Progressive audio data decoding

Hello,

Is there a way to decode the audio data in chunks? The decodeAudioData 
method accepts the data that is complete audio files. But in some cases 
the data is delivered in chunks:
   a) the XHR range requests are used to deliver the sound track data 
(see also the moz-chunked-arraybuffer alternative);
   b) or, the encoded data is stored in some other containers (e.g. see 
swf format).

The decoding of the audio data portions without saving some intermediate 
state is not possible for some formats, even if the data is split on a 
frame boundary, e.g. in MP3 format, to decode a frame you will need a 
data from at least one previous frame. Also, the "sound1.start(time1) / 
sound2.start(time1 + duration1)" workaround is needed for playback of 
the decoded chunks (does it even guarantee the absence of playback 
artifacts?) and deep understanding of the frame format is required.

Is it possible to add the createAudioDataDecoder (?) method that will 
return the decoder session / buffer, which can be used to "push" the 
encoded audio data in chunks? If the decoding will return more than one 
buffer, the automatic scheduling of the next buffer playback must be 
provided.

Thank you,
Yury Delendik

Received on Tuesday, 15 January 2013 17:53:00 UTC