I had a faculty request for a way to generate target-language audio from a provided script. In this case the language was Russian.
I tried it Gemini, ChatGPT, and Claude as a normal prompt but all of them failed pretty badly. Listen to Claude’s attempt here. ChatGPT gave up on it. Gemini got completely confused and just gave me back text.
Google’s AI Studio provides one free option. I’m not sure what the limits are but this would be an option that could be made into a tool if we wanted. It’d be easy to make a dialogue constructor and then push the content through to the API.
I already had some dialogue with character names.
That didn’t work with…
I had a faculty request for a way to generate target-language audio from a provided script. In this case the language was Russian.
I tried it Gemini, ChatGPT, and Claude as a normal prompt but all of them failed pretty badly. Listen to Claude’s attempt here. ChatGPT gave up on it. Gemini got completely confused and just gave me back text.
Google’s AI Studio provides one free option. I’m not sure what the limits are but this would be an option that could be made into a tool if we wanted. It’d be easy to make a dialogue constructor and then push the content through to the API.
I already had some dialogue with character names.
That didn’t work with the format that Google AI studio wanted. So I did a find/replace to exchange the human names with “Speaker 1” and “Speaker 2.” Note that I did not use AI to do this. AI for making voices in Russian? Pretty wild and useful. AI for doing find/replace like this? A terrible idea.
A short time later, I could download the audio file as a wav.