aboutsummaryrefslogtreecommitdiffstats
path: root/libavcodec/alacenc.c
blob: d9ff2b8e0a4346c39fb201dfb8ab3109b97b03fd (plain) (blame)
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
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
/**
 * ALAC audio encoder
 * Copyright (c) 2008  Jaikrishnan Menon <realityman@gmx.net>
 *
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * FFmpeg is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with FFmpeg; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

#include "avcodec.h"
#include "put_bits.h"
#include "dsputil.h"
#include "lpc.h"
#include "mathops.h"

#define DEFAULT_FRAME_SIZE        4096
#define DEFAULT_SAMPLE_SIZE       16
#define MAX_CHANNELS              8
#define ALAC_EXTRADATA_SIZE       36
#define ALAC_FRAME_HEADER_SIZE    55
#define ALAC_FRAME_FOOTER_SIZE    3

#define ALAC_ESCAPE_CODE          0x1FF
#define ALAC_MAX_LPC_ORDER        30
#define DEFAULT_MAX_PRED_ORDER    6
#define DEFAULT_MIN_PRED_ORDER    4
#define ALAC_MAX_LPC_PRECISION    9
#define ALAC_MAX_LPC_SHIFT        9

#define ALAC_CHMODE_LEFT_RIGHT    0
#define ALAC_CHMODE_LEFT_SIDE     1
#define ALAC_CHMODE_RIGHT_SIDE    2
#define ALAC_CHMODE_MID_SIDE      3

typedef struct RiceContext {
    int history_mult;
    int initial_history;
    int k_modifier;
    int rice_modifier;
} RiceContext;

typedef struct AlacLPCContext {
    int lpc_order;
    int lpc_coeff[ALAC_MAX_LPC_ORDER+1];
    int lpc_quant;
} AlacLPCContext;

typedef struct AlacEncodeContext {
    int compression_level;
    int min_prediction_order;
    int max_prediction_order;
    int max_coded_frame_size;
    int write_sample_size;
    int32_t sample_buf[MAX_CHANNELS][DEFAULT_FRAME_SIZE];
    int32_t predictor_buf[DEFAULT_FRAME_SIZE];
    int interlacing_shift;
    int interlacing_leftweight;
    PutBitContext pbctx;
    RiceContext rc;
    AlacLPCContext lpc[MAX_CHANNELS];
    LPCContext lpc_ctx;
    AVCodecContext *avctx;
} AlacEncodeContext;


static void init_sample_buffers(AlacEncodeContext *s, const int16_t *input_samples)
{
    int ch, i;

    for(ch=0;ch<s->avctx->channels;ch++) {
        const int16_t *sptr = input_samples + ch;
        for(i=0;i<s->avctx->frame_size;i++) {
            s->sample_buf[ch][i] = *sptr;
            sptr += s->avctx->channels;
        }
    }
}

static void encode_scalar(AlacEncodeContext *s, int x, int k, int write_sample_size)
{
    int divisor, q, r;

    k = FFMIN(k, s->rc.k_modifier);
    divisor = (1<<k) - 1;
    q = x / divisor;
    r = x % divisor;

    if(q > 8) {
        // write escape code and sample value directly
        put_bits(&s->pbctx, 9, ALAC_ESCAPE_CODE);
        put_bits(&s->pbctx, write_sample_size, x);
    } else {
        if(q)
            put_bits(&s->pbctx, q, (1<<q) - 1);
        put_bits(&s->pbctx, 1, 0);

        if(k != 1) {
            if(r > 0)
                put_bits(&s->pbctx, k, r+1);
            else
                put_bits(&s->pbctx, k-1, 0);
        }
    }
}

static void write_frame_header(AlacEncodeContext *s, int is_verbatim)
{
    put_bits(&s->pbctx, 3,  s->avctx->channels-1);          // No. of channels -1
    put_bits(&s->pbctx, 16, 0);                             // Seems to be zero
    put_bits(&s->pbctx, 1,  1);                             // Sample count is in the header
    put_bits(&s->pbctx, 2,  0);                             // FIXME: Wasted bytes field
    put_bits(&s->pbctx, 1,  is_verbatim);                   // Audio block is verbatim
    put_bits32(&s->pbctx, s->avctx->frame_size);            // No. of samples in the frame
}

static void calc_predictor_params(AlacEncodeContext *s, int ch)
{
    int32_t coefs[MAX_LPC_ORDER][MAX_LPC_ORDER];
    int shift[MAX_LPC_ORDER];
    int opt_order;

    if (s->compression_level == 1) {
        s->lpc[ch].lpc_order = 6;
        s->lpc[ch].lpc_quant = 6;
        s->lpc[ch].lpc_coeff[0] =  160;
        s->lpc[ch].lpc_coeff[1] = -190;
        s->lpc[ch].lpc_coeff[2] =  170;
        s->lpc[ch].lpc_coeff[3] = -130;
        s->lpc[ch].lpc_coeff[4] =   80;
        s->lpc[ch].lpc_coeff[5] =  -25;
    } else {
        opt_order = ff_lpc_calc_coefs(&s->lpc_ctx, s->sample_buf[ch],
                                      s->avctx->frame_size,
                                      s->min_prediction_order,
                                      s->max_prediction_order,
                                      ALAC_MAX_LPC_PRECISION, coefs, shift,
                                      AV_LPC_TYPE_LEVINSON, 0,
                                      ORDER_METHOD_EST, ALAC_MAX_LPC_SHIFT, 1);

        s->lpc[ch].lpc_order = opt_order;
        s->lpc[ch].lpc_quant = shift[opt_order-1];
        memcpy(s->lpc[ch].lpc_coeff, coefs[opt_order-1], opt_order*sizeof(int));
    }
}

static int estimate_stereo_mode(int32_t *left_ch, int32_t *right_ch, int n)
{
    int i, best;
    int32_t lt, rt;
    uint64_t sum[4];
    uint64_t score[4];

    /* calculate sum of 2nd order residual for each channel */
    sum[0] = sum[1] = sum[2] = sum[3] = 0;
    for(i=2; i<n; i++) {
        lt = left_ch[i] - 2*left_ch[i-1] + left_ch[i-2];
        rt = right_ch[i] - 2*right_ch[i-1] + right_ch[i-2];
        sum[2] += FFABS((lt + rt) >> 1);
        sum[3] += FFABS(lt - rt);
        sum[0] += FFABS(lt);
        sum[1] += FFABS(rt);
    }

    /* calculate score for each mode */
    score[0] = sum[0] + sum[1];
    score[1] = sum[0] + sum[3];
    score[2] = sum[1] + sum[3];
    score[3] = sum[2] + sum[3];

    /* return mode with lowest score */
    best = 0;
    for(i=1; i<4; i++) {
        if(score[i] < score[best]) {
            best = i;
        }
    }
    return best;
}

static void alac_stereo_decorrelation(AlacEncodeContext *s)
{
    int32_t *left = s->sample_buf[0], *right = s->sample_buf[1];
    int i, mode, n = s->avctx->frame_size;
    int32_t tmp;

    mode = estimate_stereo_mode(left, right, n);

    switch(mode)
    {
        case ALAC_CHMODE_LEFT_RIGHT:
            s->interlacing_leftweight = 0;
            s->interlacing_shift = 0;
            break;

        case ALAC_CHMODE_LEFT_SIDE:
            for(i=0; i<n; i++) {
                right[i] = left[i] - right[i];
            }
            s->interlacing_leftweight = 1;
            s->interlacing_shift = 0;
            break;

        case ALAC_CHMODE_RIGHT_SIDE:
            for(i=0; i<n; i++) {
                tmp = right[i];
                right[i] = left[i] - right[i];
                left[i] = tmp + (right[i] >> 31);
            }
            s->interlacing_leftweight = 1;
            s->interlacing_shift = 31;
            break;

        default:
            for(i=0; i<n; i++) {
                tmp = left[i];
                left[i] = (tmp + right[i]) >> 1;
                right[i] = tmp - right[i];
            }
            s->interlacing_leftweight = 1;
            s->interlacing_shift = 1;
            break;
    }
}

static void alac_linear_predictor(AlacEncodeContext *s, int ch)
{
    int i;
    AlacLPCContext lpc = s->lpc[ch];

    if(lpc.lpc_order == 31) {
        s->predictor_buf[0] = s->sample_buf[ch][0];

        for(i=1; i<s->avctx->frame_size; i++)
            s->predictor_buf[i] = s->sample_buf[ch][i] - s->sample_buf[ch][i-1];

        return;
    }

    // generalised linear predictor

    if(lpc.lpc_order > 0) {
        int32_t *samples  = s->sample_buf[ch];
        int32_t *residual = s->predictor_buf;

        // generate warm-up samples
        residual[0] = samples[0];
        for(i=1;i<=lpc.lpc_order;i++)
            residual[i] = samples[i] - samples[i-1];

        // perform lpc on remaining samples
        for(i = lpc.lpc_order + 1; i < s->avctx->frame_size; i++) {
            int sum = 1 << (lpc.lpc_quant - 1), res_val, j;

            for (j = 0; j < lpc.lpc_order; j++) {
                sum += (samples[lpc.lpc_order-j] - samples[0]) *
                        lpc.lpc_coeff[j];
            }

            sum >>= lpc.lpc_quant;
            sum += samples[0];
            residual[i] = sign_extend(samples[lpc.lpc_order+1] - sum,
                                      s->write_sample_size);
            res_val = residual[i];

            if(res_val) {
                int index = lpc.lpc_order - 1;
                int neg = (res_val < 0);

                while(index >= 0 && (neg ? (res_val < 0):(res_val > 0))) {
                    int val = samples[0] - samples[lpc.lpc_order - index];
                    int sign = (val ? FFSIGN(val) : 0);

                    if(neg)
                        sign*=-1;

                    lpc.lpc_coeff[index] -= sign;
                    val *= sign;
                    res_val -= ((val >> lpc.lpc_quant) *
                            (lpc.lpc_order - index));
                    index--;
                }
            }
            samples++;
        }
    }
}

static void alac_entropy_coder(AlacEncodeContext *s)
{
    unsigned int history = s->rc.initial_history;
    int sign_modifier = 0, i, k;
    int32_t *samples = s->predictor_buf;

    for(i=0;i < s->avctx->frame_size;) {
        int x;

        k = av_log2((history >> 9) + 3);

        x = -2*(*samples)-1;
        x ^= (x>>31);

        samples++;
        i++;

        encode_scalar(s, x - sign_modifier, k, s->write_sample_size);

        history += x * s->rc.history_mult
                   - ((history * s->rc.history_mult) >> 9);

        sign_modifier = 0;
        if(x > 0xFFFF)
            history = 0xFFFF;

        if((history < 128) && (i < s->avctx->frame_size)) {
            unsigned int block_size = 0;

            k = 7 - av_log2(history) + ((history + 16) >> 6);

            while((*samples == 0) && (i < s->avctx->frame_size)) {
                samples++;
                i++;
                block_size++;
            }
            encode_scalar(s, block_size, k, 16);

            sign_modifier = (block_size <= 0xFFFF);

            history = 0;
        }

    }
}

static void write_compressed_frame(AlacEncodeContext *s)
{
    int i, j;

    if(s->avctx->channels == 2)
        alac_stereo_decorrelation(s);
    put_bits(&s->pbctx, 8, s->interlacing_shift);
    put_bits(&s->pbctx, 8, s->interlacing_leftweight);

    for(i=0;i<s->avctx->channels;i++) {

        calc_predictor_params(s, i);

        put_bits(&s->pbctx, 4, 0);  // prediction type : currently only type 0 has been RE'd
        put_bits(&s->pbctx, 4, s->lpc[i].lpc_quant);

        put_bits(&s->pbctx, 3, s->rc.rice_modifier);
        put_bits(&s->pbctx, 5, s->lpc[i].lpc_order);
        // predictor coeff. table
        for(j=0;j<s->lpc[i].lpc_order;j++) {
            put_sbits(&s->pbctx, 16, s->lpc[i].lpc_coeff[j]);
        }
    }

    // apply lpc and entropy coding to audio samples

    for(i=0;i<s->avctx->channels;i++) {
        alac_linear_predictor(s, i);
        alac_entropy_coder(s);
    }
}

static av_cold int alac_encode_init(AVCodecContext *avctx)
{
    AlacEncodeContext *s    = avctx->priv_data;
    int ret;
    uint8_t *alac_extradata = av_mallocz(ALAC_EXTRADATA_SIZE+1);

    avctx->frame_size      = DEFAULT_FRAME_SIZE;
    avctx->bits_per_coded_sample = DEFAULT_SAMPLE_SIZE;

    if(avctx->sample_fmt != AV_SAMPLE_FMT_S16) {
        av_log(avctx, AV_LOG_ERROR, "only pcm_s16 input samples are supported\n");
        return -1;
    }

    // Set default compression level
    if(avctx->compression_level == FF_COMPRESSION_DEFAULT)
        s->compression_level = 2;
    else
        s->compression_level = av_clip(avctx->compression_level, 0, 2);

    // Initialize default Rice parameters
    s->rc.history_mult    = 40;
    s->rc.initial_history = 10;
    s->rc.k_modifier      = 14;
    s->rc.rice_modifier   = 4;

    s->max_coded_frame_size = 8 + (avctx->frame_size*avctx->channels*avctx->bits_per_coded_sample>>3);

    s->write_sample_size  = avctx->bits_per_coded_sample + avctx->channels - 1; // FIXME: consider wasted_bytes

    AV_WB32(alac_extradata,    ALAC_EXTRADATA_SIZE);
    AV_WB32(alac_extradata+4,  MKBETAG('a','l','a','c'));
    AV_WB32(alac_extradata+12, avctx->frame_size);
    AV_WB8 (alac_extradata+17, avctx->bits_per_coded_sample);
    AV_WB8 (alac_extradata+21, avctx->channels);
    AV_WB32(alac_extradata+24, s->max_coded_frame_size);
    AV_WB32(alac_extradata+28, avctx->sample_rate*avctx->channels*avctx->bits_per_coded_sample); // average bitrate
    AV_WB32(alac_extradata+32, avctx->sample_rate);

    // Set relevant extradata fields
    if(s->compression_level > 0) {
        AV_WB8(alac_extradata+18, s->rc.history_mult);
        AV_WB8(alac_extradata+19, s->rc.initial_history);
        AV_WB8(alac_extradata+20, s->rc.k_modifier);
    }

    s->min_prediction_order = DEFAULT_MIN_PRED_ORDER;
    if(avctx->min_prediction_order >= 0) {
        if(avctx->min_prediction_order < MIN_LPC_ORDER ||
           avctx->min_prediction_order > ALAC_MAX_LPC_ORDER) {
            av_log(avctx, AV_LOG_ERROR, "invalid min prediction order: %d\n", avctx->min_prediction_order);
                return -1;
        }

        s->min_prediction_order = avctx->min_prediction_order;
    }

    s->max_prediction_order = DEFAULT_MAX_PRED_ORDER;
    if(avctx->max_prediction_order >= 0) {
        if(avctx->max_prediction_order < MIN_LPC_ORDER ||
           avctx->max_prediction_order > ALAC_MAX_LPC_ORDER) {
            av_log(avctx, AV_LOG_ERROR, "invalid max prediction order: %d\n", avctx->max_prediction_order);
                return -1;
        }

        s->max_prediction_order = avctx->max_prediction_order;
    }

    if(s->max_prediction_order < s->min_prediction_order) {
        av_log(avctx, AV_LOG_ERROR, "invalid prediction orders: min=%d max=%d\n",
               s->min_prediction_order, s->max_prediction_order);
        return -1;
    }

    avctx->extradata = alac_extradata;
    avctx->extradata_size = ALAC_EXTRADATA_SIZE;

    avctx->coded_frame = avcodec_alloc_frame();
    avctx->coded_frame->key_frame = 1;

    s->avctx = avctx;
    ret = ff_lpc_init(&s->lpc_ctx, avctx->frame_size, s->max_prediction_order,
                      AV_LPC_TYPE_LEVINSON);

    return ret;
}

static int alac_encode_frame(AVCodecContext *avctx, uint8_t *frame,
                             int buf_size, void *data)
{
    AlacEncodeContext *s = avctx->priv_data;
    PutBitContext *pb = &s->pbctx;
    int i, out_bytes, verbatim_flag = 0;

    if(avctx->frame_size > DEFAULT_FRAME_SIZE) {
        av_log(avctx, AV_LOG_ERROR, "input frame size exceeded\n");
        return -1;
    }

    if(buf_size < 2*s->max_coded_frame_size) {
        av_log(avctx, AV_LOG_ERROR, "buffer size is too small\n");
        return -1;
    }

verbatim:
    init_put_bits(pb, frame, buf_size);

    if((s->compression_level == 0) || verbatim_flag) {
        // Verbatim mode
        const int16_t *samples = data;
        write_frame_header(s, 1);
        for(i=0; i<avctx->frame_size*avctx->channels; i++) {
            put_sbits(pb, 16, *samples++);
        }
    } else {
        init_sample_buffers(s, data);
        write_frame_header(s, 0);
        write_compressed_frame(s);
    }

    put_bits(pb, 3, 7);
    flush_put_bits(pb);
    out_bytes = put_bits_count(pb) >> 3;

    if(out_bytes > s->max_coded_frame_size) {
        /* frame too large. use verbatim mode */
        if(verbatim_flag || (s->compression_level == 0)) {
            /* still too large. must be an error. */
            av_log(avctx, AV_LOG_ERROR, "error encoding frame\n");
            return -1;
        }
        verbatim_flag = 1;
        goto verbatim;
    }

    return out_bytes;
}

static av_cold int alac_encode_close(AVCodecContext *avctx)
{
    AlacEncodeContext *s = avctx->priv_data;
    ff_lpc_end(&s->lpc_ctx);
    av_freep(&avctx->extradata);
    avctx->extradata_size = 0;
    av_freep(&avctx->coded_frame);
    return 0;
}

AVCodec alac_encoder = {
    "alac",
    AVMEDIA_TYPE_AUDIO,
    CODEC_ID_ALAC,
    sizeof(AlacEncodeContext),
    alac_encode_init,
    alac_encode_frame,
    alac_encode_close,
    .capabilities = CODEC_CAP_SMALL_LAST_FRAME,
    .sample_fmts = (const enum AVSampleFormat[]){ AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_NONE},
    .long_name = NULL_IF_CONFIG_SMALL("ALAC (Apple Lossless Audio Codec)"),
};