Voice Design - The First Generative AI For Audio
The inaugural model for crafting synthetic voices has arrived
Just last month, we hinted that our generative model for voice synthesis was on its way. Now, it's made its grand debut as the pioneering solution of its sort – introducing Voice Design. This tool allows you to fashion novel voices from the ground up, defining key attributes such as gender, age, and accent. Intriguingly, even when the primary parameters remain consistent, our model infuses a touch of randomness with every generation, guaranteeing unparalleled voice uniqueness. Voice Design is a testament to our commitment to providing publishers and creators with the most adaptable AI-driven storytelling resources.
Try Voice DesignVoice Design
Voice Design's underlying model emerges predominantly from our deep dives into speech synthesis and voice cloning. However, we've always harbored a fondness for the concept of a generative tool tailored for speech. While generative text-to-image and chatbot solutions have found their practical niches, a corresponding instrument for audio remained elusive. Since our inception, there's been a recurring chorus of users requesting an expanded roster of speakers. Rather than bloating our library with innumerable voices, compelling users to sift through each one, we opted for a radical approach: empower users to craft speaker identities while ensuring limitless variety within those boundaries.
We felt the imperative to introduce an element of control in voice selection, recognizing our users' precise requirements for specific vocal traits in their scripts. The quest to guarantee uniqueness in every voice generated was paramount, considering the numerous scenarios where users desire — or at the very least appreciate — an exclusive voice. Beyond offering a fresh canvas for creativity, it's pivotal to note that voices birthed via Voice Design are wholly synthetic, bearing no ties to any living individual.
The Voice Design interface is remarkably user-friendly. All you need to do is choose a few vocal characteristics and tweak the accent intensity using a toggle.
Hear some of the voices we generated with Voice Design:
Applications
Beyond the seamless transformation of written content into high-quality audio with our renowned Speech Synthesis tool, book authors can now harness Voice Design to wield creative control over their narration, tailoring each character's voice to match their envisioned persona.
For news publishers venturing into the auditory realm, selecting the right voice is paramount. Since narrators inevitably become synonymous with the brands they voice for, this selection becomes a pivotal, seldom-repeated task. With Voice Design, publishers can instantaneously sift through and evaluate a myriad of potential narrators, ensuring that their chosen voice is exclusive to their brand.
Game developers are no longer faced with the dilemma of weighing character importance against recording expenses. Now, even the most peripheral of NPCs can be endowed with distinctive voices, enhancing the depth of in-game immersion.
Be it a content creator crafting their next masterpiece or a corporate leader voicing organizational announcements, the horizon for producing authentic and captivating audio tailored to specific purposes and audiences has expanded exponentially.
Ecosystem
Voice Design is just the beginning of several narration - enhancement tools we aim to unveil this year. Up next is 'Projects' our upcoming workspace designed for organizing extensive texts, integrating pauses, refreshing segments of audio, and designating various text portions to distinct speakers. Expect 'Projects' to launch by the end of March, followed by support for intonation adjustments in the latter part of Q2.