adding audio output to avwrapper

1 files changed, 77 insertions, 6 deletions

diff --git a/rotord/libavwrapper.cpp b/rotord/libavwrapper.cpp
index 5cff8c9..1fe054e 100755
--- a/rotord/libavwrapper.cpp
+++ b/rotord/libavwrapper.cpp
@@ -368,7 +368,7 @@ bool libav::decoder::fetchFrame(int targetFrameIndex)
         if (response < 0)
             return false;
         if (response!=targetFrameIndex){
-            cerr<<"libav::decoder asked for "<<targetFrameIndex<<", got "<<response<<endl;
+            cerr<<"libav::decoder asked for "<<targetFrameIndex<<", got "<<response<<endl; //does not seem to be aware of wrong frame
         }
     }
     previousFrameIndex = targetFrameIndex;
@@ -386,10 +386,10 @@ int libav::decoder::seekToFrame(int targetFrameIndex)
     }
     int result = avformat_seek_file( container, //format context
                             videoStream,//stream id
-                            0,               //min timestamp
+                            0,               //min timestamp 0?
                             ts,              //target timestamp
                             ts,              //max timestamp
-                            0); //AVSEEK_FLAG_ANY),//flags
+                            0);//flags AVSEEK_FLAG_ANY //doesn't seem to work great
     if (result < 0)
         return -1;
 
@@ -546,11 +546,14 @@ bool libav::decoder::b_is_one_time_inited = false;
 ///////////////////////////
 
 
-libav::encoder::encoder(const char * file_name, int width, int height, enum AVCodecID codec_id)
+libav::encoder::encoder(const char * file_name, int width, int height, float _framerate,enum AVCodecID codec_id)
     : picture_yuv(NULL)
     , picture_rgb(NULL)
     , container(NULL)
 {
+    //multiply float seconds by this to get pts
+    timebase=((float)AV_TIME_BASE_Q.den)/(AV_TIME_BASE_Q.num*_framerate*3.125f); //no idea where the 3.125 comes from
+
     if (0 != (width % 2))
         cerr << "WARNING: Video width is not a multiple of 2" << endl;
     if (0 != (height % 2))
@@ -605,7 +608,7 @@ libav::encoder::encoder(const char * file_name, int width, int height, enum AVCo
         // TODO
     }
 
-    pCtx->time_base = (AVRational){1, 25};
+    pCtx->time_base = (AVRational){1, 25}; /////TODO FIX TO SUPPORT OTHER RATES
     // pCtx->time_base = (AVRational){1, 10};
     pCtx->gop_size = 12; // emit one intra frame every twelve frames
     // pCtx->max_b_frames = 0;
@@ -679,6 +682,27 @@ libav::encoder::encoder(const char * file_name, int width, int height, enum AVCo
             SWS_BICUBIC,NULL,NULL,NULL)) )
         throw std::runtime_error("");
 
+//
+//
+// added audio init
+    AVCodec * acodec = avcodec_find_encoder(AV_CODEC_ID_AAC);
+    int ret = avcodec_open2(pCtx, acodec, NULL);
+    if (ret < 0) {
+        throw std::runtime_error("Could not open audio codec:");
+        
+    }
+
+    if (pCtx->codec->capabilities & CODEC_CAP_VARIABLE_FRAME_SIZE)
+            audio_input_frame_size = 10000;
+    else
+        audio_input_frame_size = pCtx->frame_size; //is coming out at 0?
+
+    audiostep=((float)audio_input_frame_size)/(pCtx->sample_rate);
+
+
+// are we supposed to use the same codeccontext?
+//
+
     /* open the output file */
     if (!(fmt->flags & AVFMT_NOFILE))
     {
@@ -695,8 +719,10 @@ void libav::encoder::setPixelIntensity(int x, int y, int c, uint8_t value)
     *ptr = value;
 }
 
-void libav::encoder::write_frame()
+void libav::encoder::write_frame(float seconds,uint8_t *rgbdata)
 {
+    picture_rgb->data[0]=rgbdata;
+
     // convert from RGB24 to YUV
     sws_scale(Sctx,              // sws context
               picture_rgb->data,        // src slice
@@ -713,6 +739,11 @@ void libav::encoder::write_frame()
     packet.data = NULL;
     packet.size = 0;
 
+    //no time stamps as is
+    //http://dranger.com/ffmpeg/tutorial07.html
+
+    picture_yuv->pts=(uint64_t)(seconds*timebase);
+
     int got_packet;
     int ret = avcodec_encode_video2(pCtx,
                                     &packet,
@@ -727,6 +758,46 @@ void libav::encoder::write_frame()
         av_destruct_packet(&packet);
     }
 }
+void libav::encoder::write_frame(float seconds,uint16_t *audiodata){
+    audio_frame = avcodec_alloc_frame();
+    AVPacket pkt = { 0 }; // data and size must be 0;
+    int got_packet, ret;
+    av_init_packet(&pkt);
+    audio_frame->nb_samples = audio_input_frame_size;
+    uint8_t *sampleptr;
+    int bufsize=audio_input_frame_size * av_get_bytes_per_sample(pCtx->sample_fmt) *pCtx->channels;
+    if (audiodata) {
+        sampleptr=(uint8_t*)audiodata;
+    }
+    else {
+        sampleptr=new uint8_t[bufsize];
+        memset(sampleptr,0,bufsize);
+    }
+
+     avcodec_fill_audio_frame(audio_frame, pCtx->channels, pCtx->sample_fmt,
+                                 sampleptr,
+                                 audio_input_frame_size *
+                                 av_get_bytes_per_sample(pCtx->sample_fmt) *
+                                 pCtx->channels, 0); //;
+
+     audio_frame->pts=(uint64_t)(seconds*timebase);
+
+     ret = avcodec_encode_audio2(pCtx, &pkt, audio_frame, &got_packet);
+     if (!audiodata) {
+        delete[] sampleptr;
+    }
+    if (ret < 0) {
+        throw std::runtime_error("Audio encoding failed");
+    }
+
+    if (!got_packet)
+            return;
+
+    // ? pkt.stream_index = st->index; 
+
+    ret = av_interleaved_write_frame(container, &pkt);
+    avcodec_free_frame(&audio_frame);
+}
 
 /* virtual */
 libav::encoder::~encoder()