A few weeks ago, I wrote about a couple of text-to-speech applications, aimed at producing quality voice overs for various uses. I ran into CyberVoice, another application that does it, except that this time it’s a web application.
From what I understand, CyberVoice uses a different technology for voice synthesis, studying various samples of voice tones from an actor and then using it to create a synthetic voice from your text input. The web app requires you to purchase a subscription to create voices you can actually use, but they also offer the ability to test it out for free.
I didn’t get a paid subscription to test the service. I just created an account to see how the app works and to test the voices they already have available. Using the app is very simple. You just enter the text, and then you use the fine-tuning tab to tweak the sound of your speech.
When you are fine-tuning, you can change the phonemes and edit them, and also add emphasis to change the way the voice reads the text. Then, you push the microphone button and wait for the software to synthetize the voice.
Right now, there are 7 voices you can choose from: four English voices and three Russian voices. While the results could still be better. One thing I don’t quite like about it is how the interface combines English and Russian. I was using the English interface, obviously, but some of the explanations shown during hovering the mouse over a button are still shown in Russian, so I don’t understand what they are supposed to mean (and I don’t think I should be using an online translator to figure out instructions of a service being offered to English-speaking people).
If you read my previous articles covering text-to-speech solutions, you will know of the things I didn’t like about them is that they still sounded somewhat robotic. In the case of CyberVoice, that issue is still somewhat present, but it certainly doesn’t sound robotic like the previous apps. Another problem is sometimes these apps may sound very monotone, and this is a problem CyberVoice also has. I hope in the future they can add something to tweak the tones of the voices, so you can
While I don’t think CyberVoice is going to replace your voice actors right now, I do see the potential. Who knows, maybe in the future this technology can be used as some sort of AI speech generator for games and other interactive applications.
Try CyberVoice: https://cybervoice.io/en