mp3enc.c 20.3 KB
Newer Older
1
/*
2
 * MP3 muxer
3
 * Copyright (c) 2003 Fabrice Bellard
4
 *
5 6 7
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
8 9
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
10
 * version 2.1 of the License, or (at your option) any later version.
11
 *
12
 * FFmpeg is distributed in the hope that it will be useful,
13 14 15 16 17
 * 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
18
 * License along with FFmpeg; if not, write to the Free Software
19
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20
 */
21

22
#include "avformat.h"
23
#include "avio_internal.h"
24
#include "id3v1.h"
25
#include "id3v2.h"
26
#include "rawenc.h"
27
#include "libavutil/avstring.h"
28 29 30
#include "libavcodec/mpegaudio.h"
#include "libavcodec/mpegaudiodata.h"
#include "libavcodec/mpegaudiodecheader.h"
31
#include "libavutil/intreadwrite.h"
32
#include "libavutil/opt.h"
33
#include "libavutil/dict.h"
34
#include "libavutil/avassert.h"
35 36 37
#include "libavutil/crc.h"
#include "libavutil/mathematics.h"
#include "libavutil/replaygain.h"
Andreas Öman's avatar
Andreas Öman committed
38

39 40
static int id3v1_set_string(AVFormatContext *s, const char *key,
                            uint8_t *buf, int buf_size)
41
{
42 43
    AVDictionaryEntry *tag;
    if ((tag = av_dict_get(s->metadata, key, NULL, 0)))
44
        av_strlcpy(buf, tag->value, buf_size);
45 46 47 48 49
    return !!tag;
}

static int id3v1_create_tag(AVFormatContext *s, uint8_t *buf)
{
50
    AVDictionaryEntry *tag;
51
    int i, count = 0;
52 53 54 55 56

    memset(buf, 0, ID3v1_TAG_SIZE); /* fail safe */
    buf[0] = 'T';
    buf[1] = 'A';
    buf[2] = 'G';
57 58 59 60
    /* we knowingly overspecify each tag length by one byte to compensate for the mandatory null byte added by av_strlcpy */
    count += id3v1_set_string(s, "TIT2",    buf +  3, 30 + 1);       //title
    count += id3v1_set_string(s, "TPE1",    buf + 33, 30 + 1);       //author|artist
    count += id3v1_set_string(s, "TALB",    buf + 63, 30 + 1);       //album
61
    count += id3v1_set_string(s, "TDRC",    buf + 93,  4 + 1);       //date
62
    count += id3v1_set_string(s, "comment", buf + 97, 30 + 1);
63
    if ((tag = av_dict_get(s->metadata, "TRCK", NULL, 0))) { //track
64
        buf[125] = 0;
65 66
        buf[126] = atoi(tag->value);
        count++;
67
    }
68
    buf[127] = 0xFF; /* default to unknown genre */
69
    if ((tag = av_dict_get(s->metadata, "TCON", NULL, 0))) { //genre
Aurelien Jacobs's avatar
Aurelien Jacobs committed
70
        for(i = 0; i <= ID3v1_GENRE_MAX; i++) {
71
            if (!av_strcasecmp(tag->value, ff_id3v1_genre_str[i])) {
Aurelien Jacobs's avatar
Aurelien Jacobs committed
72 73 74 75
                buf[127] = i;
                count++;
                break;
            }
76 77
        }
    }
78
    return count;
79 80
}

Anton Khirnov's avatar
Anton Khirnov committed
81 82
#define XING_NUM_BAGS 400
#define XING_TOC_SIZE 100
83 84
// size of the XING/LAME data, starting from the Xing tag
#define XING_SIZE 156
85

86 87
typedef struct MP3Context {
    const AVClass *class;
88
    ID3v2EncContext id3;
89
    int id3v2_version;
90
    int write_id3v1;
91
    int write_xing;
Anton Khirnov's avatar
Anton Khirnov committed
92 93

    /* xing header */
94 95 96 97 98 99 100 101 102 103 104 105
    // a buffer containing the whole XING/LAME frame
    uint8_t *xing_frame;
    int      xing_frame_size;

    AVCRC    audio_crc;     // CRC of the audio data
    uint32_t audio_size;    // total size of the audio data

    // offset of the XING/LAME frame in the file
    int64_t  xing_frame_offset;
    // offset of the XING/INFO tag in the frame
    int xing_offset;

106 107 108 109 110
    int32_t frames;
    int32_t size;
    uint32_t want;
    uint32_t seen;
    uint32_t pos;
Anton Khirnov's avatar
Anton Khirnov committed
111
    uint64_t bag[XING_NUM_BAGS];
112 113
    int initial_bitrate;
    int has_variable_bitrate;
114 115
    int delay;
    int padding;
116

117 118 119 120
    /* index of the audio stream */
    int audio_stream_idx;
    /* number of attached pictures we still need to write */
    int pics_to_write;
121

122 123
    /* audio packets are queued here until we get all the attached pictures */
    AVPacketList *queue, *queue_end;
124
} MP3Context;
125

126
static const uint8_t xing_offtbl[2][2] = {{32, 17}, {17, 9}};
127 128 129 130 131 132 133

/*
 * Write an empty XING header and initialize respective data.
 */
static int mp3_write_xing(AVFormatContext *s)
{
    MP3Context       *mp3 = s->priv_data;
134
    AVCodecParameters *par = s->streams[mp3->audio_stream_idx]->codecpar;
135 136
    AVDictionaryEntry *enc = av_dict_get(s->streams[mp3->audio_stream_idx]->metadata, "encoder", NULL, 0);
    AVIOContext *dyn_ctx;
137 138 139
    int32_t        header;
    MPADecodeHeader  mpah;
    int srate_idx, i, channels;
Anton Khirnov's avatar
Anton Khirnov committed
140
    int bitrate_idx;
141
    int best_bitrate_idx = -1;
142
    int best_bitrate_error = INT_MAX;
143
    int ret;
144
    int ver = 0;
145
    int bytes_needed;
146

147
    if (!(s->pb->seekable & AVIO_SEEKABLE_NORMAL) || !mp3->write_xing)
148
        return 0;
149

150 151
    for (i = 0; i < FF_ARRAY_ELEMS(avpriv_mpa_freq_tab); i++) {
        const uint16_t base_freq = avpriv_mpa_freq_tab[i];
152

153 154 155
        if      (par->sample_rate == base_freq)     ver = 0x3; // MPEG 1
        else if (par->sample_rate == base_freq / 2) ver = 0x2; // MPEG 2
        else if (par->sample_rate == base_freq / 4) ver = 0x0; // MPEG 2.5
156
        else continue;
157

158 159 160
        srate_idx = i;
        break;
    }
161
    if (i == FF_ARRAY_ELEMS(avpriv_mpa_freq_tab)) {
162
        av_log(s, AV_LOG_WARNING, "Unsupported sample rate, not writing Xing header.\n");
163 164 165
        return -1;
    }

166
    switch (par->channels) {
167 168
    case 1:  channels = MPA_MONO;                                          break;
    case 2:  channels = MPA_STEREO;                                        break;
169 170
    default: av_log(s, AV_LOG_WARNING, "Unsupported number of channels, "
                    "not writing Xing header.\n");
171
             return -1;
172 173 174
    }

    /* dummy MPEG audio header */
175
    header  =  0xffU                                 << 24; // sync
176
    header |= (0x7 << 5 | ver << 3 | 0x1 << 1 | 0x1) << 16; // sync/audio-version/layer 3/no crc*/
177
    header |= (srate_idx << 2) << 8;
178 179
    header |= channels << 6;

180
    for (bitrate_idx = 1; bitrate_idx < 15; bitrate_idx++) {
181
        int bit_rate = 1000 * avpriv_mpa_bitrate_tab[ver != 3][3 - 1][bitrate_idx];
182
        int error    = FFABS(bit_rate - par->bit_rate);
183

184
        if (error < best_bitrate_error) {
185 186
            best_bitrate_error = error;
            best_bitrate_idx   = bitrate_idx;
187 188
        }
    }
189
    av_assert0(best_bitrate_idx >= 0);
190

191
    for (bitrate_idx = best_bitrate_idx; ; bitrate_idx++) {
192
        int32_t mask = bitrate_idx << (4 + 8);
193 194
        if (15 == bitrate_idx)
            return -1;
195
        header |= mask;
196

197 198
        ret = avpriv_mpegaudio_decode_header(&mpah, header);
        av_assert0(ret >= 0);
199 200
        mp3->xing_offset = xing_offtbl[mpah.lsf == 1][mpah.nb_channels == 1] + 4;
        bytes_needed     = mp3->xing_offset + XING_SIZE;
201

202
        if (bytes_needed <= mpah.frame_size)
203
            break;
204

205
        header &= ~mask;
206 207
    }

208 209
    ret = avio_open_dyn_buf(&dyn_ctx);
    if (ret < 0)
210
        return ret;
211 212

    avio_wb32(dyn_ctx, header);
Anton Khirnov's avatar
Anton Khirnov committed
213

214 215 216
    ffio_fill(dyn_ctx, 0, mp3->xing_offset - 4);
    ffio_wfourcc(dyn_ctx, "Xing");
    avio_wb32(dyn_ctx, 0x01 | 0x02 | 0x04 | 0x08);  // frames / size / TOC / vbr scale
217

Anton Khirnov's avatar
Anton Khirnov committed
218
    mp3->size = mpah.frame_size;
219 220 221
    mp3->want=1;
    mp3->seen=0;
    mp3->pos=0;
222

223 224
    avio_wb32(dyn_ctx, 0);  // frames
    avio_wb32(dyn_ctx, 0);  // size
Anton Khirnov's avatar
Anton Khirnov committed
225 226 227

    // TOC
    for (i = 0; i < XING_TOC_SIZE; i++)
228
        avio_w8(dyn_ctx, (uint8_t)(255 * i / XING_TOC_SIZE));
229 230 231 232 233 234 235 236

    // vbr quality
    // we write it, because some (broken) tools always expect it to be present
    avio_wb32(dyn_ctx, 0);

    // encoder short version string
    if (enc) {
        uint8_t encoder_str[9] = { 0 };
237 238 239 240 241 242
        if (   strlen(enc->value) > sizeof(encoder_str)
            && !strcmp("Lavc libmp3lame", enc->value)) {
            memcpy(encoder_str, "Lavf lame", 9);
        } else
            memcpy(encoder_str, enc->value, FFMIN(strlen(enc->value), sizeof(encoder_str)));

243 244
        avio_write(dyn_ctx, encoder_str, sizeof(encoder_str));
    } else
245
        avio_write(dyn_ctx, "Lavf\0\0\0\0\0", 9);
246 247 248 249 250 251

    avio_w8(dyn_ctx, 0);      // tag revision 0 / unknown vbr method
    avio_w8(dyn_ctx, 0);      // unknown lowpass filter value
    ffio_fill(dyn_ctx, 0, 8); // empty replaygain fields
    avio_w8(dyn_ctx, 0);      // unknown encoding flags
    avio_w8(dyn_ctx, 0);      // unknown abr/minimal bitrate
252
    avio_wb24(dyn_ctx, 0);    // empty encoder delay/padding
253 254 255 256 257 258 259 260 261

    avio_w8(dyn_ctx,   0); // misc
    avio_w8(dyn_ctx,   0); // mp3gain
    avio_wb16(dyn_ctx, 0); // preset

    // audio length and CRCs (will be updated later)
    avio_wb32(dyn_ctx, 0); // music length
    avio_wb16(dyn_ctx, 0); // music crc
    avio_wb16(dyn_ctx, 0); // tag crc
262

263
    ffio_fill(dyn_ctx, 0, mpah.frame_size - bytes_needed);
264

265 266 267
    mp3->xing_frame_size   = avio_close_dyn_buf(dyn_ctx, &mp3->xing_frame);
    mp3->xing_frame_offset = avio_tell(s->pb);
    avio_write(s->pb, mp3->xing_frame, mp3->xing_frame_size);
268

269
    mp3->audio_size = mp3->xing_frame_size;
270 271 272 273 274 275 276 277

    return 0;
}

/*
 * Add a frame to XING data.
 * Following lame's "VbrTag.c".
 */
Anton Khirnov's avatar
Anton Khirnov committed
278
static void mp3_xing_add_frame(MP3Context *mp3, AVPacket *pkt)
279 280 281
{
    int i;

Anton Khirnov's avatar
Anton Khirnov committed
282 283
    mp3->frames++;
    mp3->seen++;
284
    mp3->size += pkt->size;
285

Anton Khirnov's avatar
Anton Khirnov committed
286
    if (mp3->want == mp3->seen) {
287
        mp3->bag[mp3->pos] = mp3->size;
288

Anton Khirnov's avatar
Anton Khirnov committed
289
        if (XING_NUM_BAGS == ++mp3->pos) {
290
            /* shrink table to half size by throwing away each second bag. */
Anton Khirnov's avatar
Anton Khirnov committed
291
            for (i = 1; i < XING_NUM_BAGS; i += 2)
292
                mp3->bag[i >> 1] = mp3->bag[i];
293 294

            /* double wanted amount per bag. */
Anton Khirnov's avatar
Anton Khirnov committed
295
            mp3->want *= 2;
296
            /* adjust current position to half of table size. */
Anton Khirnov's avatar
Anton Khirnov committed
297
            mp3->pos = XING_NUM_BAGS / 2;
298 299
        }

300
        mp3->seen = 0;
301 302 303
    }
}

Anton Khirnov's avatar
Anton Khirnov committed
304
static int mp3_write_audio_packet(AVFormatContext *s, AVPacket *pkt)
305
{
Anton Khirnov's avatar
Anton Khirnov committed
306 307
    MP3Context  *mp3 = s->priv_data;

308
    if (pkt->data && pkt->size >= 4) {
309
        MPADecodeHeader mpah;
310
        int ret;
311
        int av_unused base;
312 313 314
        uint32_t h;

        h = AV_RB32(pkt->data);
315
        ret = avpriv_mpegaudio_decode_header(&mpah, h);
316
        if (ret >= 0) {
317
            if (!mp3->initial_bitrate)
318 319
                mp3->initial_bitrate = mpah.bit_rate;
            if ((mpah.bit_rate == 0) || (mp3->initial_bitrate != mpah.bit_rate))
320
                mp3->has_variable_bitrate = 1;
321
        } else {
322
            av_log(s, AV_LOG_WARNING, "Audio packet of size %d (starting with %08"PRIX32"...) "
323
                   "is invalid, writing it anyway.\n", pkt->size, h);
324
        }
325

326
#ifdef FILTER_VBR_HEADERS
327
        /* filter out XING and INFO headers. */
328
        base = 4 + xing_offtbl[mpah.lsf == 1][mpah.nb_channels == 1];
329 330 331 332

        if (base + 4 <= pkt->size) {
            uint32_t v = AV_RB32(pkt->data + base);

333
            if (MKBETAG('X','i','n','g') == v || MKBETAG('I','n','f','o') == v)
334 335 336 337 338 339
                return 0;
        }

        /* filter out VBRI headers. */
        base = 4 + 32;

340
        if (base + 4 <= pkt->size && MKBETAG('V','B','R','I') == AV_RB32(pkt->data + base))
341 342 343
            return 0;
#endif

344
        if (mp3->xing_offset) {
345 346 347
            uint8_t *side_data = NULL;
            int side_data_size = 0;

348
            mp3_xing_add_frame(mp3, pkt);
349 350 351
            mp3->audio_size += pkt->size;
            mp3->audio_crc   = av_crc(av_crc_get_table(AV_CRC_16_ANSI_LE),
                                      mp3->audio_crc, pkt->data, pkt->size);
352 353 354 355 356 357 358 359 360 361 362

            side_data = av_packet_get_side_data(pkt,
                                                AV_PKT_DATA_SKIP_SAMPLES,
                                                &side_data_size);
            if (side_data && side_data_size >= 10) {
                mp3->padding = FFMAX(AV_RL32(side_data + 4) + 528 + 1, 0);
                if (!mp3->delay)
                    mp3->delay =  FFMAX(AV_RL32(side_data) - 528 - 1, 0);
            } else {
                mp3->padding = 0;
            }
363
        }
364
    }
Anton Khirnov's avatar
Anton Khirnov committed
365 366

    return ff_raw_write_packet(s, pkt);
367 368
}

369 370 371
static int mp3_queue_flush(AVFormatContext *s)
{
    MP3Context *mp3 = s->priv_data;
372
    AVPacket pkt;
373 374
    int ret = 0, write = 1;

375
    ff_id3v2_finish(&mp3->id3, s->pb, s->metadata_header_padding);
376 377
    mp3_write_xing(s);

378 379 380
    while (mp3->queue) {
        ff_packet_list_get(&mp3->queue, &mp3->queue_end, &pkt);
        if (write && (ret = mp3_write_audio_packet(s, &pkt)) < 0)
381
            write = 0;
382
        av_packet_unref(&pkt);
383 384 385 386
    }
    return ret;
}

387 388 389
static void mp3_update_xing(AVFormatContext *s)
{
    MP3Context  *mp3 = s->priv_data;
390 391 392 393
    AVReplayGain *rg;
    uint16_t tag_crc;
    uint8_t *toc;
    int i, rg_size;
394
    int64_t old_pos = avio_tell(s->pb);
395 396

    /* replace "Xing" identification string with "Info" for CBR files. */
397 398
    if (!mp3->has_variable_bitrate)
        AV_WL32(mp3->xing_frame + mp3->xing_offset, MKTAG('I', 'n', 'f', 'o'));
399

400 401
    AV_WB32(mp3->xing_frame + mp3->xing_offset + 8,  mp3->frames);
    AV_WB32(mp3->xing_frame + mp3->xing_offset + 12, mp3->size);
402

403 404
    toc    = mp3->xing_frame + mp3->xing_offset + 16;
    toc[0] = 0;  // first toc entry has to be zero.
405 406 407
    for (i = 1; i < XING_TOC_SIZE; ++i) {
        int j = i * mp3->pos / XING_TOC_SIZE;
        int seek_point = 256LL * mp3->bag[j] / mp3->size;
408
        toc[i] = FFMIN(seek_point, 255);
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
    /* write replaygain */
    rg = (AVReplayGain*)av_stream_get_side_data(s->streams[0], AV_PKT_DATA_REPLAYGAIN,
                                                &rg_size);
    if (rg && rg_size >= sizeof(*rg)) {
        uint16_t val;

        AV_WB32(mp3->xing_frame + mp3->xing_offset + 131,
                av_rescale(rg->track_peak, 1 << 23, 100000));

        if (rg->track_gain != INT32_MIN) {
            val  = FFABS(rg->track_gain / 10000) & ((1 << 9) - 1);
            val |= (rg->track_gain < 0) << 9;
            val |= 1 << 13;
            AV_WB16(mp3->xing_frame + mp3->xing_offset + 135, val);
        }

        if (rg->album_gain != INT32_MIN) {
            val  = FFABS(rg->album_gain / 10000) & ((1 << 9) - 1);
            val |= (rg->album_gain < 0) << 9;
            val |= 1 << 14;
            AV_WB16(mp3->xing_frame + mp3->xing_offset + 137, val);
        }
    }

435 436 437 438 439 440 441 442 443 444 445
    /* write encoder delay/padding */
    if (mp3->delay >= 1 << 12) {
        mp3->delay = (1 << 12) - 1;
        av_log(s, AV_LOG_WARNING, "Too many samples of initial padding.\n");
    }
    if (mp3->padding >= 1 << 12) {
        mp3->padding = (1 << 12) - 1;
        av_log(s, AV_LOG_WARNING, "Too many samples of trailing padding.\n");
    }
    AV_WB24(mp3->xing_frame + mp3->xing_offset + 141, (mp3->delay << 12) + mp3->padding);

446 447 448 449 450 451 452 453
    AV_WB32(mp3->xing_frame + mp3->xing_offset + XING_SIZE - 8, mp3->audio_size);
    AV_WB16(mp3->xing_frame + mp3->xing_offset + XING_SIZE - 4, mp3->audio_crc);

    tag_crc = av_crc(av_crc_get_table(AV_CRC_16_ANSI_LE), 0, mp3->xing_frame, 190);
    AV_WB16(mp3->xing_frame + mp3->xing_offset + XING_SIZE - 2, tag_crc);

    avio_seek(s->pb,  mp3->xing_frame_offset, SEEK_SET);
    avio_write(s->pb, mp3->xing_frame, mp3->xing_frame_size);
454
    avio_seek(s->pb, old_pos, SEEK_SET);
455 456
}

457
static int mp3_write_trailer(struct AVFormatContext *s)
458 459
{
    uint8_t buf[ID3v1_TAG_SIZE];
460
    MP3Context *mp3 = s->priv_data;
461

Anton Khirnov's avatar
Anton Khirnov committed
462
    if (mp3->pics_to_write) {
463 464 465 466 467
        av_log(s, AV_LOG_WARNING, "No packets were sent for some of the "
               "attached pictures.\n");
        mp3_queue_flush(s);
    }

468
    /* write the id3v1 tag */
Anton Khirnov's avatar
Anton Khirnov committed
469
    if (mp3->write_id3v1 && id3v1_create_tag(s, buf) > 0) {
470
        avio_write(s->pb, buf, ID3v1_TAG_SIZE);
471
    }
472

473 474
    if (mp3->xing_offset)
        mp3_update_xing(s);
475

476 477
    av_freep(&mp3->xing_frame);

478 479 480
    return 0;
}

481
static int query_codec(enum AVCodecID id, int std_compliance)
482
{
483
    const CodecMime *cm= ff_id3v2_mime_tags;
484
    while(cm->id != AV_CODEC_ID_NONE) {
485 486 487 488 489 490 491
        if(id == cm->id)
            return MKTAG('A', 'P', 'I', 'C');
        cm++;
    }
    return -1;
}

492 493
static const AVOption options[] = {
    { "id3v2_version", "Select ID3v2 version to write. Currently 3 and 4 are supported.",
494
      offsetof(MP3Context, id3v2_version), AV_OPT_TYPE_INT, {.i64 = 4}, 0, 4, AV_OPT_FLAG_ENCODING_PARAM},
495
    { "write_id3v1", "Enable ID3v1 writing. ID3v1 tags are written in UTF-8 which may not be supported by most software.",
496
      offsetof(MP3Context, write_id3v1), AV_OPT_TYPE_BOOL, {.i64 = 0}, 0, 1, AV_OPT_FLAG_ENCODING_PARAM},
497
    { "write_xing",  "Write the Xing header containing file duration.",
498
      offsetof(MP3Context, write_xing),  AV_OPT_TYPE_BOOL, {.i64 = 1}, 0, 1, AV_OPT_FLAG_ENCODING_PARAM},
499 500 501 502
    { NULL },
};

static const AVClass mp3_muxer_class = {
503 504 505 506
    .class_name     = "MP3 muxer",
    .item_name      = av_default_item_name,
    .option         = options,
    .version        = LIBAVUTIL_VERSION_INT,
507 508
};

509 510 511 512 513 514 515
static int mp3_write_packet(AVFormatContext *s, AVPacket *pkt)
{
    MP3Context *mp3 = s->priv_data;

    if (pkt->stream_index == mp3->audio_stream_idx) {
        if (mp3->pics_to_write) {
            /* buffer audio packets until we get all the pictures */
516
            int ret = ff_packet_list_put(&mp3->queue, &mp3->queue_end, pkt, FF_PACKETLIST_FLAG_REF_PACKET);
517

518
            if (ret < 0) {
519 520 521 522 523
                av_log(s, AV_LOG_WARNING, "Not enough memory to buffer audio. Skipping picture streams\n");
                mp3->pics_to_write = 0;
                mp3_queue_flush(s);
                return mp3_write_audio_packet(s, pkt);
            }
524
        } else
Anton Khirnov's avatar
Anton Khirnov committed
525
            return mp3_write_audio_packet(s, pkt);
526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548
    } else {
        int ret;

        /* warn only once for each stream */
        if (s->streams[pkt->stream_index]->nb_frames == 1) {
            av_log(s, AV_LOG_WARNING, "Got more than one picture in stream %d,"
                   " ignoring.\n", pkt->stream_index);
        }
        if (!mp3->pics_to_write || s->streams[pkt->stream_index]->nb_frames >= 1)
            return 0;

        if ((ret = ff_id3v2_write_apic(s, &mp3->id3, pkt)) < 0)
            return ret;
        mp3->pics_to_write--;

        /* flush the buffered audio packets */
        if (!mp3->pics_to_write &&
            (ret = mp3_queue_flush(s)) < 0)
            return ret;
    }

    return 0;
}
549

Andreas Öman's avatar
Andreas Öman committed
550
/**
551
 * Write an ID3v2 header at beginning of stream
Andreas Öman's avatar
Andreas Öman committed
552 553
 */

554 555
static int mp3_write_header(struct AVFormatContext *s)
{
556
    MP3Context  *mp3 = s->priv_data;
557
    int ret, i;
558

559 560 561 562 563 564 565 566
    if (mp3->id3v2_version      &&
        mp3->id3v2_version != 3 &&
        mp3->id3v2_version != 4) {
        av_log(s, AV_LOG_ERROR, "Invalid ID3v2 version requested: %d. Only "
               "3, 4 or 0 (disabled) are allowed.\n", mp3->id3v2_version);
        return AVERROR(EINVAL);
    }

567 568 569 570 571
    /* check the streams -- we want exactly one audio and arbitrary number of
     * video (attached pictures) */
    mp3->audio_stream_idx = -1;
    for (i = 0; i < s->nb_streams; i++) {
        AVStream *st = s->streams[i];
572 573
        if (st->codecpar->codec_type == AVMEDIA_TYPE_AUDIO) {
            if (mp3->audio_stream_idx >= 0 || st->codecpar->codec_id != AV_CODEC_ID_MP3) {
574 575 576 577 578
                av_log(s, AV_LOG_ERROR, "Invalid audio stream. Exactly one MP3 "
                       "audio stream is required.\n");
                return AVERROR(EINVAL);
            }
            mp3->audio_stream_idx = i;
579
        } else if (st->codecpar->codec_type != AVMEDIA_TYPE_VIDEO) {
580 581 582 583 584 585 586 587 588 589
            av_log(s, AV_LOG_ERROR, "Only audio streams and pictures are allowed in MP3.\n");
            return AVERROR(EINVAL);
        }
    }
    if (mp3->audio_stream_idx < 0) {
        av_log(s, AV_LOG_ERROR, "No audio stream present.\n");
        return AVERROR(EINVAL);
    }
    mp3->pics_to_write = s->nb_streams - 1;

590 591 592 593 594 595 596 597 598 599 600 601
    if (mp3->pics_to_write && !mp3->id3v2_version) {
        av_log(s, AV_LOG_ERROR, "Attached pictures were requested, but the "
               "ID3v2 header is disabled.\n");
        return AVERROR(EINVAL);
    }

    if (mp3->id3v2_version) {
        ff_id3v2_start(&mp3->id3, s->pb, mp3->id3v2_version, ID3v2_DEFAULT_MAGIC);
        ret = ff_id3v2_write_metadata(s, &mp3->id3);
        if (ret < 0)
            return ret;
    }
Andreas Öman's avatar
Andreas Öman committed
602

603
    if (!mp3->pics_to_write) {
604
        if (mp3->id3v2_version)
605
            ff_id3v2_finish(&mp3->id3, s->pb, s->metadata_header_padding);
606
        mp3_write_xing(s);
607
    }
608

609 610 611
    return 0;
}

612
AVOutputFormat ff_mp3_muxer = {
613
    .name              = "mp3",
614
    .long_name         = NULL_IF_CONFIG_SMALL("MP3 (MPEG audio layer 3)"),
615
    .mime_type         = "audio/mpeg",
616 617
    .extensions        = "mp3",
    .priv_data_size    = sizeof(MP3Context),
618 619
    .audio_codec       = AV_CODEC_ID_MP3,
    .video_codec       = AV_CODEC_ID_PNG,
620
    .write_header      = mp3_write_header,
621
    .write_packet      = mp3_write_packet,
622
    .write_trailer     = mp3_write_trailer,
623
    .query_codec       = query_codec,
624
    .flags             = AVFMT_NOTIMESTAMPS,
625
    .priv_class        = &mp3_muxer_class,
626
};