Following Thomas's e-mail, I thought I should send a "summary report" too,
although don't expect as amazing achievements as his. :)
Over the period of of GSoC, I was able to port the WMA pro decoder from
ffmpeg. The decoder received lots of optimisations and now runs efficiently
both in memory usage and decoding speed. In the remaining time, I was able
to import libwmavoice to rockbox, but the port is still in an early stage as
it currently only works in the sim and is in floating point. By that stage,
I have achieved all primary and secondary goals on my proposal.
There's still additional work that needs to be done for both decoders,
namely, for WMA voice :
- Convert to fixed point.
- Get rid of unneeded code.
And adding support for multichannel ( >2-channel) streams in WMA pro.
For me the former is much more important and useful than the latter (as I
believe is the case for many of you), so I'll be giving my primary focus to
WMA voice. Although I have to start working on my graduation project in a
few days, and classes start in exactly one month from now, so my progress
will be definitely slower.
Project specifics aside, I sincerely thank you all for being such a great
community to work with, and special thanks go to Andree Buschmann and Nils
Wallménius for their help in optimising WMA pro. And of course special
thanks also for Michael Giacomelli for being such a great mentor for the 2nd
year in a row.
This year's GSoC has been better than the previous ones for rockbox, imho.
So, I'm glad I was part of it. This surely wouldn't have happened hadn't it
been to the successful participants/projects, so, congratulations Robert and
Received on 2010-08-16