ANNOUNCE: gst-audiodescription

[David Holroyd]

Oh hi!


See,
     https://github.com/dholroyd/gst-audiodescription


I have some WAV files containing 'WHP 198' audio description data [1].

The Internet doesn't seem know of any tools that could help me use these 
files, so I created the above plugin providing a couple of elements,

  * whp198dec - decodes and checksums the Manchester Encoded audio 
waveform per WHP 198 to produce buffers of 'AD_descriptor' structures
  * adcontrol - consumes 'AD_descriptor' structures and uses these to 
apply a 'fade' to the volume of some other audio stream

Combining these with other existing elements ('deinterleave', 
'audiomix', gaffer tape, string etc.) will eventually allow construction 
of pipelines which can combine one of these WAV files with the file for 
the video being described, and usefully render the described audio.


Code is just a proof of concept at the moment, but I would be very 
interested to hear if,

  a) anyone else uses data in these formats themselves (collaborators 
welcome!)

  b) there are any problems with the code (I'm new to gstreamer!)

  c) I missed something out there that *does* already support WHP 198 
etc. (which I've just not been smart enough to find!)


ta,
dave


[1] If you're not quite interested enough to google the spec, there is 
effectively mono description audio in the left channel, and some 
manchester-encoded metadata in the right channel.  This metadata signals 
temporary reductions in the volume of the main audio track in the 
related video content (i.e. in another file), so as to make the 
description audio clearer.


-- 
http://david.holroyd.me.uk/