Create a Perfect Digital Copy of Your Voice and Speak the Languages You Don’t!

We’re proud to release our Professional Voice Cloning (PVC) model to the wider public. Formerly available exclusively to our enterprise clients, we're now opening access at large, allowing you to create a perfect digital copy of your own voice; one that’s virtually indistinguishable from the original.

We enabled users to upload their voice data back in March and we promised to release the voices on a first-come, first-serve basis in July - they’re finally here.

PVC comes free to everyone on or above the Creator plan. What’s more, your PVC voice can also automatically speak all the languages supported by Eleven Multilingual v1!

The Process

To access PVC, simply go to VoiceLab, click on “add a new voice”, and select Professional Voice Cloning. Unlike our Instant cloning feature, PVC involves training a dedicated model on a large dataset of voice samples - 30 minutes minimum, with 3 hours being optimal.

To ensure the highest-fidelity output, there are a few things to keep in mind as you prepare your samples for fine-tuning.

Firstly, make sure the training data comprises clean audio files of a single speaker with no background noise, music or other effects. Any non-speech sounds may confuse the model and find their way to the output. If you upload multiple audio files recorded in separate sessions, make sure to match the recording conditions as closely as possible - noticeable differences in reverb or distance from the microphone etc. will likewise pollute the output.

The same is true of your speaking style - your style of delivery should be uniform across all the samples you upload. For example, if you’re planning to use your voice to record an audiobook then the training data should comprise recordings of you reading in audiobook delivery style.

PVC comes integrated with all our models, including Eleven Multilingual v1. If you speak any of the languages it supports, you can create a perfect replica of your voice and have it speak all the other languages, too!

We run the model at least once per month, depending on the number of requests, with further speed-ups expected towards the end of the quarter.

Safety

To ensure safe use of our technology and maintain strict user privacy and ethical guidelines, we’ve integrated robust security measures to make sure you can only clone your own voice.

Once you upload your speech data for training, a verification step follows. In it, you’re provided with a text captcha prompt. You are then asked to read it aloud within 10 seconds. We validate your request by comparing the voice profile from this recording with the voice contained in the data you uploaded.

If there’s a match, your request is sent for fine-tuning. If not, you have 4 verification attempts remaining. If they are all invalid, you’ll have to reach out via our help center to have your voice verified manually.

Unless you decide to share it, your voice belongs and is available only to you.

Applications

Professional Voice Cloning extends beyond simple convenience and offers a range of advantages for your personal and commercial projects. Here are just a few:

• Content Creation: Content creators can deliver their message even when they can't record in person, meaning no more disruptions in content schedule.
• Audiobooks: Clone your voice and use Projects to narrate an entire audiobook in your own delivery style, regardless of the book's length or your available recording time.
• Digital Presentations: Use your cloned voice to deliver a compelling, consistent narrative and make yourself a part of your presentations, even when you're not physically present.
• IVR Systems: Businesses can provide a more personal touch to their customer interactions by using their staff’s voices in their automated responses.
• Podcasts: Podcasters can maintain their show's schedule even when they're unable to record.

Your New Digital Self

Professional Voice Cloning allows for more control in how you represent yourself digitally. It's more than voice replication - it's an opportunity for convenience and consistency across a variety of audio use-cases!