This is the sample project from my Text-to-Speech tutorial on YouTube. The included dictionary is from the Carnegie Mellon University Computer Science Department, although I removed a version of "deja" that contained non-ASCII characters which would raise an error (There is a way to properly handle it though). Additionally, not all the speech sounds have been included.
Video: https://www.youtube.com/watch?v=KSSsVhoR7FQ
CMU Source: http://www.speech.cs.cmu.edu/cgi-bin/cmudict
See my official dev guide for detailed instructions: http://arcompware.com/python-basic-text-to-speech-engine/
US English speakers, this is your opportunity to be your own TTS so try recording your own sounds!