VOICE CLONING

NEXT GENERATION OF SYNTHETIC VOICES

USING ARTIFICIAL INTELLIGENCE

VOICE CLONING

INTRODUCTION

Automatic voice cloning aims to generate synthetic voices very similar to an original voice. Based on deep-learning techniques, this technology takes advantage of a set of audios of the original voice in order to train a model capable of generating new audios that sound alike.

Our voice cloning technology is language and gender independent, and can be quickly and easily customized to a specific or familiar voice.

Nowadays, voice cloning is being increasingly demanded by the market due to its interesting and varied applications, which include conversational assistants, smart speakers, dubbing, digital characters, gaming, audiobooks, navigation systems or voice branding. This technology helps automate and personalize many tasks carried out in these type of applications and domains

FEATURES

REALISTIC VOICE CLONING

Generation of high quality and realistic voices using deep learning techniques..

FAST DECODING

Fast generation of voices to operate in real time scenarios and natural interactions

CUSTOMIZATION AND ADAPTATION

The technology can be customized to generate the appropiate voice style for each application and domain.

API FOR DEVELOPERS

Developers and companies can easily create and integrate the voices in their products and solutions.

MARKETS

SMART ASSISTANTS

Voices of smart assistants sound increasingly natural due to the advancements in the AI field of text-to-speech technology. Voice cloning allows their personalization, through the use of specific or favorite voices to develop customized conversational assistants.

DIGITAL DUBBING AND ANIMATION

Voice cloning makes it possible to create identifiable voices for unique digital characters or avatars in human-computer interaction systems or for the production of multimedia and audiovisual contents..

VOICE BRANDING

Use your personal and particular voice as the main brand of your company within your conversational assistants and marketing activities.

AUDIOBOOKS

Take the opportunity to synthesize millions of audiobooks without human participation and including your favourite voice.

VOICE EXAMPLES

“Batxilergoa amaitu ondoren, medikuntza ikasketak egin zituen”

REAL VOICE

CLONED VOICE

SYNTHETIC VOICE EXAMPLE

¨Kapitulu bakoitzak bi zati ditu, meditazio-ipuina lehenbizi eta ipuin-jolasa jarraian.¨

¿Crees que ha renunciado a una cubertería de plata, candelabros, bandejas y fruteros?

REAL VOICE

CLONED VOICE

SYNTHETIC VOICE EXAMPLE

¨A pesar de que Estados Unidos haya levantado durante tres meses el veto a Huawei, la marca china se enfrenta ahora a un reto gigante.¨

¨Igualment, la regulació de la conservació de la història clínica permetrà resoldre casos que, a hores d’ara, van arribant d’un en un.¨

REAL VOICE

CLONED VOICE

SYNTHETIC VOICE EXAMPLE

¨Es va acostant l’estiu, i això es notarà clarament entre avui i demà.¨

 

“The invention of movable metal letters in the middle of the fifteenth century may justly be considered.”

REAL VOICE

CLONED VOICE

SYNTHETIC VOICE EXAMPLE

¨Food production and manufacturing are among sectors allowed to be open.¨

SUCCESS STORIES

This example of voice cloning constitutes a great scientific-technological challenge that Vicomtech has solved using artificial intelligence technologies applied to speech processing. XRey is an exclusive Spotify podcast produced by The Story Lab that reviews the lights and shades on the figure of the king emeritus. Franco’s cloned voice has been generated exclusively for two key moments in the script and the bonus track, which explains how its creation process has been.

SPECIFICATIONS

Module Description Platforms
Audio Preprocessing Audio normalisation, enhancement and transcodification. Machines & servers Linux
E2E synthesis Deep learning based E2E technology for the generation of voice synthesis models Machines & servers Linux
API for AI Interface to easily synthesize audios or train your own models. Machines & servers Linux