Voices

The speect.SVoice() method can load a voice that has been defined either as a JSON voice definition, or as a Python function in a file. The advantage of loading a voice through a Python function is that one can define the voice’s utterance and feature processors with the voice. Lets create a file called my_voice.py and define a function called load_voice as follows:

def load_voice():

    import speect
    import speect.uttproc_cb   # utterance callback interface

    # load voice json file
    voice = speect.SVoice("voice.hts.json") # load voice definition

    # load any required data

    ...

    # create utterance processors

    ...

    # set utterance processors in voice

    ...

    return voice

The function loading the voice must be named load_voice. On line 7 a JSON voice definition is loaded, this defines the voice processors and utterance types. Now other processors can be created and set in the voice, and the voice is returned. This file, my_voice.py, can now be used to load the voice from within Python as follows (this example shows how to do it with the Python interpreter):

$ python
>>> import speect
>>> voice = speect.SVoice("my_voice.py")

Example

A complete example can be found at Voice Loading.