mov.c

ffmpeg源码分析

    ctx.buf_end = ctx.buffer + moov_len;
    atom.type = MKTAG( 'm', 'o', 'o', 'v' );
    atom.offset = 0;
    atom.size = moov_len;
#ifdef DEBUG
//    { int fd = open("/tmp/uncompheader.mov", O_WRONLY | O_CREAT); write(fd, moov_data, moov_len); close(fd); }
#endif
    ret = mov_read_default(c, &ctx, atom);
    av_free(moov_data);
    av_free(cmov_data);

    return ret;
}
#endif

/* edit list atom */
static int mov_read_elst(MOVContext *c, ByteIOContext *pb, MOV_atom_t atom)
{
  int i, edit_count;

  get_byte(pb); /* version */
  get_byte(pb); get_byte(pb); get_byte(pb); /* flags */
  edit_count= c->streams[c->fc->nb_streams-1]->edit_count = get_be32(pb);     /* entries */

  for(i=0; i<edit_count; i++){
    get_be32(pb); /* Track duration */
    get_be32(pb); /* Media time */
    get_be32(pb); /* Media rate */
  }
  dprintf("track[%i].edit_count = %i\n", c->fc->nb_streams-1, c->streams[c->fc->nb_streams-1]->edit_count);
  return 0;
}

static const MOVParseTableEntry mov_default_parse_table[] = {
/* mp4 atoms */
{ MKTAG( 'c', 'o', '6', '4' ), mov_read_stco },
{ MKTAG( 'c', 'p', 'r', 't' ), mov_read_default },
{ MKTAG( 'c', 'r', 'h', 'd' ), mov_read_default },
{ MKTAG( 'c', 't', 't', 's' ), mov_read_ctts }, /* composition time to sample */
{ MKTAG( 'd', 'i', 'n', 'f' ), mov_read_default }, /* data information */
{ MKTAG( 'd', 'p', 'n', 'd' ), mov_read_leaf },
{ MKTAG( 'd', 'r', 'e', 'f' ), mov_read_leaf },
{ MKTAG( 'e', 'd', 't', 's' ), mov_read_default },
{ MKTAG( 'e', 'l', 's', 't' ), mov_read_elst },
{ MKTAG( 'e', 'n', 'd', 'a' ), mov_read_enda },
{ MKTAG( 'f', 'r', 'e', 'e' ), mov_read_leaf },
{ MKTAG( 'f', 't', 'y', 'p' ), mov_read_ftyp },
{ MKTAG( 'h', 'd', 'l', 'r' ), mov_read_hdlr },
{ MKTAG( 'h', 'i', 'n', 't' ), mov_read_leaf },
{ MKTAG( 'h', 'm', 'h', 'd' ), mov_read_leaf },
{ MKTAG( 'i', 'o', 'd', 's' ), mov_read_leaf },
{ MKTAG( 'j', 'p', '2', 'h' ), mov_read_jp2h },
{ MKTAG( 'm', 'd', 'a', 't' ), mov_read_mdat },
{ MKTAG( 'm', 'd', 'h', 'd' ), mov_read_mdhd },
{ MKTAG( 'm', 'd', 'i', 'a' ), mov_read_default },
{ MKTAG( 'm', 'i', 'n', 'f' ), mov_read_default },
{ MKTAG( 'm', 'o', 'o', 'v' ), mov_read_moov },
{ MKTAG( 'm', 'p', '4', 'a' ), mov_read_default },
{ MKTAG( 'm', 'p', '4', 's' ), mov_read_default },
{ MKTAG( 'm', 'p', '4', 'v' ), mov_read_default },
{ MKTAG( 'm', 'p', 'o', 'd' ), mov_read_leaf },
{ MKTAG( 'm', 'v', 'h', 'd' ), mov_read_mvhd },
{ MKTAG( 'n', 'm', 'h', 'd' ), mov_read_leaf },
{ MKTAG( 'o', 'd', 'h', 'd' ), mov_read_default },
{ MKTAG( 's', 'd', 'h', 'd' ), mov_read_default },
{ MKTAG( 's', 'k', 'i', 'p' ), mov_read_leaf },
{ MKTAG( 's', 'm', 'h', 'd' ), mov_read_leaf }, /* sound media info header */
{ MKTAG( 'S', 'M', 'I', ' ' ), mov_read_smi }, /* Sorenson extension ??? */
{ MKTAG( 'a', 'l', 'a', 'c' ), mov_read_alac }, /* alac specific atom */
{ MKTAG( 'a', 'v', 'c', 'C' ), mov_read_avcC },
{ MKTAG( 's', 't', 'b', 'l' ), mov_read_default },
{ MKTAG( 's', 't', 'c', 'o' ), mov_read_stco },
{ MKTAG( 's', 't', 'd', 'p' ), mov_read_default },
{ MKTAG( 's', 't', 's', 'c' ), mov_read_stsc },
{ MKTAG( 's', 't', 's', 'd' ), mov_read_stsd }, /* sample description */
{ MKTAG( 's', 't', 's', 'h' ), mov_read_default },
{ MKTAG( 's', 't', 's', 's' ), mov_read_stss }, /* sync sample */
{ MKTAG( 's', 't', 's', 'z' ), mov_read_stsz }, /* sample size */
{ MKTAG( 's', 't', 't', 's' ), mov_read_stts },
{ MKTAG( 't', 'k', 'h', 'd' ), mov_read_tkhd }, /* track header */
{ MKTAG( 't', 'r', 'a', 'k' ), mov_read_trak },
{ MKTAG( 't', 'r', 'e', 'f' ), mov_read_default }, /* not really */
{ MKTAG( 'u', 'd', 't', 'a' ), mov_read_leaf },
{ MKTAG( 'u', 'r', 'l', ' ' ), mov_read_leaf },
{ MKTAG( 'u', 'r', 'n', ' ' ), mov_read_leaf },
{ MKTAG( 'u', 'u', 'i', 'd' ), mov_read_leaf },
{ MKTAG( 'v', 'm', 'h', 'd' ), mov_read_leaf }, /* video media info header */
{ MKTAG( 'w', 'a', 'v', 'e' ), mov_read_wave },
/* extra mp4 */
{ MKTAG( 'M', 'D', 'E', 'S' ), mov_read_leaf },
/* QT atoms */
{ MKTAG( 'c', 'h', 'a', 'p' ), mov_read_leaf },
{ MKTAG( 'c', 'l', 'i', 'p' ), mov_read_default },
{ MKTAG( 'c', 'r', 'g', 'n' ), mov_read_leaf },
{ MKTAG( 'c', 't', 'a', 'b' ), mov_read_ctab },
{ MKTAG( 'e', 's', 'd', 's' ), mov_read_esds },
{ MKTAG( 'k', 'm', 'a', 't' ), mov_read_leaf },
{ MKTAG( 'm', 'a', 't', 't' ), mov_read_default },
{ MKTAG( 'r', 'd', 'r', 'f' ), mov_read_leaf },
{ MKTAG( 'r', 'm', 'd', 'a' ), mov_read_default },
{ MKTAG( 'r', 'm', 'd', 'r' ), mov_read_leaf },
{ MKTAG( 'r', 'm', 'r', 'a' ), mov_read_default },
{ MKTAG( 's', 'c', 'p', 't' ), mov_read_leaf },
{ MKTAG( 's', 's', 'r', 'c' ), mov_read_leaf },
{ MKTAG( 's', 'y', 'n', 'c' ), mov_read_leaf },
{ MKTAG( 't', 'c', 'm', 'd' ), mov_read_leaf },
{ MKTAG( 'w', 'i', 'd', 'e' ), mov_read_wide }, /* place holder */
//{ MKTAG( 'r', 'm', 'q', 'u' ), mov_read_leaf },
#ifdef CONFIG_ZLIB
{ MKTAG( 'c', 'm', 'o', 'v' ), mov_read_cmov },
#else
{ MKTAG( 'c', 'm', 'o', 'v' ), mov_read_leaf },
#endif
{ 0L, mov_read_leaf }
};

static void mov_free_stream_context(MOVStreamContext *sc)
{
    if(sc) {
        av_freep(&sc->chunk_offsets);
        av_freep(&sc->sample_to_chunk);
        av_freep(&sc->sample_sizes);
        av_freep(&sc->keyframes);
        av_freep(&sc->stts_data);
        av_freep(&sc->ctts_data);
        av_freep(&sc);
    }
}

static inline uint32_t mov_to_tag(uint8_t *buf)
{
    return MKTAG(buf[0], buf[1], buf[2], buf[3]);
}

static inline uint32_t to_be32(uint8_t *buf)
{
    return (buf[0] << 24) | (buf[1] << 16) | (buf[2] << 8) | buf[3];
}

/* XXX: is it sufficient ? */
static int mov_probe(AVProbeData *p)
{
    unsigned int offset;
    uint32_t tag;
    int score = 0;

    /* check file header */
    if (p->buf_size <= 12)
        return 0;
    offset = 0;
    for(;;) {
        /* ignore invalid offset */
        if ((offset + 8) > (unsigned int)p->buf_size)
            return score;
        tag = mov_to_tag(p->buf + offset + 4);
        switch(tag) {
        /* check for obvious tags */
        case MKTAG( 'j', 'P', ' ', ' ' ): /* jpeg 2000 signature */
        case MKTAG( 'm', 'o', 'o', 'v' ):
        case MKTAG( 'm', 'd', 'a', 't' ):
        case MKTAG( 'p', 'n', 'o', 't' ): /* detect movs with preview pics like ew.mov and april.mov */
        case MKTAG( 'u', 'd', 't', 'a' ): /* Packet Video PVAuthor adds this and a lot of more junk */
            return AVPROBE_SCORE_MAX;
        /* those are more common words, so rate then a bit less */
        case MKTAG( 'w', 'i', 'd', 'e' ):
        case MKTAG( 'f', 'r', 'e', 'e' ):
        case MKTAG( 'j', 'u', 'n', 'k' ):
        case MKTAG( 'p', 'i', 'c', 't' ):
            return AVPROBE_SCORE_MAX - 5;
        case MKTAG( 'f', 't', 'y', 'p' ):
        case MKTAG( 's', 'k', 'i', 'p' ):
        case MKTAG( 'u', 'u', 'i', 'd' ):
            offset = to_be32(p->buf+offset) + offset;
            /* if we only find those cause probedata is too small at least rate them */
            score = AVPROBE_SCORE_MAX - 50;
            break;
        default:
            /* unrecognized tag */
            return score;
        }
    }
    return score;
}

static int mov_read_header(AVFormatContext *s, AVFormatParameters *ap)
{
    MOVContext *mov = (MOVContext *) s->priv_data;
    ByteIOContext *pb = &s->pb;
    int i, j, nb, err;
    MOV_atom_t atom = { 0, 0, 0 };

    mov->fc = s;
    mov->parse_table = mov_default_parse_table;

    if(!url_is_streamed(pb)) /* .mov and .mp4 aren't streamable anyway (only progressive download if moov is before mdat) */
        atom.size = url_fsize(pb);
    else
        atom.size = 0x7FFFFFFFFFFFFFFFLL;

    /* check MOV header */
    err = mov_read_default(mov, pb, atom);
    if (err<0 || (!mov->found_moov && !mov->found_mdat)) {
        av_log(s, AV_LOG_ERROR, "mov: header not found !!! (err:%d, moov:%d, mdat:%d) pos:%"PRId64"\n",
                err, mov->found_moov, mov->found_mdat, url_ftell(pb));
        return -1;
    }
    dprintf("on_parse_exit_offset=%d\n", (int) url_ftell(pb));

    /* some cleanup : make sure we are on the mdat atom */
    if(!url_is_streamed(pb) && (url_ftell(pb) != mov->mdat_offset))
        url_fseek(pb, mov->mdat_offset, SEEK_SET);

    mov->next_chunk_offset = mov->mdat_offset; /* initialise reading */
    mov->total_streams = nb = s->nb_streams;

#if 1
    for(i=0; i<s->nb_streams;) {
        if(s->streams[i]->codec->codec_type == CODEC_TYPE_MOV_OTHER) {/* not audio, not video, delete */
            av_free(s->streams[i]);
            for(j=i+1; j<s->nb_streams; j++)
                s->streams[j-1] = s->streams[j];
            s->nb_streams--;
        } else
            i++;
    }
    for(i=0; i<s->nb_streams;i++) {
        MOVStreamContext *sc = (MOVStreamContext *)s->streams[i]->priv_data;

        if(!sc->time_rate)
            sc->time_rate=1;
        if(!sc->time_scale)
            sc->time_scale= mov->time_scale;
        av_set_pts_info(s->streams[i], 64, sc->time_rate, sc->time_scale);

        if(s->streams[i]->duration != AV_NOPTS_VALUE){
            assert(s->streams[i]->duration % sc->time_rate == 0);
            s->streams[i]->duration /= sc->time_rate;
        }

        sc->ffindex = i;
        sc->is_ff_stream = 1;
    }
#endif
    return 0;
}

/* Yes, this is ugly... I didn't write the specs of QT :p */
/* XXX:remove useless commented code sometime */
static int mov_read_packet(AVFormatContext *s, AVPacket *pkt)
{
    MOVContext *mov = (MOVContext *) s->priv_data;
    MOVStreamContext *sc;
    AVStream *st;
    int64_t offset = INT64_MAX;
    int64_t best_dts = INT64_MAX;
    int i, a, b, m;
    int next_sample= -99;
    int size;
    int idx;
    size = 0x0FFFFFFF;

    if (mov->partial) {
        sc = mov->partial;
        idx = sc->sample_to_chunk_index;

        if (idx < 0) return 0;
        dprintf("sc[ffid %d]->sample_size = %ld\n", sc->ffindex, sc->sample_size);
        //size = sc->sample_sizes[sc->current_sample];
        // that ain't working...
        //size = (sc->sample_size)?sc->sample_size:sc->sample_sizes[sc->current_sample];
        size = (sc->sample_size > 1)?sc->sample_size:sc->sample_sizes[sc->current_sample];

        next_sample= sc->current_sample+1;

        sc->left_in_chunk--;

        if (sc->left_in_chunk <= 0)
            mov->partial = 0;
        offset = mov->next_chunk_offset;
        /* extract the sample */

        goto readchunk;
    }

again:
    sc = 0;
    if(offset == INT64_MAX)
        best_dts= INT64_MAX;
    for(i=0; i<mov->total_streams; i++) {
        MOVStreamContext *msc = mov->streams[i];

        if ((msc->next_chunk < msc->chunk_count) && msc->next_chunk >= 0){
            if (msc->sample_to_time_index < msc->stts_count && mov->ni) {
                int64_t dts;
                int index= msc->sample_to_time_index;
                int sample= msc->sample_to_time_sample;
                int time= msc->sample_to_time_time;
                int duration = msc->stts_data[index].duration;
                int count = msc->stts_data[index].count;
                if (sample + count <= msc->current_sample) {
                    sample += count;
                    time   += count*duration;
                    index ++;
                    duration = msc->stts_data[index].duration;
                }
                dts = time + (msc->current_sample - sample) * (int64_t)duration;
                dts = av_rescale(dts, AV_TIME_BASE, msc->time_scale);
                dprintf("stream: %d dts: %"PRId64" best_dts: %"PRId64" offset: %"PRId64"\n", i, dts, best_dts, offset);
                if(dts < best_dts){
                    best_dts= dts;
                    sc = msc;
                    offset = msc->chunk_offsets[msc->next_chunk];
                }
            }else{
                if ((msc->chunk_offsets[msc->next_chunk] < offset)) {
                    sc = msc;
                    offset = msc->chunk_offsets[msc->next_chunk];
                }
            }
        }
    }
    if (!sc || offset==INT64_MAX)
        return -1;

    sc->next_chunk++;

    if(mov->next_chunk_offset < offset) { /* some meta data */
        url_fskip(&s->pb, (offset - mov->next_chunk_offset));
        mov->next_chunk_offset = offset;
    }

    if(!sc->is_ff_stream || (s->streams[sc->ffindex]->discard >= AVDISCARD_ALL)) {
        url_fskip(&s->pb, (offset - mov->next_chunk_offset));
        mov->next_chunk_offset = offset;
        offset = INT64_MAX;
        goto again;
    }

    /* now get the chunk size... */

    for(i=0; i<mov->total_streams; i++) {
        MOVStreamContext *msc = mov->streams[i];
        if ((msc->next_chunk < msc->chunk_count)
            && msc->chunk_offsets[msc->next_chunk] - offset < size
            && msc->chunk_offsets[msc->next_chunk] > offset)
            size = msc->chunk_offsets[msc->next_chunk] - offset;
    }

#ifdef MOV_MINOLTA_FIX
    //Make sure that size is according to sample_size (Needed by .mov files
    //created on a Minolta Dimage Xi where audio chunks contains waste data in the end)
    //Maybe we should really not only check sc->sample_size, but also sc->sample_sizes
    //but I have no such movies
    if (sc->sample_size > 0) {
        int foundsize=0;
        for(i=0; i<(sc->sample_to_chunk_sz); i++) {
            if( (sc->sample_to_chunk[i].first)<=(sc->next_chunk) )
            {
                foundsize=sc->sample_to_chunk[i].count*sc->sample_size;
            }
            dprintf("sample_to_chunk first=%ld count=%ld, id=%ld\n", sc->sample_to_chunk[i].first, sc->sample_to_chunk[i].count, sc->sample_to_chunk[i].id);
        }
        if( (foundsize>0) && (foundsize<size) )
        {
            size=foundsize;
        }
    }
#endif //MOV_MINOLTA_FIX

    idx = sc->sample_to_chunk_index;
    if (idx + 1 < sc->sample_to_chunk_sz && sc->next_chunk >= sc->sample_to_chunk[idx + 1].first)