[mad-dev] Clear input buffer with high-level API?

Rob Leslie rob@mars.org
Tue, 5 Nov 2002 20:20:12 -0800


On Tuesday, November 5, 2002, at 10:39  AM, Eric Lyons wrote:
> I'm trying to figure out how to tell the decoding process to flush the
> input buffer after a MAD_FLOW_STOP is returned from decode_input_read
> (in madplay.c).
>
> It stops playing fine, but when I decode() again, it picks up from
> where it left off in the buffer.  I want to retain this behavior too,
> but also have the ability to kill the current buffer when it restarts.
>
> Is this not possible with the high-level API?  Perhaps I can add a
> MAD_FLOW_FLUSH?

Before you start decoding again, if you want to decode something 
different, you should call mad_stream_buffer() to reset the stream 
location. You may also want to reset the internal decoding state by 
calling mad_frame_mute() and mad_synth_mute().

-- 
Rob Leslie
rob@mars.org