Integration of Text-to-Speech (TTS) Feature for Enhanced Immersion

[SverreNystad]

While the GPT Dungeon Master provides textual responses, having an auditory component can significantly enhance the immersion for players, especially during intense or dramatic moments in the game.

I propose the integration of a Text-to-Speech (TTS) feature that can convert the GPT DM's textual responses into spoken words. This would allow players to hear the DM's narration, making the gameplay experience more engaging and lifelike.

Proposed Features:

• Voice Selection: Provide options for different voice tones (e.g., deep, mystical, cheerful) to match the mood of the game.
• Volume Control: Allow users to adjust the TTS volume independently of other game sounds.
• Speed Adjustment: Give users the ability to speed up or slow down the TTS to their preference.
• Pause/Resume: Implement a feature to pause and resume the TTS, especially useful during longer narrations.
• Emotions: Make the model
• Skip: Many players find long dialog irritating, we should have a method to skip current Speech.

[SverreNystad]

OpenAIs guild on speech to text: https://platform.openai.com/docs/guides/speech-to-text
Watsons tts: https://www.youtube.com/watch?v=8k8S5ruFAUs

[SverreNystad]

I started out using watson but lovo seems better as it allows us to change the speed and emotion of the voice. We can use classifiers or the LLM with function calling to get it to use correct emotion for voice and speed of it
https://api-doc.lovo.ai/

[SverreNystad]

We need to discuss what is the best text to speech bot