Merge changes from Hanson Robotics

[linas]

No description provided.

[linas]

err, don't merge yet ... there are some problem here, I will fix them shortly.

[linas]

OK, I fixed the few obvious packaging bugs. This is now ready to be pulled. Vytas made several changes I am not clear about:
-- using wav instead of/in addition to mp3
-- something to do with the jaw
-- publish of jaw movement for the dmitri robot

I don't know if/how any of these changes might break compatiblity

[linas]

See also: https://github.com/hansonrobotics/tts_speech/issues/1

[jdddog]

This line should be changed back to how it was; its bad to hard code the topic name to produce data for a particular system.

The dmitry jaw controller should instead subscribe to the topic speech_strength.

[linas]

Update per comments. There's not much that is interesting that is left ...

[linas]

Do the mp3->wav conversions look reasonable? If not, let me know, else please merge.