* use larger buffer - the size should be calculcated from the bitstream

  data rate and outgoing samples (i.e. WMA could generate a lot of samples
  from 4096KB chunks)
  FIXME

Originally committed as revision 1169 to svn://svn.ffmpeg.org/ffmpeg/trunk

diff --git a/ffmpeg.c b/ffmpeg.c
index b9dc932..00809c6 100644 (file)
--- a/ffmpeg.c
+++ b/ffmpeg.c
@@ -301,7 +301,7 @@ static void do_audio_out(AVFormatContext *s,
 {
     UINT8 *buftmp;
     UINT8 audio_buf[2*MAX_AUDIO_PACKET_SIZE]; /* XXX: allocate it */
-    UINT8 audio_out[MAX_AUDIO_PACKET_SIZE]; /* XXX: allocate it */
+    UINT8 audio_out[4*MAX_AUDIO_PACKET_SIZE]; /* XXX: allocate it - yep really WMA */
     int size_out, frame_bytes, ret;
     AVCodecContext *enc;
 
@@ -346,7 +346,7 @@ static void do_audio_out(AVFormatContext *s,
             break;
         }
         ret = avcodec_encode_audio(enc, audio_out, size_out, 
-                                   (short *)buftmp);
+                                  (short *)buftmp);
         av_write_frame(s, ost->index, audio_out, ret);
     }
 }