vidby, a Swiss startup offering AI-powered software for fast and accurate video translation and YouTube’s recommended vendor, has announced the cross-lingual voice cloning tool. The new feature will be helpful to users who need to translate and dub videos with their own voices.
To utilize the new feature, the users will be required to submit a voice recording of at least 5 minutes in any language when placing an order through the platform. The training data requires clean audio files from a single speaker, devoid of background noise, music, or other effects. Once the sample is uploaded, vidby’s advanced processing generates the digital voice replica within minutes. Afterwards, the voice sample can be used for video dubbing. Presently, the platform supports cloned voice dubbing for 28 languages. They include some of the most commonly spoken ones, such as English, German, Spanish, Mandarin, and Hindi.
“You’ll find plenty of voice cloning solutions in the market for the same language, but when it comes to cross-language cloning there’s a scarcity of them because it is quite a challenge. Each language has its own unique sounds that might not exist in other languages, so poor interlingual cloning can lead to a thick accent. This isn’t just a machine problem; humans face it too. Think about when someone speaks a non-native language. That’s why nailing interlingual cloning isn’t just about mimicking the voice’s tone and sound, but also about mastering the language’s pronunciation without that distinct accent,” explains Alexander Konovalov, CEO and co-founder of vidby. “You can easily dub videos in multiple languages with the necessary voice without compromising the quality. All you need is access to our platform and an audio recording of the speaker. We take care of the rest.”
With this upgrade, vidby aims to provide its customers with an essentially limitless library of voices and to democratize cutting-edge technology, making it affordable for everyone. Digital voice creation is priced at a reasonable $10 USD, while voiceover services using the cloned voice are priced at $3 USD per minute.
To utilize the feature, the user must confirm that the provided voice belongs to them by pressing the confirmation button. This restriction reflects vidby’s ethical approach to the use of AI. Once generated, the voice is securely stored within the user account, exclusively accessible to the creator.