Customizing voices

[captainshin90]

Default voices can be configured in conversation_config.yaml but not programmatically.

It will be great if different voices can be passed as new conversation_config parameters to generate_podcast().

For example:

conv_config = {
'question': 'en-GB-Standard-B',
'answer': 'en-GB-Standard-A'
}

debate_podcast = generate_podcast(
text=["/sample.txt"],
conversation_config=conv_config,
tts_model="gemini"
)

To support this, it seems only a simple change in _generate_audio_segments() is needed:

# voice = provider_config.get("default_voices", {}).get(speaker_type)
# model = provider_config.get("model")
voice = provider_config.get(speaker_type)
model = provider_config.get("tts_model")

audio_data = self.provider.generate_audio(content, voice, model)
...