The best voices generated with AI technology!
THE BEST VO + LEGAL AI
A 100% legal tool to assist audio/video professionals.
VO+AI PROJECT
VO+AI is a professional voiceovers and Poland’s first online voice bank Mikrofonika partnership project. The first tool to combine the possibilities and experience of popular and well-liked voices with the advantages of artificial intelligence.
We present to you a unique system that allows the use of voiceovers well known to audiences from radio, television and the Internet. It is based on legal sources, fully safe for both sides of the partnership. Each model is fully autonomous, created from scratch, preserving the individual characteristics of each talent. Voice timbre, intonation, interpretation and even breaths – a possibly faithful copy of our best voiceovers.
Introduce AI-generated recordings to your customers, based on the perfectly familiar sound they’ve been trusting for decades.
Listen how VO+AI sounds
BLOG - Step by step
We trained a “studio” class voicebot for Anita Maroszek. How does it differ from a standard ‘basic’ class avatar? In a nutshell, …
The day has finally come. We are ready to show you how we are creating voice avatars of the best voiceovers! We …
The 8th National Voiceovers Meeting, organized by the Mikrofonika voice bank, is the best place to present our technology. A year ago, …
Ireneusz is the first voiceover artist in our voicebase for whom we have trained a ‘studio’ class voicebot. How does it differ …
Ethics and Legality
The development of AI technology carries a special responsibility and presents us with new challenges. The world is changing, but there are certain principles that remain constant. Each person’s voice is their exclusive property, just like other distinctive characteristics of every voiceover: interpretation, accentuation, phrasing, and emission patterns developed over years of work.
It is no secret that technology companies do not respect the basic rights of people whose voice appears in public. The companies use digital resources without the consent of the publishers, broadcasters and voiceovers – the owners of these recordings. This is bound to create legal consequences.
The VO+AI project shows that it is possible to reconcile the interests of all sides and present a product that is 100% compliant with the law and basic principles. Every voiceover, whose voice we have used in this project, is sure that he will have full control over its use, and any profits from the use of the resulting voice model will be shared fairly. This is guaranteed to the talent, thanks to the clear and transparent contract.
We consult our development philosophy with the voiceovers themselves, presenting it at the National Meeting of Voiceover Artists. Our agreements have been presented to the Voiceover Association of the Republic of Poland, ZASP and the Professional Union of Dubbing Artists.
VO+AI is not intended to replace voiceover with technology. Its purpose is to facilitate clients in choosing the right performer; assist editors in correcting tedious errors; make producers aware of the differences and multitude of options available. It is also designed to help voiceovers when they are unable to work, saving time and increasing profits for all market participants. Co-create better productions.
Our partners can rest assured that we will advocate for and protect their rights, stand up in their defense and safeguard their voices against any unauthorized use.
Technology
In a nutshell, our project allows for the creation of an accurate digital copy of any voice. Based on recorded samples, we use AI to analyze the sound, articulation, interpretation, phrasing, and breathing patterns of individual reader. We strive to faithfully reproduce these variables when generating speech sounds from a given text file. In English, this is referred as TTS (text-to-speech).
Our work is based on publicly available technologies. We use MIT-licensed solutions such as GradTTS and HiFi-GAN, among others.We modify and improve the proprietary engine in an effort to get the best possible results. Our voice models are continuously improve and expanded with new recordings. All of this takes place within a closed, secure local network, and no materials are sent to external company servers.
The work of our team is led by Marcin Hajdukiewicz, one of the pioneers of TTS technology in Poland. He has been developing solutions based on AI technology since 2021.
The project is supported by
