I'm looking to create a series of short audio episodes (about 5 minutes each) for students that cover US History themes. I've been using ChatGPT effectively to produce materials, but I'm now interested in a straightforward way to generate audio files or podcasts with AI voices. I'd love your suggestions on the best solutions for this. Additionally, I want to create longer audio content (around 15-30 minutes) for personal learning on various topics that I'm also exploring with ChatGPT. What tools do you recommend?
2 Answers
I think Eleven Labs is one of the best options right now for generating educational audio. It provides pretty high-quality voices, which can help keep students engaged, although I’ve heard it doesn’t always emote perfectly. If emotion in the delivery is important to you, that could be a factor to consider!
You should definitely check out the GitHub repo called awesome-TTS-Colab! It has different TTS models you can run on your computer. Kokoro is great for a single narrator if you want high-quality output, and for dialogues with some natural sounds, Dia 1.6B is worth a shot. If you're looking for a ready-to-use tool, Google's NotebookLM might be suitable, but you can't edit the audio afterwards. Have you thought about Eleven Labs or Fliki? I'm curious how those compare to what I suggested!
Thanks for the suggestion! I’m leaning towards NotebookLM for its ease, but I do want a single narrator for my scripts, which is where Eleven Labs or Fliki come in. Do you think they offer better flexibility compared to your methods?