tts-speak: implement embedded speech commands

[ghost]

Apple tts has a nice sugarcandy which would be incredibly useful in shell scripts - the possibility to embed commands into the text such as [[emph +]], [[slnc 6000]], [[volm +0.4]]. Could also use [[beep duration pitch]] or [[play file.ogg]] which apple does not seem to have.

https://developer.apple.com/library/mac/documentation/UserExperience/Conceptual/SpeechSynthesisProgrammingGuide/FineTuning/FineTuning.html

I found the Android Pico TTS supports it (Chapter 7 of its documentation) and a few related stackoverflow posts:
http://stackoverflow.com/questions/3495301/best-practice-for-specifying-pronunciation-for-android-tts-engine
http://stackoverflow.com/questions/3495301/best-practice-for-specifying-pronunciation-for-android-tts-engine

Not sure if it can work without speacial help of termux-api?