, LPC_ORDER * sizeof(*filter_signal));
898
899 for (i = LPC_ORDER, j = 0; j < SUBFRAMES; i += SUBFRAME_LEN, j++) {
900 for (k = 0; k < LPC_ORDER; k++) {
901 filter_coef[0][k] = (-lpc[k] * postfilter_tbl[0][k] +
902 (1 << 14)) >> 15;
903 filter_coef[1][k] = (-lpc[k] * postfilter_tbl[1][k] +
904 (1 << 14)) >> 15;
905 }
906 iir_filter(filter_coef[0], filter_coef[1], buf + i,
907 filter_signal + i, 1);
908 lpc += LPC_ORDER;
909 }
910
911 memcpy(p->fir_mem, buf + FRAME_LEN, LPC_ORDER * sizeof(int16_t));
912 memcpy(p->iir_mem, filter_signal + FRAME_LEN, LPC_ORDER * sizeof(int));
913
914 buf += LPC_ORDER;
915 signal_ptr = filter_signal + LPC_ORDER;
916 for (i = 0; i < SUBFRAMES; i++) {
917 int temp;
918 int auto_corr[2];
919 int scale, energy;
920
921 /* Normalize */
922 scale = scale_vector(dst, buf, SUBFRAME_LEN);
923
924 /* Compute auto correlation coefficients */
925 auto_corr[0] = dot_product(dst, dst + 1, SUBFRAME_LEN - 1);
926 auto_corr[1] = dot_product(dst, dst, SUBFRAME_LEN);
927
928 /* Compute reflection coefficient */
929 temp = auto_corr[1] >> 16;
930 if (temp) {
931 temp = (auto_corr[0] >> 2) / temp;
932 }
933 p->reflection_coef = (3 * p->reflection_coef + temp + 2) >> 2;
934 temp = -p->reflection_coef >> 1 & ~3;
935
936 /* Compensation filter */
937 for (j = 0; j < SUBFRAME_LEN; j++) {
938 dst[j] = av_sat_dadd32(signal_ptr[j],
939 (signal_ptr[j - 1] >> 16) * temp) >> 16;
940 }
941
942 /* Compute normalized signal energy */
943 temp = 2 * scale + 4;
944 if (temp < 0) {
945 energy = av_clipl_int32((int64_t)auto_corr[1] << -temp);
946 } else
947 energy = auto_corr[1] >> temp;
948
949 gain_scale(p, dst, energy);
950
951 buf += SUBFRAME_LEN;
952 signal_ptr += SUBFRAME_LEN;
953 dst += SUBFRAME_LEN;
954 }
955}
956
957static int sid_gain_to_lsp_index(int gain)
958{
959 if (gain < 0x10)
960 return gain << 6;
961 else if (gain < 0x20)
962 return gain - 8 << 7;
963 else
964 return gain - 20 << 8;
965}
966
967static inline int cng_rand(int *state, int base)
968{
969 *state = (*state * 521 + 259) & 0xFFFF;
970 return (*state & 0x7FFF) * base >> 15;
971}
972
973static int estimate_sid_gain(G723_1_Context *p)
974{
975 int i, shift, seg, seg2, t, val, val_add, x, y;
976
977 shift = 16 - p->cur_gain * 2;
978 if (shift > 0)
979 t = p->sid_gain << shift;
980 else
981 t = p->sid_gain >> -shift;
982 x = t * cng_filt[0] >> 16;
983
984 if (x >= cng_bseg[2])
985 return 0x3F;
986
987 if (x >= cng_bseg[1]) {
988 shift = 4;
989 seg = 3;
990 } else {
991 shift = 3;
992 seg = (x >= cng_bseg[0]);
993 }
994 seg2 = FFMIN(seg, 3);
995
996 val = 1 << shift;
997 val_add = val >> 1;
998 for (i = 0; i < shift; i++) {
999 t = seg * 32 + (val << seg2);
1000 t *= t;
1001 if (x >= t)
1002 val += val_add;
1003 else
1004 val -= val_add;
1005 val_add >>= 1;
1006 }
1007
1008 t = seg * 32 + (val << seg2);
1009 y = t * t - x;
1010 if (y <= 0) {
1011 t = seg * 32 + (val + 1 << seg2);
1012 t = t * t - x;
1013 val = (seg2 - 1 << 4) + val;
1014 if (t >= y)
1015 val++;
1016 } else {
1017 t = seg * 32 + (val - 1 << seg2);
1018 t = t * t - x;
1019 val = (seg2 - 1 << 4) + val;
1020 if (t >= y)
1021 val--;
1022 }
1023
1024 return val;
1025}
1026
1027static void generate_noise(G723_1_Context *p)
1028{
1029 int i, j, idx, t;
1030 int off[SUBFRAMES];
1031 int signs[SUBFRAMES / 2 * 11], pos[SUBFRAMES / 2 * 11];
1032 int tmp[SUBFRAME_LEN * 2];
1033 int16_t *vector_ptr;
1034 int64_t sum;
1035 int b0, c, delta, x, shift;
1036
1037 p->pitch_lag[0] = cng_rand(&p->cng_random_seed, 21) + 123;
1038 p->pitch_lag[1] = cng_rand(&p->cng_random_seed, 19) + 123;
1039
1040 for (i = 0; i < SUBFRAMES; i++) {
1041 p->subframe[i].ad_cb_gain = cng_rand(&p->cng_random_seed, 50) + 1;
1042 p->subframe[i].ad_cb_lag = cng_adaptive_cb_lag[i];
1043 }
1044
1045 for (i = 0; i < SUBFRAMES / 2; i++) {
1046 t = cng_rand(&p->cng_random_seed, 1 << 13);
1047 off[i * 2] = t & 1;
1048 off[i * 2 + 1] = ((t >> 1) & 1) + SUBFRAME_LEN;
1049 t >>= 2;
1050 for (j = 0; j < 11; j++) {
1051 signs[i * 11 + j] = (t & 1) * 2 - 1 << 14;
1052 t >>= 1;
1053 }
1054 }
1055
1056 idx = 0;
1057 for (i = 0; i < SUBFRAMES; i++) {
1058 for (j = 0; j < SUBFRAME_LEN / 2; j++)
1059 tmp[j] = j;
1060 t = SUBFRAME_LEN / 2;
1061 for (j = 0; j < pulses[i]; j++, idx++) {
1062 int idx2 = cng_rand(&p->cng_random_seed, t);
1063
1064 pos[idx] = tmp[idx2] * 2 + off[i];
1065 tmp[idx2] = tmp[--t];
1066 }
1067 }
1068
1069 vector_ptr = p->audio + LPC_ORDER;
1070 memcpy(vector_ptr, p->prev_excitation,
1071 PITCH_MAX * sizeof(*p->excitation));
1072 for (i = 0; i < SUBFRAMES; i += 2) {
1073 gen_acb_excitation(vector_ptr, vector_ptr,
1074 p->pitch_lag[i >> 1], &p->subframe[i],
1075 p->cur_rate);
1076 gen_acb_excitation(vector_ptr + SUBFRAME_LEN,
1077 vector_ptr + SUBFRAME_LEN,
1078 p->pitch_lag[i >> 1], &p->subframe[i + 1],
1079 p->cur_rate);
1080
1081 t = 0;
1082 for (j = 0; j < SUBFRAME_LEN * 2; j++)
1083 t |= FFABS(vector_ptr[j]);
1084 t = FFMIN(t, 0x7FFF);
1085 if (!t) {
1086 shift = 0;
1087 } else {
1088 shift = -10 + av_log2(t);
1089 if (shift < -2)
1090 shift = -2;
1091 }
1092 sum = 0;
1093 if (shift < 0) {
1094 for (j = 0; j < SUBFRAME_LEN * 2; j++) {
1095 t = vector_ptr[j] << -shift;
1096 sum += t * t;
1097 tmp[j] = t;
1098 }
1099 } else {
1100 for (j = 0; j < SUBFRAME_LEN * 2; j++) {
1101 t = vector_ptr[j] >> shift;
1102 sum += t * t;
1103 tmp[j] = t;
1104 }
1105 }
1106
1107 b0 = 0;
1108 for (j = 0; j < 11; j++)
1109 b0 += tmp[pos[(i / 2) * 11 + j]] * signs[(i / 2) * 11 + j];
1110 b0 = b0 * 2 * 2979LL + (1 << 29) >> 30; // approximated division by 11
1111
1112 c = p->cur_gain * (p->cur_gain * SUBFRAME_LEN >> 5);
1113 if (shift * 2 + 3 >= 0)
1114 c >>= shift * 2 + 3;
1115 else
1116 c <<= -(shift * 2 + 3);
1117 c = (av_clipl_int32(sum << 1) - c) * 2979LL >> 15;
1118
1119 delta = b0 * b0 * 2 - c;
1120 if (delta <= 0) {
1121 x = -b0;
1122 } else {
1123 delta = square_root(delta);
1124 x = delta - b0;
1125 t = delta + b0;
1126 if (FFABS(t) < FFABS(x))
1127 x = -t;
1128 }
1129 shift++;
1130 if (shift < 0)
1131 x >>= -shift;
1132 else
1133 x <<= shift;
1134 x = av_clip(x, -10000, 10000);
1135
1136 for (j = 0; j < 11; j++) {
1137 idx = (i / 2) * 11 + j;
1138 vector_ptr[pos[idx]] = av_clip_int16(vector_ptr[pos[idx]] +
1139 (x * signs[idx] >> 15));
1140 }
1141
1142 /* copy decoded data to serve as a history for the next decoded subframes */
1143 memcpy(vector_ptr + PITCH_MAX, vector_ptr,
1144 sizeof(*vector_ptr) * SUBFRAME_LEN * 2);
1145 vector_ptr += SUBFRAME_LEN * 2;
1146 }
1147 /* Save the excitation for the next frame */
1148 memcpy(p->prev_excitation, p->audio + LPC_ORDER + FRAME_LEN,
1149 PITCH_MAX * sizeof(*p->excitation));
1150}
1151
1152static int g723_1_decode_frame(AVCodecContext *avctx, void *data,
1153 int *got_frame_ptr, AVPacket *avpkt)
1154{
1155 G723_1_Context *p = avctx->priv_data;
1156 AVFrame *frame = data;
1157 const uint8_t *buf = avpkt->data;
1158 int buf_size = avpkt->size;
1159 int dec_mode = buf[0] & 3;
1160
1161 PPFParam ppf[SUBFRAMES];
1162 int16_t cur_lsp[LPC_ORDER];
1163 int16_t lpc[SUBFRAMES * LPC_ORDER];
1164 int16_t acb_vector[SUBFRAME_LEN];
1165 int16_t *out;
1166 int bad_frame = 0, i, j, ret;
1167 int16_t *audio = p->audio;
1168
1169 if (buf_size < frame_size[dec_mode]) {
1170 if (buf_size)
1171 av_log(avctx, AV_LOG_WARNING,
1172 "Expected %d bytes, got %d - skipping packet\n",
1173 frame_size[dec_mode], buf_size);
1174 *got_frame_ptr = 0;
1175 return buf_size;
1176 }
1177
1178 if (unpack_bitstream(p, buf, buf_size) < 0) {
1179 bad_frame = 1;
1180 if (p->past_frame_type == ACTIVE_FRAME)
1181 p->cur_frame_type = ACTIVE_FRAME;
1182 else
1183 p->cur_frame_type = UNTRANSMITTED_FRAME;
1184 }
1185
1186 frame->nb_samples = FRAME_LEN;
1187 if ((ret = ff_get_buffer(avctx, frame, 0)) < 0)
1188 return ret;
1189
1190 out = (int16_t *)frame->data[0];
1191
1192 if (p->cur_frame_type == ACTIVE_FRAME) {
1193 if (!bad_frame)
1194 p->erased_frames = 0;
1195 else if (p->erased_frames != 3)
1196 p->erased_frames++;
1197
1198 inverse_quant(cur_lsp, p->prev_lsp, p->lsp_index, bad_frame);
1199 lsp_interpolate(lpc, cur_lsp, p->prev_lsp);
1200
1201 /* Save the lsp_vector for the next frame */
1202 memcpy(p->prev_lsp, cur_lsp, LPC_ORDER * sizeof(*p->prev_lsp));
1203
1204 /* Generate the excitation for the frame */
1205 memcpy(p->excitation, p->prev_excitation,
1206 PITCH_MAX * sizeof(*p->excitation));
1207 if (!p->erased_frames) {
1208 int16_t *vector_ptr = p->excitation + PITCH_MAX;
1209
1210 /* Update interpolation gain memory */
1211 p->interp_gain = fixed_cb_gain[(p->subframe[2].amp_index +
1212 p->subframe[3].amp_index) >> 1];
1213 for (i = 0; i < SUBFRAMES; i++) {
1214 gen_fcb_excitation(vector_ptr, &p->subframe[i], p->cur_rate,
1215 p->pitch_lag[i >> 1], i);
1216 gen_acb_excitation(acb_vector, &p->excitation[SUBFRAME_LEN * i],
1217 p->pitch_lag[i >> 1], &p->subframe[i],
1218 p->cur_rate);
1219 /* Get the total excitation */
1220 for (j = 0; j < SUBFRAME_LEN; j++) {
1221 int v = av_clip_int16(vector_ptr[j] << 1);
1222 vector_ptr[j] = av_clip_int16(v + acb_vector[j]);
1223 }
1224 vector_ptr += SUBFRAME_LEN;
1225 }
1226
1227 vector_ptr = p->excitation + PITCH_MAX;
1228
1229 p->interp_index = comp_interp_index(p, p->pitch_lag[1],
1230 &p->sid_gain, &p->cur_gain);
1231
1232 /* Peform pitch postfiltering */
1233 if (p->postfilter) {
1234 i = PITCH_MAX;
1235 for (j = 0; j < SUBFRAMES; i += SUBFRAME_LEN, j++)
1236 comp_ppf_coeff(p, i, p->pitch_lag[j >> 1],
1237 ppf + j, p->cur_rate);
1238
1239 for (i = 0, j = 0; j < SUBFRAMES; i += SUBFRAME_LEN, j++)
1240 ff_acelp_weighted_vector_sum(p->audio + LPC_ORDER + i,
1241 vector_ptr + i,
1242 vector_ptr + i + ppf[j].index,
1243 ppf[j].sc_gain,
1244 ppf[j].opt_gain,
1245 1 << 14, 15, SUBFRAME_LEN);
1246 } else {
1247 audio = vector_ptr - LPC_ORDER;
1248 }
1249
1250 /* Save the excitation for the next frame */
1251 memcpy(p->prev_excitation, p->excitation + FRAME_LEN,
1252 PITCH_MAX * sizeof(*p->excitation));
1253 } else {
1254 p->interp_gain = (p->interp_gain * 3 + 2) >> 2;
1255 if (p->erased_frames == 3) {
1256 /* Mute output */
1257 memset(p->excitation, 0,
1258 (FRAME_LEN + PITCH_MAX) * sizeof(*p->excitation));
1259 memset(p->prev_excitation, 0,
1260 PITCH_MAX * sizeof(*p->excitation));
1261 memset(frame->data[0], 0,
1262 (FRAME_LEN + LPC_ORDER) * sizeof(int16_t));
1263 } else {
1264 int16_t *buf = p->audio + LPC_ORDER;
1265
1266 /* Regenerate frame */
1267 residual_interp(p->excitation, buf, p->interp_index,
1268 p->interp_gain, &p->random_seed);
1269
1270 /* Save the excitation for the next frame */
1271 memcpy(p->prev_excitation, buf + (FRAME_LEN - PITCH_MAX),
1272 PITCH_MAX * sizeof(*p->excitation));
1273 }
1274 }
1275 p->cng_random_seed = CNG_RANDOM_SEED;
1276 } else {
1277 if (p->cur_frame_type == SID_FRAME) {
1278 p->sid_gain = sid_gain_to_lsp_index(p->subframe[0].amp_index);
1279 inverse_quant(p->sid_lsp, p->prev_lsp, p->lsp_index, 0);
1280 } else if (p->past_frame_type == ACTIVE_FRAME) {
1281 p->sid_gain = estimate_sid_gain(p);
1282 }
1283
1284 if (p->past_frame_type == ACTIVE_FRAME)
1285 p->cur_gain = p->sid_gain;
1286 else
1287 p->cur_gain = (p->cur_gain * 7 + p->sid_gain) >> 3;
1288 generate_noise(p);
1289 lsp_interpolate(lpc, p->sid_lsp, p->prev_lsp);
1290 /* Save the lsp_vector for the next frame */
1291 memcpy(p->prev_lsp, p->sid_lsp, LPC_ORDER * sizeof(*p->prev_lsp));
1292 }
1293
1294 p->past_frame_type = p->cur_frame_type;
1295
1296 memcpy(p->audio, p->synth_mem, LPC_ORDER * sizeof(*p->audio));
1297 for (i = LPC_ORDER, j = 0; j < SUBFRAMES; i += SUBFRAME_LEN, j++)
1298 ff_celp_lp_synthesis_filter(p->audio + i, &lpc[j * LPC_ORDER],
1299 audio + i, SUBFRAME_LEN, LPC_ORDER,
1300 0, 1, 1 << 12);
1301 memcpy(p->synth_mem, p->audio + FRAME_LEN, LPC_ORDER * sizeof(*p->audio));
1302
1303 if (p->postfilter) {
1304 formant_postfilter(p, lpc, p->audio, out);
1305 } else { // if output is not postfiltered it should be scaled by 2
1306 for (i = 0; i < FRAME_LEN; i++)
1307 out[i] = av_clip_int16(p->audio[LPC_ORDER + i] << 1);
1308 }
1309
1310 *got_frame_ptr = 1;
1311
1312 return frame_size[dec_mode];
1313}
1314
1315#define OFFSET(x) offsetof(G723_1_Context, x)
1316#define AD AV_OPT_FLAG_AUDIO_PARAM | AV_OPT_FLAG_DECODING_PARAM
1317
1318static const AVOption options[] = {
1319 { "postfilter", "postfilter on/off", OFFSET(postfilter), AV_OPT_TYPE_INT,
1320 { .i64 = 1 }, 0, 1, AD },
1321 { NULL }
1322};
1323
1324
1325static const AVClass g723_1dec_class = {
1326 .class_name = "G.723.1 decoder",
1327 .item_name = av_default_item_name,
1328 .option = options,
1329 .version = LIBAVUTIL_VERSION_INT,
1330};
1331
1332AVCodec ff_g723_1_decoder = {
1333 .name = "g723_1",
1334 .long_name = NULL_IF_CONFIG_SMALL("G.723.1"),
1335 .type = AVMEDIA_TYPE_AUDIO,
1336 .id = AV_CODEC_ID_G723_1,
1337 .priv_data_size = sizeof(G723_1_Context),
1338 .init = g723_1_decode_init,
1339 .decode = g723_1_decode_frame,
1340 .capabilities = CODEC_CAP_SUBFRAMES | CODEC_CAP_DR1,
1341 .priv_class = &g723_1dec_class,
1342};
1343
1344#if CONFIG_G723_1_ENCODER
1345#define BITSTREAM_WRITER_LE
1346#include "put_bits.h"
1347
1348static av_cold int g723_1_encode_init(AVCodecContext *avctx)
1349{
1350 G723_1_Context *p = avctx->priv_data;
1351
1352 if (avctx->sample_rate != 8000) {
1353 av_log(avctx, AV_LOG_ERROR, "Only 8000Hz sample rate supported\n");
1354 return -1;
1355 }
1356
1357 if (avctx->channels != 1) {
1358 av_log(avctx, AV_LOG_ERROR, "Only mono supported\n");
1359 return AVERROR(EINVAL);
1360 }
1361
1362 if (avctx->bit_rate == 6300) {
1363 p->cur_rate = RATE_6300;
1364 } else if (avctx->bit_rate == 5300) {
1365 av_log(avctx, AV_LOG_ERROR, "Bitrate not supported yet, use 6.3k\n");
1366 return AVERROR_PATCHWELCOME;
1367 } else {
1368 av_log(avctx, AV_LOG_ERROR,
1369 "Bitrate not supported, use 6.3k\n");
1370 return AVERROR(EINVAL);
1371 }
1372 avctx->frame_size = 240;
1373 memcpy(p->prev_lsp, dc_lsp, LPC_ORDER * sizeof(int16_t));
1374
1375 return 0;
1376}
1377
1378/**
1379 * Remove DC component from the input signal.
1380 *
1381 * @param buf input signal
1382 * @param fir zero memory
1383 * @param iir pole memory
1384 */
1385static void highpass_filter(int16_t *buf, int16_t *fir, int *iir)
1386{
1387 int i;
1388 for (i = 0; i < FRAME_LEN; i++) {
1389 *iir = (buf[i] << 15) + ((-*fir) << 15) + MULL2(*iir, 0x7f00);
1390 *fir = buf[i];
1391 buf[i] = av_clipl_int32((int64_t)*iir + (1 << 15)) >> 16;
1392 }
1393}
1394
1395/**
1396 * Estimate autocorrelation of the input vector.
1397 *
1398 * @param buf input buffer
1399 * @param autocorr autocorrelation coefficients vector
1400 */
1401static void comp_autocorr(int16_t *buf, int16_t *autocorr)
1402{
1403 int i, scale, temp;
1404 int16_t vector[LPC_FRAME];
1405
1406 scale_vector(vector, buf, LPC_FRAME);
1407
1408 /* Apply the Hamming window */
1409 for (i = 0; i < LPC_FRAME; i++)
1410 vector[i] = (vector[i] * hamming_window[i] + (1 << 14)) >> 15;
1411
1412 /* Compute the first autocorrelation coefficient */
1413 temp = ff_dot_product(vector, vector, LPC_FRAME);
1414
1415 /* Apply a white noise correlation factor of (1025/1024) */
1416 temp += temp >> 10;
1417
1418 /* Normalize */
1419 scale = normalize_bits_int32(temp);
1420 autocorr[0] = av_clipl_int32((int64_t)(temp << scale) +
1421 (1 << 15)) >> 16;
1422
1423 /* Compute the remaining coefficients */
1424 if (!autocorr[0]) {
1425 memset(autocorr + 1, 0, LPC_ORDER * sizeof(int16_t));
1426 } else {
1427 for (i = 1; i <= LPC_ORDER; i++) {
1428 temp = ff_dot_product(vector, vector + i, LPC_FRAME - i);
1429 temp = MULL2((temp << scale), binomial_window[i - 1]);
1430 autocorr[i] = av_clipl_int32((int64_t)temp + (1 << 15)) >> 16;
1431 }
1432 }
1433}
1434
1435/**
1436 * Use Levinson-Durbin recursion to compute LPC coefficients from
1437 * autocorrelation values.
1438 *
1439 * @param lpc LPC coefficients vector
1440 * @param autocorr autocorrelation coefficients vector
1441 * @param error prediction error
1442 */
1443static void levinson_durbin(int16_t *lpc, int16_t *autocorr, int16_t error)
1444{
1445 int16_t vector[LPC_ORDER];
1446 int16_t partial_corr;
1447 int i, j, temp;
1448
1449 memset(lpc, 0, LPC_ORDER * sizeof(int16_t));
1450
1451 for (i = 0; i < LPC_ORDER; i++) {
1452 /* Compute the partial correlation coefficient */
1453 temp = 0;
1454 for (j = 0; j < i; j++)
1455 temp -= lpc[j] * autocorr[i - j - 1];
1456 temp = ((autocorr[i] << 13) + temp) << 3;
1457
1458 if (FFABS(temp) >= (error << 16))
1459 break;
1460
1461 partial_corr = temp / (error << 1);
1462
1463 lpc[i] = av_clipl_int32((int64_t)(partial_corr << 14) +
1464 (1 << 15)) >> 16;
1465
1466 /* Update the prediction error */
1467 temp = MULL2(temp, partial_corr);
1468 error = av_clipl_int32((int64_t)(error << 16) - temp +
1469 (1 << 15)) >> 16;
1470
1471 memcpy(vector, lpc, i * sizeof(int16_t));
1472 for (j = 0; j < i; j++) {
1473 temp = partial_corr * vector[i - j - 1] << 1;
1474 lpc[j] = av_clipl_int32((int64_t)(lpc[j] << 16) - temp +
1475 (1 << 15)) >> 16;
1476 }
1477 }
1478}
1479
1480/**
1481 * Calculate LPC coefficients for the current frame.
1482 *
1483 * @param buf current frame
1484 * @param prev_data 2 trailing subframes of the previous frame
1485 * @param lpc LPC coefficients vector
1486 */
1487static void comp_lpc_coeff(int16_t *buf, int16_t *lpc)
1488{
1489 int16_t autocorr[(LPC_ORDER + 1) * SUBFRAMES];
1490 int16_t *autocorr_ptr = autocorr;
1491 int16_t *lpc_ptr = lpc;
1492 int i, j;
1493
1494 for (i = 0, j = 0; j < SUBFRAMES; i += SUBFRAME_LEN, j++) {
1495 comp_autocorr(buf + i, autocorr_ptr);
1496 levinson_durbin(lpc_ptr, autocorr_ptr + 1, autocorr_ptr[0]);
1497
1498 lpc_ptr += LPC_ORDER;
1499 autocorr_ptr += LPC_ORDER + 1;
1500 }
1501}
1502
1503static void lpc2lsp(int16_t *lpc, int16_t *prev_lsp, int16_t *lsp)
1504{
1505 int f[LPC_ORDER + 2]; ///< coefficients of the sum and difference
1506 ///< polynomials (F1, F2) ordered as
1507 ///< f1[0], f2[0], ...., f1[5], f2[5]
1508
1509 int max, shift, cur_val, prev_val, count, p;
1510 int i, j;
1511 int64_t temp;
1512
1513 /* Initialize f1[0] and f2[0] to 1 in Q25 */
1514 for (i = 0; i < LPC_ORDER; i++)
1515 lsp[i] = (lpc[i] * bandwidth_expand[i] + (1 << 14)) >> 15;
1516
1517 /* Apply bandwidth expansion on the LPC coefficients */
1518 f[0] = f[1] = 1 << 25;
1519
1520 /* Compute the remaining coefficients */
1521 for (i = 0; i < LPC_ORDER / 2; i++) {
1522 /* f1 */
1523 f[2 * i + 2] = -f[2 * i] - ((lsp[i] + lsp[LPC_ORDER - 1 - i]) << 12);
1524 /* f2 */
1525 f[2 * i + 3] = f[2 * i + 1] - ((lsp[i] - lsp[LPC_ORDER - 1 - i]) << 12);
1526 }
1527
1528 /* Divide f1[5] and f2[5] by 2 for use in polynomial evaluation */
1529 f[LPC_ORDER] >>= 1;
1530 f[LPC_ORDER + 1] >>= 1;
1531
1532 /* Normalize and shorten */
1533 max = FFABS(f[0]);
1534 for (i = 1; i < LPC_ORDER + 2; i++)
1535 max = FFMAX(max, FFABS(f[i]));
1536
1537 shift = normalize_bits_int32(max);
1538
1539 for (i = 0; i < LPC_ORDER + 2; i++)
1540 f[i] = av_clipl_int32((int64_t)(f[i] << shift) + (1 << 15)) >> 16;
1541
1542 /**
1543 * Evaluate F1 and F2 at uniform intervals of pi/256 along the
1544 * unit circle and check for zero crossings.
1545 */
1546 p = 0;
1547 temp = 0;
1548 for (i = 0; i <= LPC_ORDER / 2; i++)
1549 temp += f[2 * i] * cos_tab[0];
1550 prev_val = av_clipl_int32(temp << 1);
1551 count = 0;
1552 for ( i = 1; i < COS_TBL_SIZE / 2; i++) {
1553 /* Evaluate */
1554 temp = 0;
1555 for (j = 0; j <= LPC_ORDER / 2; j++)
1556 temp += f[LPC_ORDER - 2 * j + p] * cos_tab[i * j % COS_TBL_SIZE];
1557 cur_val = av_clipl_int32(temp << 1);
1558
1559 /* Check for sign change, indicating a zero crossing */
1560 if ((cur_val ^ prev_val) < 0) {
1561 int abs_cur = FFABS(cur_val);
1562 int abs_prev = FFABS(prev_val);
1563 int sum = abs_cur + abs_prev;
1564
1565 shift = normalize_bits_int32(sum);
1566 sum <<= shift;
1567 abs_prev = abs_prev << shift >> 8;
1568 lsp[count++] = ((i - 1) << 7) + (abs_prev >> 1) / (sum >> 16);
1569
1570 if (count == LPC_ORDER)
1571 break;
1572
1573 /* Switch between sum and difference polynomials */
1574 p ^= 1;
1575
1576 /* Evaluate */
1577 temp = 0;
1578 for (j = 0; j <= LPC_ORDER / 2; j++){
1579 temp += f[LPC_ORDER - 2 * j + p] *
1580 cos_tab[i * j % COS_TBL_SIZE];
1581 }
1582 cur_val = av_clipl_int32(temp<<1);
1583 }
1584 prev_val = cur_val;
1585 }
1586
1587 if (count != LPC_ORDER)
1588 memcpy(lsp, prev_lsp, LPC_ORDER * sizeof(int16_t));
1589}
1590
1591/**
1592 * Quantize the current LSP subvector.
1593 *
1594 * @param num band number
1595 * @param offset offset of the current subvector in an LPC_ORDER vector
1596 * @param size size of the current subvector
1597 */
1598#define get_index(num, offset, size) \
1599{\
1600 int error, max = -1;\
1601 int16_t temp[4];\
1602 int i, j;\
1603 for (i = 0; i < LSP_CB_SIZE; i++) {\
1604 for (j = 0; j < size; j++){\
1605 temp[j] = (weight[j + (offset)] * lsp_band##num[i][j] +\
1606 (1 << 14)) >> 15;\
1607 }\
1608 error = dot_product(lsp + (offset), temp, size) << 1;\
1609 error -= dot_product(lsp_band##num[i], temp, size);\
1610 if (error > max) {\
1611 max = error;\
1612 lsp_index[num] = i;\
1613 }\
1614 }\
1615}
1616
1617/**
1618 * Vector quantize the LSP frequencies.
1619 *
1620 * @param lsp the current lsp vector
1621 * @param prev_lsp the previous lsp vector
1622 */
1623static void lsp_quantize(uint8_t *lsp_index, int16_t *lsp, int16_t *prev_lsp)
1624{
1625 int16_t weight[LPC_ORDER];
1626 int16_t min, max;
1627 int shift, i;
1628
1629 /* Calculate the VQ weighting vector */
1630 weight[0] = (1 << 20) / (lsp[1] - lsp[0]);
1631 weight[LPC_ORDER - 1] = (1 << 20) /
1632 (lsp[LPC_ORDER - 1] - lsp[LPC_ORDER - 2]);
1633
1634 for (i = 1; i < LPC_ORDER - 1; i++) {
1635 min = FFMIN(lsp[i] - lsp[i - 1], lsp[i + 1] - lsp[i]);
1636 if (min > 0x20)
1637 weight[i] = (1 << 20) / min;
1638 else
1639 weight[i] = INT16_MAX;
1640 }
1641
1642 /* Normalize */
1643 max = 0;
1644 for (i = 0; i < LPC_ORDER; i++)
1645 max = FFMAX(weight[i], max);
1646
1647 shift = normalize_bits_int16(max);
1648 for (i = 0; i < LPC_ORDER; i++) {
1649 weight[i] <<= shift;
1650 }
1651
1652 /* Compute the VQ target vector */
1653 for (i = 0; i < LPC_ORDER; i++) {
1654 lsp[i] -= dc_lsp[i] +
1655 (((prev_lsp[i] - dc_lsp[i]) * 12288 + (1 << 14)) >> 15);
1656 }
1657
1658 get_index(0, 0, 3);
1659 get_index(1, 3, 3);
1660 get_index(2, 6, 4);
1661}
1662
1663/**
1664 * Apply the formant perceptual weighting filter.
1665 *
1666 * @param flt_coef filter coefficients
1667 * @param unq_lpc unquantized lpc vector
1668 */
1669static void perceptual_filter(G723_1_Context *p, int16_t *flt_coef,
1670 int16_t *unq_lpc, int16_t *buf)
1671{
1672 int16_t vector[FRAME_LEN + LPC_ORDER];
1673 int i, j, k, l = 0;
1674
1675 memcpy(buf, p->iir_mem, sizeof(int16_t) * LPC_ORDER);
1676 memcpy(vector, p->fir_mem, sizeof(int16_t) * LPC_ORDER);
1677 memcpy(vector + LPC_ORDER, buf + LPC_ORDER, sizeof(int16_t) * FRAME_LEN);
1678
1679 for (i = LPC_ORDER, j = 0; j < SUBFRAMES; i += SUBFRAME_LEN, j++) {
1680 for (k = 0; k < LPC_ORDER; k++) {
1681 flt_coef[k + 2 * l] = (unq_lpc[k + l] * percept_flt_tbl[0][k] +
1682 (1 << 14)) >> 15;
1683 flt_coef[k + 2 * l + LPC_ORDER] = (unq_lpc[k + l] *
1684 percept_flt_tbl[1][k] +
1685 (1 << 14)) >> 15;
1686 }
1687 iir_filter(flt_coef + 2 * l, flt_coef + 2 * l + LPC_ORDER, vector + i,
1688 buf + i, 0);
1689 l += LPC_ORDER;
1690 }
1691 memcpy(p->iir_mem, buf + FRAME_LEN, sizeof(int16_t) * LPC_ORDER);
1692 memcpy(p->fir_mem, vector + FRAME_LEN, sizeof(int16_t) * LPC_ORDER);
1693}
1694
1695/**
1696 * Estimate the open loop pitch period.
1697 *
1698 * @param buf perceptually weighted speech
1699 * @param start estimation is carried out from this position
1700 */
1701static int estimate_pitch(int16_t *buf, int start)
1702{
1703 int max_exp = 32;
1704 int max_ccr = 0x4000;
1705 int max_eng = 0x7fff;
1706 int index = PITCH_MIN;
1707 int offset = start - PITCH_MIN + 1;
1708
1709 int ccr, eng, orig_eng, ccr_eng, exp;
1710 int diff, temp;
1711
1712 int i;
1713
1714 orig_eng = ff_dot_product(buf + offset, buf + offset, HALF_FRAME_LEN);
1715
1716 for (i = PITCH_MIN; i <= PITCH_MAX - 3; i++) {
1717 offset--;
1718
1719 /* Update energy and compute correlation */
1720 orig_eng += buf[offset] * buf[offset] -
1721 buf[offset + HALF_FRAME_LEN] * buf[offset + HALF_FRAME_LEN];
1722 ccr = ff_dot_product(buf + start, buf + offset, HALF_FRAME_LEN);
1723 if (ccr <= 0)
1724 continue;
1725
1726 /* Split into mantissa and exponent to maintain precision */
1727 exp = normalize_bits_int32(ccr);
1728 ccr = av_clipl_int32((int64_t)(ccr << exp) + (1 << 15)) >> 16;
1729 exp <<= 1;
1730 ccr *= ccr;
1731 temp = normalize_bits_int32(ccr);
1732 ccr = ccr << temp >> 16;
1733 exp += temp;
1734
1735 temp = normalize_bits_int32(orig_eng);
1736 eng = av_clipl_int32((int64_t)(orig_eng << temp) + (1 << 15)) >> 16;
1737 exp -= temp;
1738
1739 if (ccr >= eng) {
1740 exp--;
1741 ccr >>= 1;
1742 }
1743 if (exp > max_exp)
1744 continue;
1745
1746 if (exp + 1 < max_exp)
1747 goto update;
1748
1749 /* Equalize exponents before comparison */
1750 if (exp + 1 == max_exp)
1751 temp = max_ccr >> 1;
1752 else
1753 temp = max_ccr;
1754 ccr_eng = ccr * max_eng;
1755 diff = ccr_eng - eng * temp;
1756 if (diff > 0 && (i - index < PITCH_MIN || diff > ccr_eng >> 2)) {
1757update:
1758 index = i;
1759 max_exp = exp;
1760 max_ccr = ccr;
1761 max_eng = eng;
1762 }
1763 }
1764 return index;
1765}
1766
1767/**
1768 * Compute harmonic noise filter parameters.
1769 *
1770 * @param buf perceptually weighted speech
1771 * @param pitch_lag open loop pitch period
1772 * @param hf harmonic filter parameters
1773 */
1774static void comp_harmonic_coeff(int16_t *buf, int16_t pitch_lag, HFParam *hf)
1775{
1776 int ccr, eng, max_ccr, max_eng;
1777 int exp, max, diff;
1778 int energy[15];
1779 int i, j;
1780
1781 for (i = 0, j = pitch_lag - 3; j <= pitch_lag + 3; i++, j++) {
1782 /* Compute residual energy */
1783 energy[i << 1] = ff_dot_product(buf - j, buf - j, SUBFRAME_LEN);
1784 /* Compute correlation */
1785 energy[(i << 1) + 1] = ff_dot_product(buf, buf - j, SUBFRAME_LEN);
1786 }
1787
1788 /* Compute target energy */
1789 energy[14] = ff_dot_product(buf, buf, SUBFRAME_LEN);
1790
1791 /* Normalize */
1792 max = 0;
1793 for (i = 0; i < 15; i++)
1794 max = FFMAX(max, FFABS(energy[i]));
1795
1796 exp = normalize_bits_int32(max);
1797 for (i = 0; i < 15; i++) {
1798 energy[i] = av_clipl_int32((int64_t)(energy[i] << exp) +
1799 (1 << 15)) >> 16;
1800 }
1801
1802 hf->index = -1;
1803 hf->gain = 0;
1804 max_ccr = 1;
1805 max_eng = 0x7fff;
1806
1807 for (i = 0; i <= 6; i++) {
1808 eng = energy[i << 1];
1809 ccr = energy[(i << 1) + 1];
1810
1811 if (ccr <= 0)
1812 continue;
1813
1814 ccr = (ccr * ccr + (1 << 14)) >> 15;
1815 diff = ccr * max_eng - eng * max_ccr;
1816 if (diff > 0) {
1817 max_ccr = ccr;
1818 max_eng = eng;
1819 hf->index = i;
1820 }
1821 }
1822
1823 if (hf->index == -1) {
1824 hf->index = pitch_lag;
1825 return;
1826 }
1827
1828 eng = energy[14] * max_eng;
1829 eng = (eng >> 2) + (eng >> 3);
1830 ccr = energy[(hf->index << 1) + 1] * energy[(hf->index << 1) + 1];
1831 if (eng < ccr) {
1832 eng = energy[(hf->index << 1) + 1];
1833
1834 if (eng >= max_eng)
1835 hf->gain = 0x2800;
1836 else
1837 hf->gain = ((eng << 15) / max_eng * 0x2800 + (1 << 14)) >> 15;
1838 }
1839 hf->index += pitch_lag - 3;
1840}
1841
1842/**
1843 * Apply the harmonic noise shaping filter.
1844 *
1845 * @param hf filter parameters
1846 */
1847static void harmonic_filter(HFParam *hf, const int16_t *src, int16_t *dest)
1848{
1849 int i;
1850
1851 for (i = 0; i < SUBFRAME_LEN; i++) {
1852 int64_t temp = hf->gain * src[i - hf->index] << 1;
1853 dest[i] = av_clipl_int32((src[i] << 16) - temp + (1 << 15)) >> 16;
1854 }
1855}
1856
1857static void harmonic_noise_sub(HFParam *hf, const int16_t *src, int16_t *dest)
1858{
1859 int i;
1860 for (i = 0; i < SUBFRAME_LEN; i++) {
1861 int64_t temp = hf->gain * src[i - hf->index] << 1;
1862 dest[i] = av_clipl_int32(((dest[i] - src[i]) << 16) + temp +
1863 (1 << 15)) >> 16;
1864
1865 }
1866}
1867
1868/**
1869 * Combined synthesis and formant perceptual weighting filer.
1870 *
1871 * @param qnt_lpc quantized lpc coefficients
1872 * @param perf_lpc perceptual filter coefficients
1873 * @param perf_fir perceptual filter fir memory
1874 * @param perf_iir perceptual filter iir memory
1875 * @param scale the filter output will be scaled by 2^scale
1876 */
1877static void synth_percept_filter(int16_t *qnt_lpc, int16_t *perf_lpc,
1878 int16_t *perf_fir, int16_t *perf_iir,
1879 const int16_t *src, int16_t *dest, int scale)
1880{
1881 int i, j;
1882 int16_t buf_16[SUBFRAME_LEN + LPC_ORDER];
1883 int64_t buf[SUBFRAME_LEN];
1884
1885 int16_t *bptr_16 = buf_16 + LPC_ORDER;
1886
1887 memcpy(buf_16, perf_fir, sizeof(int16_t) * LPC_ORDER);
1888 memcpy(dest - LPC_ORDER, perf_iir, sizeof(int16_t) * LPC_ORDER);
1889
1890 for (i = 0; i < SUBFRAME_LEN; i++) {
1891 int64_t temp = 0;
1892 for (j = 1; j <= LPC_ORDER; j++)
1893 temp -= qnt_lpc[j - 1] * bptr_16[i - j];
1894
1895 buf[i] = (src[i] << 15) + (temp << 3);
1896 bptr_16[i] = av_clipl_int32(buf[i] + (1 << 15)) >> 16;
1897 }
1898
1899 for (i = 0; i < SUBFRAME_LEN; i++) {
1900 int64_t fir = 0, iir = 0;
1901 for (j = 1; j <= LPC_ORDER; j++) {
1902 fir -= perf_lpc[j - 1] * bptr_16[i - j];
1903 iir += perf_lpc[j + LPC_ORDER - 1] * dest[i - j];
1904 }
1905 dest[i] = av_clipl_int32(((buf[i] + (fir << 3)) << scale) + (iir << 3) +
1906 (1 << 15)) >> 16;
1907 }
1908 memcpy(perf_fir, buf_16 + SUBFRAME_LEN, sizeof(int16_t) * LPC_ORDER);
1909 memcpy(perf_iir, dest + SUBFRAME_LEN - LPC_ORDER,
1910 sizeof(int16_t) * LPC_ORDER);
1911}
1912
1913/**
1914 * Compute the adaptive codebook contribution.
1915 *
1916 * @param buf input signal
1917 * @param index the current subframe index
1918 */
1919static void acb_search(G723_1_Context *p, int16_t *residual,
1920 int16_t *impulse_resp, const int16_t *buf,
1921 int index)
1922{
1923
1924 int16_t flt_buf[PITCH_ORDER][SUBFRAME_LEN];
1925
1926 const int16_t *cb_tbl = adaptive_cb_gain85;
1927
1928 int ccr_buf[PITCH_ORDER * SUBFRAMES << 2];
1929
1930 int pitch_lag = p->pitch_lag[index >> 1];
1931 int acb_lag = 1;
1932 int acb_gain = 0;
1933 int odd_frame = index & 1;
1934 int iter = 3 + odd_frame;
1935 int count = 0;
1936 int tbl_size = 85;
1937
1938 int i, j, k, l, max;
1939 int64_t temp;
1940
1941 if (!odd_frame) {
1942 if (pitch_lag == PITCH_MIN)
1943 pitch_lag++;
1944 else
1945 pitch_lag = FFMIN(pitch_lag, PITCH_MAX - 5);
1946 }
1947
1948 for (i = 0; i < iter; i++) {
1949 get_residual(residual, p->prev_excitation, pitch_lag + i - 1);
1950
1951 for (j = 0; j < SUBFRAME_LEN; j++) {
1952 temp = 0;
1953 for (k = 0; k <= j; k++)
1954 temp += residual[PITCH_ORDER - 1 + k] * impulse_resp[j - k];
1955 flt_buf[PITCH_ORDER - 1][j] = av_clipl_int32((temp << 1) +
1956 (1 << 15)) >> 16;
1957 }
1958
1959 for (j = PITCH_ORDER - 2; j >= 0; j--) {
1960 flt_buf[j][0] = ((residual[j] << 13) + (1 << 14)) >> 15;
1961 for (k = 1; k < SUBFRAME_LEN; k++) {
1962 temp = (flt_buf[j + 1][k - 1] << 15) +
1963 residual[j] * impulse_resp[k];
1964 flt_buf[j][k] = av_clipl_int32((temp << 1) + (1 << 15)) >> 16;
1965 }
1966 }
1967
1968 /* Compute crosscorrelation with the signal */
1969 for (j = 0; j < PITCH_ORDER; j++) {
1970 temp = ff_dot_product(buf, flt_buf[j], SUBFRAME_LEN);
1971 ccr_buf[count++] = av_clipl_int32(temp << 1);
1972 }
1973
1974 /* Compute energies */
1975 for (j = 0; j < PITCH_ORDER; j++) {
1976 ccr_buf[count++] = dot_product(flt_buf[j], flt_buf[j],
1977 SUBFRAME_LEN);
1978 }
1979
1980 for (j = 1; j < PITCH_ORDER; j++) {
1981 for (k = 0; k < j; k++) {
1982 temp = ff_dot_product(flt_buf[j], flt_buf[k], SUBFRAME_LEN);
1983 ccr_buf[count++] = av_clipl_int32(temp<<2);
1984 }
1985 }
1986 }
1987
1988 /* Normalize and shorten */
1989 max = 0;
1990 for (i = 0; i < 20 * iter; i++)
1991 max = FFMAX(max, FFABS(ccr_buf[i]));
1992
1993 temp = normalize_bits_int32(max);
1994
1995 for (i = 0; i < 20 * iter; i++){
1996 ccr_buf[i] = av_clipl_int32((int64_t)(ccr_buf[i] << temp) +
1997 (1 << 15)) >> 16;
1998 }
1999
2000 max = 0;
2001 for (i = 0; i < iter; i++) {
2002 /* Select quantization table */
2003 if (!odd_frame && pitch_lag + i - 1 >= SUBFRAME_LEN - 2 ||
2004 odd_frame && pitch_lag >= SUBFRAME_LEN - 2) {
2005 cb_tbl = adaptive_cb_gain170;
2006 tbl_size = 170;
2007 }
2008
2009 for (j = 0, k = 0; j < tbl_size; j++, k += 20) {
2010 temp = 0;
2011 for (l = 0; l < 20; l++)
2012 temp += ccr_buf[20 * i + l] * cb_tbl[k + l];
2013 temp = av_clipl_int32(temp);
2014
2015 if (temp > max) {
2016 max = temp;
2017 acb_gain = j;
2018 acb_lag = i;
2019 }
2020 }
2021 }
2022
2023 if (!odd_frame) {
2024 pitch_lag += acb_lag - 1;
2025 acb_lag = 1;
2026 }
2027
2028 p->pitch_lag[index >> 1] = pitch_lag;
2029 p->subframe[index].ad_cb_lag = acb_lag;
2030 p->subframe[index].ad_cb_gain = acb_gain;
2031}
2032
2033/**
2034 * Subtract the adaptive codebook contribution from the input
2035 * to obtain the residual.
2036 *
2037 * @param buf target vector
2038 */
2039static void sub_acb_contrib(const int16_t *residual, const int16_t *impulse_resp,
2040 int16_t *buf)
2041{
2042 int i, j;
2043 /* Subtract adaptive CB contribution to obtain the residual */
2044 for (i = 0; i < SUBFRAME_LEN; i++) {
2045 int64_t temp = buf[i] << 14;
2046 for (j = 0; j <= i; j++)
2047 temp -= residual[j] * impulse_resp[i - j];
2048
2049 buf[i] = av_clipl_int32((temp << 2) + (1 << 15)) >> 16;
2050 }
2051}
2052
2053/**
2054 * Quantize the residual signal using the fixed codebook (MP-MLQ).
2055 *
2056 * @param optim optimized fixed codebook parameters
2057 * @param buf excitation vector
2058 */
2059static void get_fcb_param(FCBParam *optim, int16_t *impulse_resp,
2060 int16_t *buf, int pulse_cnt, int pitch_lag)
2061{
2062 FCBParam param;
2063 int16_t impulse_r[SUBFRAME_LEN];
2064 int16_t temp_corr[SUBFRAME_LEN];
2065 int16_t impulse_corr[SUBFRAME_LEN];
2066
2067 int ccr1[SUBFRAME_LEN];
2068 int ccr2[SUBFRAME_LEN];
2069 int amp, err, max, max_amp_index, min, scale, i, j, k, l;
2070
2071 int64_t temp;
2072
2073 /* Update impulse response */
2074 memcpy(impulse_r, impulse_resp, sizeof(int16_t) * SUBFRAME_LEN);
2075 param.dirac_train = 0;
2076 if (pitch_lag < SUBFRAME_LEN - 2) {
2077 param.dirac_train = 1;
2078 gen_dirac_train(impulse_r, pitch_lag);
2079 }
2080
2081 for (i = 0; i < SUBFRAME_LEN; i++)
2082 temp_corr[i] = impulse_r[i] >> 1;
2083
2084 /* Compute impulse response autocorrelation */
2085 temp = dot_product(temp_corr, temp_corr, SUBFRAME_LEN);
2086
2087 scale = normalize_bits_int32(temp);
2088 impulse_corr[0] = av_clipl_int32((temp << scale) + (1 << 15)) >> 16;
2089
2090 for (i = 1; i < SUBFRAME_LEN; i++) {
2091 temp = dot_product(temp_corr + i, temp_corr, SUBFRAME_LEN - i);
2092 impulse_corr[i] = av_clipl_int32((temp << scale) + (1 << 15)) >> 16;
2093 }
2094
2095 /* Compute crosscorrelation of impulse response with residual signal */
2096 scale -= 4;
2097 for (i = 0; i < SUBFRAME_LEN; i++){
2098 temp = dot_product(buf + i, impulse_r, SUBFRAME_LEN - i);
2099 if (scale < 0)
2100 ccr1[i] = temp >> -scale;
2101 else
2102 ccr1[i] = av_clipl_int32(temp << scale);
2103 }
2104
2105 /* Search loop */
2106 for (i = 0; i < GRID_SIZE; i++) {
2107 /* Maximize the crosscorrelation */
2108 max = 0;
2109 for (j = i; j < SUBFRAME_LEN; j += GRID_SIZE) {
2110 temp = FFABS(ccr1[j]);
2111 if (temp >= max) {
2112 max = temp;
2113 param.pulse_pos[0] = j;
2114 }
2115 }
2116
2117 /* Quantize the gain (max crosscorrelation/impulse_corr[0]) */
2118 amp = max;
2119 min = 1 << 30;
2120 max_amp_index = GAIN_LEVELS - 2;
2121 for (j = max_amp_index; j >= 2; j--) {
2122 temp = av_clipl_int32((int64_t)fixed_cb_gain[j] *
2123 impulse_corr[0] << 1);
2124 temp = FFABS(temp - amp);
2125 if (temp < min) {
2126 min = temp;
2127 max_amp_index = j;
2128 }
2129 }
2130
2131 max_amp_index--;
2132 /* Select additional gain values */
2133 for (j = 1; j < 5; j++) {
2134 for (k = i; k < SUBFRAME_LEN; k += GRID_SIZE) {
2135 temp_corr[k] = 0;
2136 ccr2[k] = ccr1[k];
2137 }
2138 param.amp_index = max_amp_index + j - 2;
2139 amp = fixed_cb_gain[param.amp_index];
2140
2141 param.pulse_sign[0] = (ccr2[param.pulse_pos[0]] < 0) ? -amp : amp;
2142 temp_corr[param.pulse_pos[0]] = 1;
2143
2144 for (k = 1; k < pulse_cnt; k++) {
2145 max = -1 << 30;
2146 for (l = i; l < SUBFRAME_LEN; l += GRID_SIZE) {
2147 if (temp_corr[l])
2148 continue;
2149 temp = impulse_corr[FFABS(l - param.pulse_pos[k - 1])];
2150 temp = av_clipl_int32((int64_t)temp *
2151 param.pulse_sign[k - 1] << 1);
2152 ccr2[l] -= temp;
2153 temp = FFABS(ccr2[l]);
2154 if (temp > max) {
2155 max = temp;
2156 param.pulse_pos[k] = l;
2157 }
2158 }
2159
2160 param.pulse_sign[k] = (ccr2[param.pulse_pos[k]] < 0) ?
2161 -amp : amp;
2162 temp_corr[param.pulse_pos[k]] = 1;
2163 }
2164
2165 /* Create the error vector */
2166 memset(temp_corr, 0, sizeof(int16_t) * SUBFRAME_LEN);
2167
2168 for (k = 0; k < pulse_cnt; k++)
2169 temp_corr[param.pulse_pos[k]] = param.pulse_sign[k];
2170
2171 for (k = SUBFRAME_LEN - 1; k >= 0; k--) {
2172 temp = 0;
2173 for (l = 0; l <= k; l++) {
2174 int prod = av_clipl_int32((int64_t)temp_corr[l] *
2175 impulse_r[k - l] << 1);
2176 temp = av_clipl_int32(temp + prod);
2177 }
2178 temp_corr[k] = temp << 2 >> 16;
2179 }
2180
2181 /* Compute square of error */
2182 err = 0;
2183 for (k = 0; k < SUBFRAME_LEN; k++) {
2184 int64_t prod;
2185 prod = av_clipl_int32((int64_t)buf[k] * temp_corr[k] << 1);
2186 err = av_clipl_int32(err - prod);
2187 prod = av_clipl_int32((int64_t)temp_corr[k] * temp_corr[k]);
2188 err = av_clipl_int32(err + prod);
2189 }
2190
2191 /* Minimize */
2192 if (err < optim->min_err) {
2193 optim->min_err = err;
2194 optim->grid_index = i;
2195 optim->amp_index = param.amp_index;
2196 optim->dirac_train = param.dirac_train;
2197
2198 for (k = 0; k < pulse_cnt; k++) {
2199 optim->pulse_sign[k] = param.pulse_sign[k];
2200 optim->pulse_pos[k] = param.pulse_pos[k];
2201 }
2202 }
2203 }
2204 }
2205}
2206
2207/**
2208 * Encode the pulse position and gain of the current subframe.
2209 *
2210 * @param optim optimized fixed CB parameters
2211 * @param buf excitation vector
2212 */
2213static void pack_fcb_param(G723_1_Subframe *subfrm, FCBParam *optim,
2214 int16_t *buf, int pulse_cnt)
2215{
2216 int i, j;
2217
2218 j = PULSE_MAX - pulse_cnt;
2219
2220 subfrm->pulse_sign = 0;
2221 subfrm->pulse_pos = 0;
2222
2223 for (i = 0; i < SUBFRAME_LEN >> 1; i++) {
2224 int val = buf[optim->grid_index + (i << 1)];
2225 if (!val) {
2226 subfrm->pulse_pos += combinatorial_table[j][i];
2227 } else {
2228 subfrm->pulse_sign <<= 1;
2229 if (val < 0) subfrm->pulse_sign++;
2230 j++;
2231
2232 if (j == PULSE_MAX) break;
2233 }
2234 }
2235 subfrm->amp_index = optim->amp_index;
2236 subfrm->grid_index = optim->grid_index;
2237 subfrm->dirac_train = optim->dirac_train;
2238}
2239
2240/**
2241 * Compute the fixed codebook excitation.
2242 *
2243 * @param buf target vector
2244 * @param impulse_resp impulse response of the combined filter
2245 */
2246static void fcb_search(G723_1_Context *p, int16_t *impulse_resp,
2247 int16_t *buf, int index)
2248{
2249 FCBParam optim;
2250 int pulse_cnt = pulses[index];
2251 int i;
2252
2253 optim.min_err = 1 << 30;
2254 get_fcb_param(&optim, impulse_resp, buf, pulse_cnt, SUBFRAME_LEN);
2255
2256 if (p->pitch_lag[index >> 1] < SUBFRAME_LEN - 2) {
2257 get_fcb_param(&optim, impulse_resp, buf, pulse_cnt,
2258 p->pitch_lag[index >> 1]);
2259 }
2260
2261 /* Reconstruct the excitation */
2262 memset(buf, 0, sizeof(int16_t) * SUBFRAME_LEN);
2263 for (i = 0; i < pulse_cnt; i++)
2264 buf[optim.pulse_pos[i]] = optim.pulse_sign[i];
2265
2266 pack_fcb_param(&p->subframe[index], &optim, buf, pulse_cnt);
2267
2268 if (optim.dirac_train)
2269 gen_dirac_train(buf, p->pitch_lag[index >> 1]);
2270}
2271
2272/**
2273 * Pack the frame parameters into output bitstream.
2274 *
2275 * @param frame output buffer
2276 * @param size size of the buffer
2277 */
2278static int pack_bitstream(G723_1_Context *p, unsigned char *frame, int size)
2279{
2280 PutBitContext pb;
2281 int info_bits, i, temp;
2282
2283 init_put_bits(&pb, frame, size);
2284
2285 if (p->cur_rate == RATE_6300) {
2286 info_bits = 0;
2287 put_bits(&pb, 2, info_bits);
2288 }else
2289 av_assert0(0);
2290
2291 put_bits(&pb, 8, p->lsp_index[2]);
2292 put_bits(&pb, 8, p->lsp_index[1]);
2293 put_bits(&pb, 8, p->lsp_index[0]);
2294
2295 put_bits(&pb, 7, p->pitch_lag[0] - PITCH_MIN);
2296 put_bits(&pb, 2, p->subframe[1].ad_cb_lag);
2297 put_bits(&pb, 7, p->pitch_lag[1] - PITCH_MIN);
2298 put_bits(&pb, 2, p->subframe[3].ad_cb_lag);
2299
2300 /* Write 12 bit combined gain */
2301 for (i = 0; i < SUBFRAMES; i++) {
2302 temp = p->subframe[i].ad_cb_gain * GAIN_LEVELS +
2303 p->subframe[i].amp_index;
2304 if (p->cur_rate == RATE_6300)
2305 temp += p->subframe[i].dirac_train << 11;
2306 put_bits(&pb, 12, temp);
2307 }
2308
2309 put_bits(&pb, 1, p->subframe[0].grid_index);
2310 put_bits(&pb, 1, p->subframe[1].grid_index);
2311 put_bits(&pb, 1, p->subframe[2].grid_index);
2312 put_bits(&pb, 1, p->subframe[3].grid_index);
2313
2314 if (p->cur_rate == RATE_6300) {
2315 skip_put_bits(&pb, 1); /* reserved bit */
2316
2317 /* Write 13 bit combined position index */
2318 temp = (p->subframe[0].pulse_pos >> 16) * 810 +
2319 (p->subframe[1].pulse_pos >> 14) * 90 +
2320 (p->subframe[2].pulse_pos >> 16) * 9 +
2321 (p->subframe[3].pulse_pos >> 14);
2322 put_bits(&pb, 13, temp);
2323
2324 put_bits(&pb, 16, p->subframe[0].pulse_pos & 0xffff);
2325 put_bits(&pb, 14, p->subframe[1].pulse_pos & 0x3fff);
2326 put_bits(&pb, 16, p->subframe[2].pulse_pos & 0xffff);
2327 put_bits(&pb, 14, p->subframe[3].pulse_pos & 0x3fff);
2328
2329 put_bits(&pb, 6, p->subframe[0].pulse_sign);
2330 put_bits(&pb, 5, p->subframe[1].pulse_sign);
2331 put_bits(&pb, 6, p->subframe[2].pulse_sign);
2332 put_bits(&pb, 5, p->subframe[3].pulse_sign);
2333 }
2334
2335 flush_put_bits(&pb);
2336 return frame_size[info_bits];
2337}
2338
2339static int g723_1_encode_frame(AVCodecContext *avctx, AVPacket *avpkt,
2340 const AVFrame *frame, int *got_packet_ptr)
2341{
2342 G723_1_Context *p = avctx->priv_data;
2343 int16_t unq_lpc[LPC_ORDER * SUBFRAMES];
2344 int16_t qnt_lpc[LPC_ORDER * SUBFRAMES];
2345 int16_t cur_lsp[LPC_ORDER];
2346 int16_t weighted_lpc[LPC_ORDER * SUBFRAMES << 1];
2347 int16_t vector[FRAME_LEN + PITCH_MAX];
2348 int offset, ret;
2349 int16_t *in_orig = av_memdup(frame->data[0], frame->nb_samples * sizeof(int16_t));
2350 int16_t *in = in_orig;
2351
2352 HFParam hf[4];
2353 int i, j;
2354
2355 if (!in)
2356 return AVERROR(ENOMEM);
2357
2358 highpass_filter(in, &p->hpf_fir_mem, &p->hpf_iir_mem);
2359
2360 memcpy(vector, p->prev_data, HALF_FRAME_LEN * sizeof(int16_t));
2361 memcpy(vector + HALF_FRAME_LEN, in, FRAME_LEN * sizeof(int16_t));
2362
2363 comp_lpc_coeff(vector, unq_lpc);
2364 lpc2lsp(&unq_lpc[LPC_ORDER * 3], p->prev_lsp, cur_lsp);
2365 lsp_quantize(p->lsp_index, cur_lsp, p->prev_lsp);
2366
2367 /* Update memory */
2368 memcpy(vector + LPC_ORDER, p->prev_data + SUBFRAME_LEN,
2369 sizeof(int16_t) * SUBFRAME_LEN);
2370 memcpy(vector + LPC_ORDER + SUBFRAME_LEN, in,
2371 sizeof(int16_t) * (HALF_FRAME_LEN + SUBFRAME_LEN));
2372 memcpy(p->prev_data, in + HALF_FRAME_LEN,
2373 sizeof(int16_t) * HALF_FRAME_LEN);
2374 memcpy(in, vector + LPC_ORDER, sizeof(int16_t) * FRAME_LEN);
2375
2376 perceptual_filter(p, weighted_lpc, unq_lpc, vector);
2377
2378 memcpy(in, vector + LPC_ORDER, sizeof(int16_t) * FRAME_LEN);
2379 memcpy(vector, p->prev_weight_sig, sizeof(int16_t) * PITCH_MAX);
2380 memcpy(vector + PITCH_MAX, in, sizeof(int16_t) * FRAME_LEN);
2381
2382 scale_vector(vector, vector, FRAME_LEN + PITCH_MAX);
2383
2384 p->pitch_lag[0] = estimate_pitch(vector, PITCH_MAX);
2385 p->pitch_lag[1] = estimate_pitch(vector, PITCH_MAX + HALF_FRAME_LEN);
2386
2387 for (i = PITCH_MAX, j = 0; j < SUBFRAMES; i += SUBFRAME_LEN, j++)
2388 comp_harmonic_coeff(vector + i, p->pitch_lag[j >> 1], hf + j);
2389
2390 memcpy(vector, p->prev_weight_sig, sizeof(int16_t) * PITCH_MAX);
2391 memcpy(vector + PITCH_MAX, in, sizeof(int16_t) * FRAME_LEN);
2392 memcpy(p->prev_weight_sig, vector + FRAME_LEN, sizeof(int16_t) * PITCH_MAX);
2393
2394 for (i = 0, j = 0; j < SUBFRAMES; i += SUBFRAME_LEN, j++)
2395 harmonic_filter(hf + j, vector + PITCH_MAX + i, in + i);
2396
2397 inverse_quant(cur_lsp, p->prev_lsp, p->lsp_index, 0);
2398 lsp_interpolate(qnt_lpc, cur_lsp, p->prev_lsp);
2399
2400 memcpy(p->prev_lsp, cur_lsp, sizeof(int16_t) * LPC_ORDER);
2401
2402 offset = 0;
2403 for (i = 0; i < SUBFRAMES; i++) {
2404 int16_t impulse_resp[SUBFRAME_LEN];
2405 int16_t residual[SUBFRAME_LEN + PITCH_ORDER - 1];
2406 int16_t flt_in[SUBFRAME_LEN];
2407 int16_t zero[LPC_ORDER], fir[LPC_ORDER], iir[LPC_ORDER];
2408
2409 /**
2410 * Compute the combined impulse response of the synthesis filter,
2411 * formant perceptual weighting filter and harmonic noise shaping filter
2412 */
2413 memset(zero, 0, sizeof(int16_t) * LPC_ORDER);
2414 memset(vector, 0, sizeof(int16_t) * PITCH_MAX);
2415 memset(flt_in, 0, sizeof(int16_t) * SUBFRAME_LEN);
2416
2417 flt_in[0] = 1 << 13; /* Unit impulse */
2418 synth_percept_filter(qnt_lpc + offset, weighted_lpc + (offset << 1),
2419 zero, zero, flt_in, vector + PITCH_MAX, 1);
2420 harmonic_filter(hf + i, vector + PITCH_MAX, impulse_resp);
2421
2422 /* Compute the combined zero input response */
2423 flt_in[0] = 0;
2424 memcpy(fir, p->perf_fir_mem, sizeof(int16_t) * LPC_ORDER);
2425 memcpy(iir, p->perf_iir_mem, sizeof(int16_t) * LPC_ORDER);
2426
2427 synth_percept_filter(qnt_lpc + offset, weighted_lpc + (offset << 1),
2428 fir, iir, flt_in, vector + PITCH_MAX, 0);
2429 memcpy(vector, p->harmonic_mem, sizeof(int16_t) * PITCH_MAX);
2430 harmonic_noise_sub(hf + i, vector + PITCH_MAX, in);
2431
2432 acb_search(p, residual, impulse_resp, in, i);
2433 gen_acb_excitation(residual, p->prev_excitation,p->pitch_lag[i >> 1],
2434 &p->subframe[i], p->cur_rate);
2435 sub_acb_contrib(residual, impulse_resp, in);
2436
2437 fcb_search(p, impulse_resp, in, i);
2438
2439 /* Reconstruct the excitation */
2440 gen_acb_excitation(impulse_resp, p->prev_excitation, p->pitch_lag[i >> 1],
2441 &p->subframe[i], RATE_6300);
2442
2443 memmove(p->prev_excitation, p->prev_excitation + SUBFRAME_LEN,
2444 sizeof(int16_t) * (PITCH_MAX - SUBFRAME_LEN));
2445 for (j = 0; j < SUBFRAME_LEN; j++)
2446 in[j] = av_clip_int16((in[j] << 1) + impulse_resp[j]);
2447 memcpy(p->prev_excitation + PITCH_MAX - SUBFRAME_LEN, in,
2448 sizeof(int16_t) * SUBFRAME_LEN);
2449
2450 /* Update filter memories */
2451 synth_percept_filter(qnt_lpc + offset, weighted_lpc + (offset << 1),
2452 p->perf_fir_mem, p->perf_iir_mem,
2453 in, vector + PITCH_MAX, 0);
2454 memmove(p->harmonic_mem, p->harmonic_mem + SUBFRAME_LEN,
2455 sizeof(int16_t) * (PITCH_MAX - SUBFRAME_LEN));
2456 memcpy(p->harmonic_mem + PITCH_MAX - SUBFRAME_LEN, vector + PITCH_MAX,
2457 sizeof(int16_t) * SUBFRAME_LEN);
2458
2459 in += SUBFRAME_LEN;
2460 offset += LPC_ORDER;
2461 }
2462
2463 av_freep(&in_orig); in = NULL;
2464
2465 if ((ret = ff_alloc_packet2(avctx, avpkt, 24)) < 0)
2466 return ret;
2467
2468 *got_packet_ptr = 1;
2469 avpkt->size = pack_bitstream(p, avpkt->data, avpkt->size);
2470 return 0;
2471}
2472
2473AVCodec ff_g723_1_encoder = {
2474 .name = "g723_1",
2475 .long_name = NULL_IF_CONFIG_SMALL("G.723.1"),
2476 .type = AVMEDIA_TYPE_AUDIO,
2477 .id = AV_CODEC_ID_G723_1,
2478 .priv_data_size = sizeof(G723_1_Context),
2479 .init = g723_1_encode_init,
2480 .encode2 = g723_1_encode_frame,
2481 .sample_fmts = (const enum AVSampleFormat[]){AV_SAMPLE_FMT_S16,
2482 AV_SAMPLE_FMT_NONE},
2483};
2484#endif
ffmpeg Piment Noir Debian package repository