ASF roundup

William Pitcock nenolod at
Sat Jan 6 12:51:52 CET 2007


asfheader.txt is not based on any Microsoft technical documents -- but 
instead cleanroom reverse engineering. In the United States and the EU, it 
is believed that cleanroom reveng is legal under the pretense of allowing 

The taglib-wma patch does not actually contain source from mplayer, but 
instead from ffmpeg. ffmpeg is fabrice bellard's project -- and as such is 
run slightly differently than mplayer. i'm pretty certain that the ffmpeg 
implementation is cleanroom because the asfdec.c code contains blank spots 
too. However, the asfdec.c parsecode is abysmal at best. That said, FFmpeg 
is LGPL.

By the way, on another note, this is how a WMAv9* looks:

Got ASF GUID: { 75b22630-668e-11cf-a6d90aa062ce6c }
Got File container GUID: { 8cabdca1-a947-11cf-8ee40c0c205365 }
ASF Payload size: 104
Media's GUID: { 44d1da1f-b1d4-480e-97bafd713a36516c }
Got Header1 GUID: { 5fbf03b5-a92e-11cf-8ee30c0c205365 }
ASF Payload size: 2829
Ext comment GUID (unsupported for retag at present):
   { d2d0a440-e307-11d2-97f00a0c95ea850 }
ASF Payload size: 1504
Codec Comment GUID: { 86d15240-311d-11d0-a3a40a0c9348f6 }
ASF Payload size: 186
Stream GUID: { b7dc0791-a9b7-11cf-8ee60c0c205365 }
ASF Payload size: 114
Unknown GUID: { 7bf875ce-468d-11d1-8d8206097c9a2b2 }
ASF Payload size: 32
Comment GUID: { 75b22633-668e-11cf-a6d90aa062ce6c }
ASF Payload size: 132

.. and next is the media GUID and codec-data afterwards.

* I'm not sure if this is a WMAv9 or not -- the person I acquired it from
   generated it using WMP9, though.

- William

On Sat, 6 Jan 2007, Xavier Duret wrote:

> I am quite happy about how the discussion is evolving.
> There are now 4 issues that need to be solved to fix the problem of ASF support using taglib.
> - The topic of the source specification. William, can you tell us what you did to create your "asfheader.txt" document? In particular, did you read Microsoft specification? As a side note, can you specify whether the different fields are big-endian or little-endian?
> - The topic of the patent. Looking at William's document, I claim that what is in there is not covered by US Patent 6,041,345 on the ground that there is no "error correcting data". Thankfully, you do not have to take my word for it if Scott manages to get one of the lawyers doing pro-bono work for KDE to have a look at the issue.
> - The topic of the license. The implementation currently in Amarok is GPL Are the copyright owners willing to relicense it to LGPL?
> - The topic of the repository. There is currently no central repository for the "taglib plugins". Project that wish to use taglib and ASF need to use the taglib delivery and rip the ASF code from Amarok. Would it be possible to have them somewhere else like "/kde/trunk/kdesupport/taglib-plugins" with the proper configure script? Is it also possible to some mention of the plugins on the taglib webpage?
> Life is fully of surprises. I do not have any music in WMA, I dislike proprietary formats but I find myself in the position of advocating for the inclusion of WMA support in a free software project...
> Gagnez tous les jours un iPod nano. Inscrivez-vous au courriel gratuit CaraMail. Et bénéficiez de 10 SMS gratuits ! Cliquez ici pour gagner :

More information about the taglib-devel mailing list