[Matroska-devel] [Fwd: Re: Musepack support]
Christian HJ Wiesner
chris at matroska.org
Fri Feb 8 14:05:17 CET 2008
allow me one more question : How stable is the current spec ? Is it
worth for the muxing app devs, mosu and alex, to start looking at it now ?
Mosu, Alex : any chance to inform us about your current shedules for
your apps ? Can the musepack apps expect some alpha MKA muxing anytime
-------- Original-Nachricht --------
Betreff: Re: Musepack support
Datum: Fri, 08 Feb 2008 11:59:08 +0100
Von: Nicolas <r2d at musepack.net>
An: chris at matroska.org, Discussion about the current and future
development of Matroska <matroska-devel at lists.matroska.org>
Referenzen: <fi0u64$q7v$1 at ger.gmane.org> <47A9ED92.6070707 at matroska.org>
I Christian, thanks for your answer.
Christian HJ Wiesner a écrit :
I think I prefer this one
> - a very clear spec of the format itself, and the way the audio packets
> should be muxed into matroska, including headers etc.
> - if possible some example code for a demuxer and muxer (for
> mkvextract), preferably in C or C++
The SV8 spec is here :
The only necessary header for playing the stream is the "Stream Header Packet", so this one must be in the codec private data. Other header packets (RG, EI) are not necessary for playing the file, so may be put somewhere else (or not put at all).
The main data packet is the "Audio packet".
For building a seeking table, you will need to read the "Audio block frames" field in the Stream Header Packet, which defines the number of audio frames (1152 samples) per audio packet.
A C implementation of the demuxer is here :
or see libmpcdec in the libmpc svn :
More information about the Matroska-devel