A bit like a search engine for voices, Vocalid users can find and download their own speech tools from the startup’s massive database of global voices. And you can help.
“There are still relatively few voices that exist in the text-to-speech space,” Vocalid founder Rupal Patel told Digital Trends. “It’s not uncommon to have a scenario where a little girl is using an adult man’s voice, for example. One of the people we work with has ALS, and he told me about going to an ALS fundraiser with his wife. Virtually all the men there were using the same voice, and it made it impossible for my wife to find him if they got separated because she kept hearing his voice coming from everywhere. For most of us, this kind of thing doesn’t affect us. The way that we use text-to-speech is for information transmission. We want to get directions for GPS, or to find out which stop to get off on the train. It’s not your avatar or your voice that’s being conveyed. That’s the difference.”
The voices that Vocalid create aren’t completely human sounding. They are synthetic composites of speech parts that are recorded from users around the world and then stitched together to create fully digital voices. But the technology is getting better all the time, and — most importantly — there is enough variety to make sure that users can pick a voice which sounds appropriate to them.
Customers find them by logging onto Vocalid’s system, uploading short vocal sounds (even short utterances like “ahh” are enough to perform a search) and locating ones which match their tone and accent.
This is where you come in. At present, the Human Voicebank has more than 14,000 members all around the globe, each contributing millions of sentences to the platform. However, Patel says they’re always looking for more and this can be done for free by any user willing to give up a short amount of time to read a few sentences into a microphone from their home computer.
It’s an amazing project and the idea that your act as a “voice donor” could transform someone’s life for the better is exactly the kind of awesome use-case which makes us glad to be living in the digital age.
Now get talking!
- How to set up speech-to-text in Windows 10
- TiVo takes personalization to the next level with voice ID
- How artists and activists are using deepfakes as a force for good
- His granddad lost the ability to read, so he built a DIY text-to-speech rig
- Twitter tests Spaces, an audio-only chatroom