Blend and Morph Multiple Voices in Real Time with Vocoflex from Dreamtonics
Trailblazing Tokyo-based music tech start-up Dreamtonics, is excited to announce the full release of Vocoflex, its real-time, AI-powered voice creation and transformation plugin.
Equipping creators with three bundles combining Vocoflex, their flagship Synthesizer V Studio Pro, and a variety of realistic-sounding voice databases, Dreamtonics provides state-of-the-art AI-powered voice synthesis tools to the community while prioritizing their ethical development and use.
“Singing voice conversion (SVC) technology has been around for a few years now. As a technical person, I am genuinely excited by the technical advances, but at the same time feeling disappointed and betrayed that more often than not, it’s only used to create (often ethically questionable) covers, and fails to produce anything original or make a lasting, positive influence on modern-day culture,” admits Kanru Hua, CEO of Dreamtonics. “This untapped potential is why soon after releasing Synthesizer V, we decided to move toward the SVC space.”
While Synthesizer V originated from Vocaloid culture in Japan, it has gained mainstream adoption. In a similar way, Dreamtonics counts on its next product Vocoflex to bring ethical SVC to the creative masses.
Using some clever statistical tricks to map multiple voice characteristics onto a 2D space, the Dreamtonics team found a way to interpolate between multiple models to create entirely new voices. Reducing the latency of the rendering unlocked real-time voice conversion. Like simultaneously DJing and singing a karaoke duet, a single performer can tap into several voices during one live performance using Vocoflex. For a more tactile experience, Vocoflex even allows users to connect and map their MIDI controller’s buttons, knobs, or sliders to adjust their voice.
While Vocoflex and all associated bundles are designed for creative applications, Dreamtonics understands the potential risks and ethical implications of its use beyond these intended purposes.
The voice morphing tool allows users to train their own AI voice models with as little as 10 seconds of source audio and operate the model on their own local computer — drastically reducing the time it takes to render new voices and effects. To prevent unauthorized usage of such a powerful tool, Vocoflex requires ID verification through Dreamtonics’ KYC-service partner and incorporates an advanced tamper-proof audio watermarking feature.
Undetectable to the ear, this watermark embeds a user-unique license ID into the voice, enabling Dreamtonics to identify both the use of Vocoflex and the creator of the audio sample, deterring users from importing voice samples without the singer or users consent — a practice prohibited by their usage terms.
“Dreamtronics adheres to ethical AI usage, never blurring any moral and legal boundaries, and encourages users to use Vocoflex to expand their creative possibilities with the artist’s, actor’s, vocalist’s or user’s full knowl
