[mad-user] Stream Decode

[john cooper]

Alexey V Paramonov wrote:
> Hi!
> 
> I have the following question - is it possible to decode a stream using 
> MAD?

Yes, of course.

> For example: I have shoutcast stream and I read it in small portions
> (say, 40960 bytes), decode it and encode back to mp3 using LAME
> (chaning bitrate, samplerate, etc...).
> 
> The problem is that I have to call mad_stream_buffer with each piece of
> stream data, decode/encode it and then if I listen the resulting mp3
> file - I notice "noise" between theese small decoded pieces.
> Seems like it happen not only with VBR files.

Haven't looked at madplay (etc..) in years but originally
it was mmap()'ing an mp3 file into memory which works
fine assuming the mmap() region could grow for as long as
you need to decode the respective file.

If rather you are trying to decode into a fixed buffer
you'll need to do some adjustment when your decode nears
the end of the static buffer.  This means detecting when
you don't have enough data left at the buffer tail to
do a valid frame decode (ie: mad_frame_decode() returns
MAD_ERROR_BUFLEN).  In that case you need to copy the
partial frame at the buffer's tail, back at/near the head
of the buffer (depending on your stream management logic)
and readjust libmad's notion of where it is via a
mad_stream_buffer() call.

> Pseudo-code:
> 
> while ( bytes = read(file_id, buff, SHOUT_BUFFER_LENGTH ) )  {
> 
>    mad_stream_buffer(&stream, buff, bytes );
> 
>    while(1){
>    mad_frame_decode(&frame,&stream)
>    ....
>    int ret = lame_encode_buffer( lameGlobalFlags,
>                              left,
>                              right,
>                              nsamples,
>                              mp3Buf,
>                              mp3Size );
>    // Play resulting mp3 data
> 
>   }
> 
> }

I really don't understand what you are trying to
accomplish via the back-to-back decode/encode but
I suppose that is unrelated to your question.

> Is it possible to decode a stream without theese "edge effects" ?

Yes but it can be a little tricky to get right in the
case you allow a user to manipulate the stream pointer
and perform arbitrary seeks within the stream.

> May be I should not re-initialize MAD buffer after each read in some way?

If you are altering the [memory address] seek point
within the current buffer you'll need to reset the
decoder's notion of decode location via a mad_stream_buffer()
call.

-john

-- 
john.cooper at third-harmonic.com