(in KHz) + fixed the bitrate ("$B") variable to use only the audio bit rate.
* winamp2.xml: updated the winamp2 skin

where the user normally use LANG, LANGUAGE, etc to define his/her
language(s). It didn't work anyway (refs #638).

* Support for Topfield .rec recordings. It's MPEG-TS with an exceptionally large (1MB) header in front of it.
  I tried adding the info in the header to meta, but the existing documention seems incorrect or at least incomplete. Instead we just accept the start of the header as identifying a MPEG-TS file, and skipping the header.

  BTW this header is again an example of very bad format design. It is using fixed length strings, for which the used buffers clearly are not null'ed before written with new information. As such i was able to determine which programs were running before this program was recorded :D
  Also a PID printed to a field was clearly not printed to null'ed memory, causing an invalid (too high) PID number to be in the field. You'd think the least someone could do when designing something like this, would be to analyze a hexdump of your own custom header.

*** NOT MEANT FOR BACKPORTING ***

* What is the point in having a detection for libvc1 if we don't have a libvc1 module ? Moreover, that prevents jlj's patch with the module from applying.

* Marillat changed the path to the win32 codecs in his latest packages, so look for the codecs there too

   you are supposed to pass frequency * 16000 instead of frequency * 16.
   This is sooo coherent.

I chose to re-add the align paramter to the subpicture region struct, since adding a style struct for every single subpicture for just an alignment seemed to be overkill to me.