I'm curious about how voice technology powered by artificial intelligence has gained momentum lately. What changes have occurred, especially in projects transitioning from text-to-speech (TTS)? I'm researching for a small project I'm developing, but I'm getting lost among the constantly changing tools. What do you guys use? Any open-source or free options out there? Should we take a look together? 😅
How do I keep up with AI voiceover trends?
👁️ 10 views💬 3 replies❤️ 0 likes
3 Replies
The latest tests on the so-called E123 model really surprised me with its improved audio quality and extremely low hardware requirements. I’ve been experimenting with open-source tools like DenoiSe, Coqui TTS, and Piper, and reading through the experiences of local users in those communities.
For example, as digital voice synthesis improves, TTS tools have started to better capture tone and emotion. I think the once robotic-sounding voices now perform almost as well as real human voices. In your projects, you can use open-source tools like RVC (Retrieval-Based Voice Conversion) to change the style of a voice—for example, making one voice sound like another. 🎙️
The TTS world is changing really fast. I'm always looking for new tools for small projects, just like you. I used to rely only on Google's TTS API, but lately, I've noticed that open-source options have come a long way. I'm especially impressed by projects like **Coqui TTS** and **Mozilla TTS**—they deliver high-quality results and let you train the model with your own voice data.
In my experience, I've worked with models like **Bark** and **VITS**. Bark excels at natural intonation, while VITS runs well even on limited hardware. Among free options, Amazon Polly's free tier is available, but it's limited in voice choices. What language are you working with? For Turkish voiceovers, I sometimes test the **VITS-turkish** model, and the results are promising. 😃