flvdec.c 16.4 KB
Newer Older
1
/*
2
 * FLV demuxer
3
 * Copyright (c) 2003 The FFmpeg Project
4
 *
5 6 7 8 9
 * This demuxer will generate a 1 byte extradata for VP6F content.
 * It is composed of:
 *  - upper 4bits: difference between encoded width and visible width
 *  - lower 4bits: difference between encoded height and visible height
 *
10 11 12
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
13 14
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
15
 * version 2.1 of the License, or (at your option) any later version.
16
 *
17
 * FFmpeg is distributed in the hope that it will be useful,
18 19 20 21 22
 * 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
23
 * License along with FFmpeg; if not, write to the Free Software
24
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
25 26
 */
#include "avformat.h"
27
#include "flv.h"
28

Baptiste Coudurier's avatar
Baptiste Coudurier committed
29 30 31 32
typedef struct {
    int wrong_dts; ///< wrong dts due to negative cts
} FLVContext;

33 34 35 36 37
static int flv_probe(AVProbeData *p)
{
    const uint8_t *d;

    d = p->buf;
38 39
    if (d[0] == 'F' && d[1] == 'L' && d[2] == 'V' && d[3] < 5 && d[5]==0) {
        return AVPROBE_SCORE_MAX;
40 41 42 43
    }
    return 0;
}

Allan Hsu's avatar
Allan Hsu committed
44 45 46 47
static void flv_set_audio_codec(AVFormatContext *s, AVStream *astream, int flv_codecid) {
    AVCodecContext *acodec = astream->codec;
    switch(flv_codecid) {
        //no distinction between S16 and S8 PCM codec flags
48
        case FLV_CODECID_PCM:
49
            acodec->codec_id = acodec->bits_per_coded_sample == 8 ? CODEC_ID_PCM_S8 :
Michael Niedermayer's avatar
Michael Niedermayer committed
50 51 52 53 54 55
#ifdef WORDS_BIGENDIAN
                                CODEC_ID_PCM_S16BE;
#else
                                CODEC_ID_PCM_S16LE;
#endif
            break;
Allan Hsu's avatar
Allan Hsu committed
56
        case FLV_CODECID_PCM_LE:
57
            acodec->codec_id = acodec->bits_per_coded_sample == 8 ? CODEC_ID_PCM_S8 : CODEC_ID_PCM_S16LE; break;
58
        case FLV_CODECID_AAC  : acodec->codec_id = CODEC_ID_AAC;                                    break;
Allan Hsu's avatar
Allan Hsu committed
59
        case FLV_CODECID_ADPCM: acodec->codec_id = CODEC_ID_ADPCM_SWF;                              break;
60 61 62 63
        case FLV_CODECID_SPEEX:
            acodec->codec_id = CODEC_ID_SPEEX;
            acodec->sample_rate = 16000;
            break;
Aurelien Jacobs's avatar
Aurelien Jacobs committed
64
        case FLV_CODECID_MP3  : acodec->codec_id = CODEC_ID_MP3      ; astream->need_parsing = AVSTREAM_PARSE_FULL; break;
65
        case FLV_CODECID_NELLYMOSER_8KHZ_MONO:
Allan Hsu's avatar
Allan Hsu committed
66 67
            acodec->sample_rate = 8000; //in case metadata does not otherwise declare samplerate
        case FLV_CODECID_NELLYMOSER:
Benjamin Larsson's avatar
Benjamin Larsson committed
68 69
            acodec->codec_id = CODEC_ID_NELLYMOSER;
            break;
Allan Hsu's avatar
Allan Hsu committed
70 71 72 73 74 75 76 77 78 79 80 81
        default:
            av_log(s, AV_LOG_INFO, "Unsupported audio codec (%x)\n", flv_codecid >> FLV_AUDIO_CODECID_OFFSET);
            acodec->codec_tag = flv_codecid >> FLV_AUDIO_CODECID_OFFSET;
    }
}

static int flv_set_video_codec(AVFormatContext *s, AVStream *vstream, int flv_codecid) {
    AVCodecContext *vcodec = vstream->codec;
    switch(flv_codecid) {
        case FLV_CODECID_H263  : vcodec->codec_id = CODEC_ID_FLV1   ; break;
        case FLV_CODECID_SCREEN: vcodec->codec_id = CODEC_ID_FLASHSV; break;
        case FLV_CODECID_VP6   : vcodec->codec_id = CODEC_ID_VP6F   ;
Aurelien Jacobs's avatar
Aurelien Jacobs committed
82 83 84
        case FLV_CODECID_VP6A  :
            if(flv_codecid == FLV_CODECID_VP6A)
                vcodec->codec_id = CODEC_ID_VP6A;
Allan Hsu's avatar
Allan Hsu committed
85 86 87 88
            if(vcodec->extradata_size != 1) {
                vcodec->extradata_size = 1;
                vcodec->extradata = av_malloc(1);
            }
89
            vcodec->extradata[0] = get_byte(s->pb);
Allan Hsu's avatar
Allan Hsu committed
90
            return 1; // 1 byte body size adjustment for flv_read_packet()
91 92 93
        case FLV_CODECID_H264:
            vcodec->codec_id = CODEC_ID_H264;
            return 3; // not 4, reading packet type will consume one byte
Allan Hsu's avatar
Allan Hsu committed
94 95 96 97 98 99 100 101
        default:
            av_log(s, AV_LOG_INFO, "Unsupported video codec (%x)\n", flv_codecid);
            vcodec->codec_tag = flv_codecid;
    }

    return 0;
}

102
static int amf_get_string(ByteIOContext *ioc, char *buffer, int buffsize) {
Michael Niedermayer's avatar
Michael Niedermayer committed
103
    int length = get_be16(ioc);
104 105
    if(length >= buffsize) {
        url_fskip(ioc, length);
Michael Niedermayer's avatar
Michael Niedermayer committed
106
        return -1;
107 108 109 110 111 112 113 114 115
    }

    get_buffer(ioc, buffer, length);

    buffer[length] = '\0';

    return length;
}

Pascal Massimino's avatar
Pascal Massimino committed
116
static int amf_parse_object(AVFormatContext *s, AVStream *astream, AVStream *vstream, const char *key, int64_t max_pos, int depth) {
Allan Hsu's avatar
Allan Hsu committed
117 118 119 120 121 122 123
    AVCodecContext *acodec, *vcodec;
    ByteIOContext *ioc;
    AMFDataType amf_type;
    char str_val[256];
    double num_val;

    num_val = 0;
124
    ioc = s->pb;
Allan Hsu's avatar
Allan Hsu committed
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

    amf_type = get_byte(ioc);

    switch(amf_type) {
        case AMF_DATA_TYPE_NUMBER:
            num_val = av_int2dbl(get_be64(ioc)); break;
        case AMF_DATA_TYPE_BOOL:
            num_val = get_byte(ioc); break;
        case AMF_DATA_TYPE_STRING:
            if(amf_get_string(ioc, str_val, sizeof(str_val)) < 0)
                return -1;
            break;
        case AMF_DATA_TYPE_OBJECT: {
            unsigned int keylen;

            while(url_ftell(ioc) < max_pos - 2 && (keylen = get_be16(ioc))) {
                url_fskip(ioc, keylen); //skip key string
                if(amf_parse_object(s, NULL, NULL, NULL, max_pos, depth + 1) < 0)
                    return -1; //if we couldn't skip, bomb out.
            }
            if(get_byte(ioc) != AMF_END_OF_OBJECT)
                return -1;
        }
            break;
        case AMF_DATA_TYPE_NULL:
        case AMF_DATA_TYPE_UNDEFINED:
        case AMF_DATA_TYPE_UNSUPPORTED:
            break; //these take up no additional space
        case AMF_DATA_TYPE_MIXEDARRAY:
            url_fskip(ioc, 4); //skip 32-bit max array index
            while(url_ftell(ioc) < max_pos - 2 && amf_get_string(ioc, str_val, sizeof(str_val)) > 0) {
                //this is the only case in which we would want a nested parse to not skip over the object
                if(amf_parse_object(s, astream, vstream, str_val, max_pos, depth + 1) < 0)
                    return -1;
            }
            if(get_byte(ioc) != AMF_END_OF_OBJECT)
                return -1;
            break;
        case AMF_DATA_TYPE_ARRAY: {
            unsigned int arraylen, i;

            arraylen = get_be32(ioc);
            for(i = 0; i < arraylen && url_ftell(ioc) < max_pos - 1; i++) {
                if(amf_parse_object(s, NULL, NULL, NULL, max_pos, depth + 1) < 0)
                    return -1; //if we couldn't skip, bomb out.
            }
        }
            break;
        case AMF_DATA_TYPE_DATE:
            url_fskip(ioc, 8 + 2); //timestamp (double) and UTC offset (int16)
            break;
        default: //unsupported type, we couldn't skip
            return -1;
    }

    if(depth == 1 && key) { //only look for metadata values when we are not nested and key != NULL
        acodec = astream ? astream->codec : NULL;
        vcodec = vstream ? vstream->codec : NULL;

        if(amf_type == AMF_DATA_TYPE_BOOL) {
            if(!strcmp(key, "stereo") && acodec) acodec->channels = num_val > 0 ? 2 : 1;
        } else if(amf_type == AMF_DATA_TYPE_NUMBER) {
            if(!strcmp(key, "duration")) s->duration = num_val * AV_TIME_BASE;
188 189
//            else if(!strcmp(key, "width")  && vcodec && num_val > 0) vcodec->width  = num_val;
//            else if(!strcmp(key, "height") && vcodec && num_val > 0) vcodec->height = num_val;
190 191
            else if(!strcmp(key, "videodatarate") && vcodec && 0 <= (int)(num_val * 1024.0))
                vcodec->bit_rate = num_val * 1024.0;
192 193 194 195 196
            else if(!strcmp(key, "audiocodecid") && acodec && 0 <= (int)num_val)
                flv_set_audio_codec(s, astream, (int)num_val << FLV_AUDIO_CODECID_OFFSET);
            else if(!strcmp(key, "videocodecid") && vcodec && 0 <= (int)num_val)
                flv_set_video_codec(s, vstream, (int)num_val);
            else if(!strcmp(key, "audiosamplesize") && acodec && 0 < (int)num_val) {
197
                acodec->bits_per_coded_sample = num_val;
Allan Hsu's avatar
Allan Hsu committed
198 199 200 201 202 203
                //we may have to rewrite a previously read codecid because FLV only marks PCM endianness.
                if(num_val == 8 && (acodec->codec_id == CODEC_ID_PCM_S16BE || acodec->codec_id == CODEC_ID_PCM_S16LE))
                    acodec->codec_id = CODEC_ID_PCM_S8;
            }
            else if(!strcmp(key, "audiosamplerate") && acodec && num_val >= 0) {
                //some tools, like FLVTool2, write consistently approximate metadata sample rates
204
                if (!acodec->sample_rate) {
Benjamin Larsson's avatar
Benjamin Larsson committed
205 206 207 208 209 210 211
                    switch((int)num_val) {
                        case 44000: acodec->sample_rate = 44100  ; break;
                        case 22000: acodec->sample_rate = 22050  ; break;
                        case 11000: acodec->sample_rate = 11025  ; break;
                        case 5000 : acodec->sample_rate = 5512   ; break;
                        default   : acodec->sample_rate = num_val;
                    }
212
                }
Allan Hsu's avatar
Allan Hsu committed
213 214 215 216 217 218 219
            }
        }
    }

    return 0;
}

Pascal Massimino's avatar
Pascal Massimino committed
220
static int flv_read_metabody(AVFormatContext *s, int64_t next_pos) {
Allan Hsu's avatar
Allan Hsu committed
221 222 223 224 225 226 227 228 229
    AMFDataType type;
    AVStream *stream, *astream, *vstream;
    ByteIOContext *ioc;
    int i, keylen;
    char buffer[11]; //only needs to hold the string "onMetaData". Anything longer is something we don't want.

    astream = NULL;
    vstream = NULL;
    keylen = 0;
230
    ioc = s->pb;
Allan Hsu's avatar
Allan Hsu committed
231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250

    //first object needs to be "onMetaData" string
    type = get_byte(ioc);
    if(type != AMF_DATA_TYPE_STRING || amf_get_string(ioc, buffer, sizeof(buffer)) < 0 || strcmp(buffer, "onMetaData"))
        return -1;

    //find the streams now so that amf_parse_object doesn't need to do the lookup every time it is called.
    for(i = 0; i < s->nb_streams; i++) {
        stream = s->streams[i];
        if     (stream->codec->codec_type == CODEC_TYPE_AUDIO) astream = stream;
        else if(stream->codec->codec_type == CODEC_TYPE_VIDEO) vstream = stream;
    }

    //parse the second object (we want a mixed array)
    if(amf_parse_object(s, astream, vstream, buffer, next_pos, 0) < 0)
        return -1;

    return 0;
}

251 252 253 254 255
static AVStream *create_stream(AVFormatContext *s, int is_audio){
    AVStream *st = av_new_stream(s, is_audio);
    if (!st)
        return NULL;
    st->codec->codec_type = is_audio ? CODEC_TYPE_AUDIO : CODEC_TYPE_VIDEO;
Baptiste Coudurier's avatar
Baptiste Coudurier committed
256
    av_set_pts_info(st, 32, 1, 1000); /* 32 bit pts in ms */
257 258 259
    return st;
}

260 261 262
static int flv_read_header(AVFormatContext *s,
                           AVFormatParameters *ap)
{
263
    int offset, flags;
264

265 266
    url_fskip(s->pb, 4);
    flags = get_byte(s->pb);
267 268 269 270 271 272
    /* old flvtool cleared this field */
    /* FIXME: better fix needed */
    if (!flags) {
        flags = FLV_HEADER_FLAG_HASVIDEO | FLV_HEADER_FLAG_HASAUDIO;
        av_log(s, AV_LOG_WARNING, "Broken FLV file, which says no streams present, this might fail\n");
    }
273

Michael Niedermayer's avatar
Michael Niedermayer committed
274 275 276 277
    if((flags & (FLV_HEADER_FLAG_HASVIDEO|FLV_HEADER_FLAG_HASAUDIO))
             != (FLV_HEADER_FLAG_HASVIDEO|FLV_HEADER_FLAG_HASAUDIO))
        s->ctx_flags |= AVFMTCTX_NOHEADER;

278
    if(flags & FLV_HEADER_FLAG_HASVIDEO){
279
        if(!create_stream(s, 0))
280
            return AVERROR(ENOMEM);
281 282
    }
    if(flags & FLV_HEADER_FLAG_HASAUDIO){
283
        if(!create_stream(s, 1))
284
            return AVERROR(ENOMEM);
285 286
    }

287 288
    offset = get_be32(s->pb);
    url_fseek(s->pb, offset, SEEK_SET);
289

290 291
    s->start_time = 0;

292 293 294
    return 0;
}

295 296 297 298 299 300 301 302 303 304 305
static int flv_get_extradata(AVFormatContext *s, AVStream *st, int size)
{
    av_free(st->codec->extradata);
    st->codec->extradata = av_mallocz(size + FF_INPUT_BUFFER_PADDING_SIZE);
    if (!st->codec->extradata)
        return AVERROR(ENOMEM);
    st->codec->extradata_size = size;
    get_buffer(s->pb, st->codec->extradata, st->codec->extradata_size);
    return 0;
}

306 307
static int flv_read_packet(AVFormatContext *s, AVPacket *pkt)
{
Baptiste Coudurier's avatar
Baptiste Coudurier committed
308
    FLVContext *flv = s->priv_data;
Pascal Massimino's avatar
Pascal Massimino committed
309 310
    int ret, i, type, size, flags, is_audio;
    int64_t next, pos;
Baptiste Coudurier's avatar
Baptiste Coudurier committed
311
    int64_t dts, pts = AV_NOPTS_VALUE;
Alex Beregszaszi's avatar
Alex Beregszaszi committed
312
    AVStream *st = NULL;
313

Michael Niedermayer's avatar
cleanup  
Michael Niedermayer committed
314
 for(;;){
315 316 317 318
    pos = url_ftell(s->pb);
    url_fskip(s->pb, 4); /* size of previous packet */
    type = get_byte(s->pb);
    size = get_be24(s->pb);
319 320 321
    dts = get_be24(s->pb);
    dts |= get_byte(s->pb) << 24;
//    av_log(s, AV_LOG_DEBUG, "type:%d, size:%d, dts:%d\n", type, size, dts);
322
    if (url_feof(s->pb))
323
        return AVERROR(EIO);
324
    url_fskip(s->pb, 3); /* stream id, always 0 */
325
    flags = 0;
326

Michael Niedermayer's avatar
cleanup  
Michael Niedermayer committed
327 328
    if(size == 0)
        continue;
329

330
    next= size + url_ftell(s->pb);
331

332
    if (type == FLV_TAG_TYPE_AUDIO) {
Michael Niedermayer's avatar
cleanup  
Michael Niedermayer committed
333
        is_audio=1;
334
        flags = get_byte(s->pb);
335
        size--;
336
    } else if (type == FLV_TAG_TYPE_VIDEO) {
Michael Niedermayer's avatar
cleanup  
Michael Niedermayer committed
337
        is_audio=0;
338
        flags = get_byte(s->pb);
339
        size--;
340 341
        if ((flags & 0xf0) == 0x50) /* video info / command frame */
            goto skip;
342
    } else {
Allan Hsu's avatar
Allan Hsu committed
343 344 345 346
        if (type == FLV_TAG_TYPE_META && size > 13+1+4)
            flv_read_metabody(s, next);
        else /* skip packet */
            av_log(s, AV_LOG_ERROR, "skipping flv packet: type %d, size %d, flags %d\n", type, size, flags);
347
    skip:
348
        url_fseek(s->pb, next, SEEK_SET);
349
        return AVERROR(EAGAIN);
350 351
    }

352 353 354 355
    /* skip empty data packets */
    if (!size)
        continue;

356 357 358
    /* now find stream */
    for(i=0;i<s->nb_streams;i++) {
        st = s->streams[i];
Michael Niedermayer's avatar
cleanup  
Michael Niedermayer committed
359 360
        if (st->id == is_audio)
            break;
361
    }
Michael Niedermayer's avatar
cleanup  
Michael Niedermayer committed
362
    if(i == s->nb_streams){
363
        av_log(s, AV_LOG_ERROR, "invalid stream\n");
364
        st= create_stream(s, is_audio);
365
        s->ctx_flags &= ~AVFMTCTX_NOHEADER;
366
    }
367
//    av_log(s, AV_LOG_DEBUG, "%d %X %d \n", is_audio, flags, st->discard);
368 369
    if(  (st->discard >= AVDISCARD_NONKEY && !((flags & FLV_VIDEO_FRAMETYPE_MASK) == FLV_FRAME_KEY ||         is_audio))
       ||(st->discard >= AVDISCARD_BIDIR  &&  ((flags & FLV_VIDEO_FRAMETYPE_MASK) == FLV_FRAME_DISP_INTER && !is_audio))
370 371
       || st->discard >= AVDISCARD_ALL
       ){
372
        url_fseek(s->pb, next, SEEK_SET);
373
        return AVERROR(EAGAIN);
374
    }
375
    if ((flags & FLV_VIDEO_FRAMETYPE_MASK) == FLV_FRAME_KEY)
376
        av_add_index_entry(st, pos, dts, size, 0, AVINDEX_KEYFRAME);
Michael Niedermayer's avatar
cleanup  
Michael Niedermayer committed
377 378 379
    break;
 }

380
    // if not streamed and no duration from metadata then seek to end to find the duration from the timestamps
381
    if(!url_is_streamed(s->pb) && s->duration==AV_NOPTS_VALUE){
382
        int size;
Pascal Massimino's avatar
Pascal Massimino committed
383 384
        const int64_t pos= url_ftell(s->pb);
        const int64_t fsize= url_fsize(s->pb);
385 386 387 388 389
        url_fseek(s->pb, fsize-4, SEEK_SET);
        size= get_be32(s->pb);
        url_fseek(s->pb, fsize-3-size, SEEK_SET);
        if(size == get_be24(s->pb) + 11){
            s->duration= get_be24(s->pb) * (int64_t)AV_TIME_BASE / 1000;
390
        }
391
        url_fseek(s->pb, pos, SEEK_SET);
392 393
    }

Michael Niedermayer's avatar
cleanup  
Michael Niedermayer committed
394
    if(is_audio){
395
        if(!st->codec->channels || !st->codec->sample_rate || !st->codec->bits_per_coded_sample) {
396
            st->codec->channels = (flags & FLV_AUDIO_CHANNEL_MASK) == FLV_STEREO ? 2 : 1;
397
            st->codec->sample_rate = (44100 << ((flags & FLV_AUDIO_SAMPLERATE_MASK) >> FLV_AUDIO_SAMPLERATE_OFFSET) >> 3);
398
            st->codec->bits_per_coded_sample = (flags & FLV_AUDIO_SAMPLESIZE_MASK) ? 16 : 8;
399 400
        }
        if(!st->codec->codec_id){
Allan Hsu's avatar
Allan Hsu committed
401
            flv_set_audio_codec(s, st, flags & FLV_AUDIO_CODECID_MASK);
Michael Niedermayer's avatar
cleanup  
Michael Niedermayer committed
402 403
        }
    }else{
Allan Hsu's avatar
Allan Hsu committed
404
        size -= flv_set_video_codec(s, st, flags & FLV_VIDEO_CODECID_MASK);
Michael Niedermayer's avatar
Michael Niedermayer committed
405 406
    }

407 408 409 410 411
    if (st->codec->codec_id == CODEC_ID_AAC ||
        st->codec->codec_id == CODEC_ID_H264) {
        int type = get_byte(s->pb);
        size--;
        if (st->codec->codec_id == CODEC_ID_H264) {
Baptiste Coudurier's avatar
Baptiste Coudurier committed
412 413 414 415 416 417 418 419
            int32_t cts = (get_be24(s->pb)+0xff800000)^0xff800000; // sign extension
            pts = dts + cts;
            if (cts < 0) { // dts are wrong
                flv->wrong_dts = 1;
                av_log(s, AV_LOG_WARNING, "negative cts, previous timestamps might be wrong\n");
            }
            if (flv->wrong_dts)
                dts = AV_NOPTS_VALUE;
420 421
        }
        if (type == 0) {
422
            if ((ret = flv_get_extradata(s, st, size)) < 0)
423
                return ret;
424
            return AVERROR(EAGAIN);
425 426 427
        }
    }

428
    ret= av_get_packet(s->pb, pkt, size);
429
    if (ret <= 0) {
430
        return AVERROR(EIO);
431 432 433 434
    }
    /* note: we need to modify the packet size here to handle the last
       packet */
    pkt->size = ret;
435
    pkt->dts = dts;
Baptiste Coudurier's avatar
Baptiste Coudurier committed
436
    pkt->pts = pts == AV_NOPTS_VALUE ? dts : pts;
437
    pkt->stream_index = st->index;
438

439
    if (is_audio || ((flags & FLV_VIDEO_FRAMETYPE_MASK) == FLV_FRAME_KEY))
440
        pkt->flags |= PKT_FLAG_KEY;
441

442 443 444
    return ret;
}

445
AVInputFormat flv_demuxer = {
446
    "flv",
447
    NULL_IF_CONFIG_SMALL("FLV format"),
Baptiste Coudurier's avatar
Baptiste Coudurier committed
448
    sizeof(FLVContext),
449 450 451 452 453 454
    flv_probe,
    flv_read_header,
    flv_read_packet,
    .extensions = "flv",
    .value = CODEC_ID_FLV1,
};