Merged vbns-espeak and vbns-ao2 in the same project

[jmdaweb]

The following changes have been made in this pull request:

• Added again msvc libraries (only crt).
• Added espeak modules and libraries.
• Updated setup.py to take into account the previous changes when building.
• Updated readme.
• New command line argument: --espeak-voice or -ev. Defaults to english, with an empty string as const.
• If espeak variable is not empty, use Espeak as the output synthesizer. Otherwise, use Accessible output / Sapi.
• As a result, there are three new if statements in the code to choose which synth must be silenced, which one must be used for speaking and which rate_map should be defined.

Please, review carefully before merging, I haven't tested anything yet.

[sukiletxe]

Lgtm, merging. Thank you very much and sorry for the waiting!