New API for audio encoders

[rom1mouret]

• AudioEncoders take as input a list of buffers instead of a single one
• New class: AudioBuffer, expected to be removed from the TTS modules
• AudioEncoders return a Guava's optional instead of null strings
• Updated the versions in the BOM.

[rdeltour]

LGTM except for the two minor inline comments.

The change introduces an API break, but it's OK AFAICS. Alternatively we could keep the old signature and add a new method accepting a list of buffers if that makes sense, your call.

[rom1mouret]

I prefer the API break option. I changed the versions to make it compile, but I thought actual version increments had to be done when merging PRs (http://code.google.com/p/daisy-pipeline/wiki/Versioning) so that it makes sense to merge PR in any order?

[rdeltour]

I prefer the API break option.

OK, fine by me.

I changed the versions to make it compile, but I thought actual version increments had to be done when merging PRs

Yes, you're right. Just wanted to make sure you were aware of that.

[rdeltour]

Merged as 04dfabb0b144b5e2f5fe722245bfc6dccfa588e0