add FLV decoder metadata parsing

patch by Allan Hsu % allan A counterpop P net %
date: Dec 12, 2006 12:19 PM
subject: Re: [Ffmpeg-devel] [PATCH] FLV decoder metadata reading

Originally committed as revision 7286 to svn://svn.ffmpeg.org/ffmpeg/trunk

1 files changed, 175 insertions, 66 deletions

diff --git a/libavformat/flvdec.c b/libavformat/flvdec.c
index 2906b1d4ec..4aaeae6b9f 100644
--- a/libavformat/flvdec.c
+++ b/libavformat/flvdec.c
@@ -40,6 +40,45 @@ static int flv_probe(AVProbeData *p)
     return 0;
 }
 
+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
+        case FLV_CODECID_PCM_BE:
+            acodec->codec_id = acodec->bits_per_sample == 8 ? CODEC_ID_PCM_S8 : CODEC_ID_PCM_S16BE; break;
+        case FLV_CODECID_PCM_LE:
+            acodec->codec_id = acodec->bits_per_sample == 8 ? CODEC_ID_PCM_S8 : CODEC_ID_PCM_S16LE; break;
+        case FLV_CODECID_ADPCM: acodec->codec_id = CODEC_ID_ADPCM_SWF;                              break;
+        case FLV_CODECID_MP3  : acodec->codec_id = CODEC_ID_MP3      ; astream->need_parsing = 1  ; break;
+        case FLV_CODECID_NELLYMOSER_8HZ_MONO:
+            acodec->sample_rate = 8000; //in case metadata does not otherwise declare samplerate
+        case FLV_CODECID_NELLYMOSER:
+        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   ;
+            if(vcodec->extradata_size != 1) {
+                vcodec->extradata_size = 1;
+                vcodec->extradata = av_malloc(1);
+            }
+            vcodec->extradata[0] = get_byte(&s->pb);
+            return 1; // 1 byte body size adjustment for flv_read_packet()
+        default:
+            av_log(s, AV_LOG_INFO, "Unsupported video codec (%x)\n", flv_codecid);
+            vcodec->codec_tag = flv_codecid;
+    }
+
+    return 0;
+}
+
 static int amf_get_string(ByteIOContext *ioc, char *buffer, int buffsize) {
     int length = get_be16(ioc);
     if(length >= buffsize) {
@@ -54,6 +93,135 @@ static int amf_get_string(ByteIOContext *ioc, char *buffer, int buffsize) {
     return length;
 }
 
+static int amf_parse_object(AVFormatContext *s, AVStream *astream, AVStream *vstream, const char *key, unsigned int max_pos, int depth) {
+    AVCodecContext *acodec, *vcodec;
+    ByteIOContext *ioc;
+    AMFDataType amf_type;
+    char str_val[256];
+    double num_val;
+
+    num_val = 0;
+    ioc = &s->pb;
+
+    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;
+            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;
+            else if(!strcmp(key, "audiocodecid") && acodec) flv_set_audio_codec(s, astream, (int)num_val << FLV_AUDIO_CODECID_OFFSET);
+            else if(!strcmp(key, "videocodecid") && vcodec) flv_set_video_codec(s, vstream, (int)num_val);
+            else if(!strcmp(key, "audiosamplesize") && acodec && num_val >= 0) {
+                acodec->bits_per_sample = num_val;
+                //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
+                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;
+                }
+            }
+        }
+    }
+
+    return 0;
+}
+
+static int flv_read_metabody(AVFormatContext *s, unsigned int next_pos) {
+    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;
+    ioc = &s->pb;
+
+    //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;
+}
+
 static int flv_read_header(AVFormatContext *s,
                            AVFormatParameters *ap)
 {
@@ -114,43 +282,11 @@ static int flv_read_packet(AVFormatContext *s, AVPacket *pkt)
     } else if (type == FLV_TAG_TYPE_VIDEO) {
         is_audio=0;
         flags = get_byte(&s->pb);
-    } else if (type == FLV_TAG_TYPE_META && size > 13+1+4) {
-        url_fskip(&s->pb, 13); //onMetaData blah
-        if(get_byte(&s->pb) == 8){
-            url_fskip(&s->pb, 4);
-        }
-        while(url_ftell(&s->pb) + 5 < next){
-            char tmp[128];
-            int type, len;
-            double d= 0;
-
-            if(amf_get_string(&s->pb, tmp, sizeof(tmp))<0)
-                break;
-
-            type= get_byte(&s->pb);
-            if(type == AMF_DATA_TYPE_NUMBER){
-                d= av_int2dbl(get_be64(&s->pb));
-            }else if(type == AMF_DATA_TYPE_STRING){
-                amf_get_string(&s->pb, NULL, 0);
-            }else if(type == AMF_DATA_TYPE_MIXEDARRAY){
-                //array
-                break;
-            }else if(type == AMF_DATA_TYPE_DATE){
-                d= av_int2dbl(get_be64(&s->pb));
-                get_be16(&s->pb);
-            }
-
-            if(!strcmp(tmp, "duration")){
-                s->duration = d*AV_TIME_BASE;
-            }else if(!strcmp(tmp, "videodatarate")){
-            }else if(!strcmp(tmp, "audiodatarate")){
-            }
-        }
-        url_fseek(&s->pb, next, SEEK_SET);
-        continue;
     } else {
-        /* skip packet */
-        av_log(s, AV_LOG_ERROR, "skipping flv packet: type %d, size %d, flags %d\n", type, size, flags);
+        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);
         url_fseek(&s->pb, next, SEEK_SET);
         continue;
     }
@@ -194,44 +330,17 @@ static int flv_read_packet(AVFormatContext *s, AVPacket *pkt)
     }
 
     if(is_audio){
-        if(st->codec->sample_rate == 0){
+        if(!st->codec->sample_rate || !st->codec->bits_per_sample || (!st->codec->codec_id && !st->codec->codec_tag)) {
             st->codec->channels = (flags & FLV_AUDIO_CHANNEL_MASK) == FLV_STEREO ? 2 : 1;
             if((flags & FLV_AUDIO_CODECID_MASK) == FLV_CODECID_NELLYMOSER_8HZ_MONO)
                 st->codec->sample_rate= 8000;
             else
                 st->codec->sample_rate = (44100 << ((flags & FLV_AUDIO_SAMPLERATE_MASK) >> FLV_AUDIO_SAMPLERATE_OFFSET) >> 3);
-            switch(flags & FLV_AUDIO_CODECID_MASK) {
-            case FLV_CODECID_PCM_BE: if (flags & FLV_AUDIO_SAMPLESIZE_MASK) st->codec->codec_id = CODEC_ID_PCM_S16BE;
-                    else st->codec->codec_id = CODEC_ID_PCM_S8; break;
-            case FLV_CODECID_ADPCM : st->codec->codec_id = CODEC_ID_ADPCM_SWF;                       break;
-            case FLV_CODECID_MP3   : st->codec->codec_id = CODEC_ID_MP3      ; st->need_parsing = 1; break;
-            // this is not listed at FLV but at SWF, strange...
-            case FLV_CODECID_PCM_LE: if (flags & FLV_AUDIO_SAMPLESIZE_MASK) st->codec->codec_id = CODEC_ID_PCM_S16LE;
-                    else st->codec->codec_id = CODEC_ID_PCM_S8; break;
-            default:
-                    av_log(s, AV_LOG_INFO, "Unsupported audio codec (%x)\n", flags >> 4);
-                st->codec->codec_tag= (flags >> 4);
-            }
             st->codec->bits_per_sample = (flags & FLV_AUDIO_SAMPLESIZE_MASK) ? 16 : 8;
+            flv_set_audio_codec(s, st, flags & FLV_AUDIO_CODECID_MASK);
         }
     }else{
-            switch(flags & FLV_VIDEO_CODECID_MASK){
-            case FLV_CODECID_H263  : st->codec->codec_id = CODEC_ID_FLV1   ; break;
-            case FLV_CODECID_SCREEN: st->codec->codec_id = CODEC_ID_FLASHSV; break;
-            case FLV_CODECID_VP6   :
-                st->codec->codec_id = CODEC_ID_VP6F;
-                if (st->codec->extradata_size != 1) {
-                    st->codec->extradata_size = 1;
-                    st->codec->extradata = av_malloc(1);
-                }
-                /* width and height adjustment */
-                st->codec->extradata[0] = get_byte(&s->pb);
-                size--;
-                break;
-            default:
-                av_log(s, AV_LOG_INFO, "Unsupported video codec (%x)\n", flags & FLV_VIDEO_CODECID_MASK);
-                st->codec->codec_tag = flags & FLV_VIDEO_CODECID_MASK;
-            }
+        size -= flv_set_video_codec(s, st, flags & FLV_VIDEO_CODECID_MASK);
     }
 
     ret= av_get_packet(&s->pb, pkt, size - 1);