Voice Demo

Voice Demo

Explore the most natural and diverse text to speech voices on the market!

X Close

Voice Creation FAQs

Q: How do I go about banking or cloning my voice with CereProc?

A: CereProc offer a voice cloning service - CereVoice Me. There is more information, together with samples generated from a CereVoice Me voice on this page

The recordings for CereVoice Me can be carried out at home, provided the user has a suitable quiet, non-echoic room to record in. Users need an up-to-date Window or Mac laptop or desktop, running the most recent version of Google Chrome or Firefox. CereProc provides a headset and mic to carry out the recordings and this is included in the price.

In English, there are 619 sentences to record for CereVoice Me. In user testing, we found it took university students 1.5 - 2 hours to complete the recordings. Users can stop at any point and start again later from the same point.

At the end of the recording process, CereProc delivers a voice that can either be installed on Windows SAPI5 or MacOS. Voices install as system voices that speech-enabled applications can use. There is a wide range of applications, including AAC (Augmentative and alternative communication) apps that can utilise a CereVoice Me voice. Please contact our support team for further information on this.

CereVoice Me is currently available in English, Spanish, French, Swedish, Italian, Polish and Romanian

Q: I no longer have the power of speech but I have an archive of recordings of my voice. Are CereProc able to build a synthesis voice from my recordings?

CereProc are able to build voices from archive recordings and have done so using recordings of well-known figures such as JFK and Roger Ebert and recordings made by private individuals. In order to build a voice, CereProc requires a collection high-quality, preferably studio-quality recordings and if they are available, transcriptions of the recordings. If possible, the recordings should be free of background noise, interjections from other speakers and disfluencies. Generally speaking, the better quality the recordings, the better quality the synthesis voice will be.

If you wish to build a voice from recordings you believe to be suitable, please contact us to discuss, via our contact form As a first step, CereProc would undertake a review of the data to determine whether building a voice would be possible.