Implement the klatt/holmes synthesizer from klatt and rsynth

[rhdunn]

The Klatt synthesizer (https://github.com/rhdunn/klatt) is a formant synthesiser. It cannot speak phonemes on its own, but uses parameters that describe formants from speech to artificially reconstruct the waveforms. As such, it does not use any recorded data like a diphone synthesiser does.

The klatt synthesiser is also used by the eSpeak and rsynth programs. In eSpeak, it is used as a complementary option to generate waveforms. In rsynth, it is used to synthesize all phonemes.

Rsynth provides its own imlpementation of the Klatt algorithm in the opsynth.c file. It uses the Holmes "vocal tract parameters" system to simplify interfacing with Klatt, implemented in holmes.c. The Holmes system is used to describe phonemes in the Elements.def file which is processed by elements.c. The phtoelem.c and phtoelem.def files support mapping SAMPA phonemes to these elements.

--- Want to back this issue? **[Post a bounty on it!](https://www.bountysource.com/issues/1026787-implement-the-klatt-holmes-synthesizer-from-klatt-and-rsynth?utm_campaign=plugin&utm_content=tracker%2F254961&utm_medium=issues&utm_source=github)** We accept bounties via [Bountysource](https://www.bountysource.com/?utm_campaign=plugin&utm_content=tracker%2F254961&utm_medium=issues&utm_source=github).

[rhdunn]

Depends on #38 (Phoneme Model)