2023 Speech Industry Award Winner: ReadSpeaker Embeds TTS in Many More Platforms
With almost 25 years of experience in developing text-to-speech solutions, ReadSpeaker today offers one of the largest selections of expressive, humanlike voices in the industry. Its roughly 200 voices supporting dozens of languages are available via plugins or software development kits for use in any application or device, embedded, on premises, or in the cloud.
The company, based in the Netherlands, has speech-enabled more than 10,000 applications worldwide using its next-generation deep neural network (DNN) technology to structurally improve synthetic voice quality for more natural and engaging conversational experiences. Its portfolio of web-based TTS solutions spans websites, mobile sites, mobile apps, RSS feeds, online documents and forms, consumer electronics, and more, and the list is constantly growing, as we saw again this year.
Central Japan Railway, for example, implemented ReadSpeaker’s TTS software for announcements about its train approaches, arrivals, and departures at platforms and concourses along its Tokaido Shinkansen, one of Japan’s busiest and most popular bullet train lines. Using speechMaker Desktop, ReadSpeaker’s artificial intelligence voice authoring tool, the railway operator can generate TTS travel announcements automatically in English and Japanese.
ReadSpeaker also lent its Dynamic Runtime Neural TTS plugin for the Unity and Unreal gaming engines to Microsoft’s Xbox platform, enabling developers to embed AI-driven runtime TTS voices into any game.
ReadSpeaker’s Deep Neural Network (DNN) TTS engine allows neural TTS to run as a stand-alone solution locally on the device for in-game narration, audio cues, dynamic non-player character conversations, and more. If game developers prefer to work in an engine other than Unity or Unreal, ReadSpeaker can provide the same high-performance neural voice capability for other custom game engines through the ReadSpeaker speechEngine SDK. Developers can add custom synthetic voices based on a voice actor or narrator and produce humanlike speech that expresses a range of emotions, including joy, sadness, and anger.
The Game Engine plugin can generate dynamic, runtime neural speech synthesis in more than 30 languages.
“Integrating ReadSpeaker’s text-to-speech plugin took around 20 minutes and automatically gave a voice to each line of dialogue created within my character engine. There was a lot of nuancing to get the right tone, timing, and intonation, but essentially Acolyte went from a text-only experience to a fully voiced game with almost infinite lines of [voice-over] in less than half an hour,” said Jamin Smith, director and lead designer at Superstring, in a statement.
ReadSpeaker also integrated its TTS with Moodle Workplace, a learning management system. With this partnership, employees going through training can listen to content directly within the Moodle Workplace LMS anywhere, any time, and on any device, and they can select from hundreds of voices in nearly 70 languages. They can personalize the reading speed, select highlighting, font, text size, and color, and also use features to help them concentrate more easily, such as page masking.
“As a global LMS platform, Moodle recognizes the growing need for enhanced textual support across all aspects of education and training,” said Carles Aguiló-Collado, technology partner manager at Moodle, in a statement. “ReadSpeaker’s text-to-speech and additional tools add to Moodle Workplace’s functionality and play a vital role in fostering effective communication during the training process. They ensure that learners remain engaged and comprehend the intended message, allowing all subsequent activities and training efforts to build upon a solid foundation of understanding.”
And when Sonos, the audio systems provider, launched a French-speaking version of its voice assistant, the synthetic voice was created by ReadSpeaker using its deep neural networking models. French voice talent participated in recording sessions to train the AI.
“We collaborated with ReadSpeaker to create the French voice for Sonos Voice Control. Their expertise in AI-driven text-to-speech and their approach to personalized voice is both innovative and extremely thorough, which is exactly what we needed for this project,” said Sonos’s vice president of voice experience, Joseph Dureau, in a statement. “The collaboration with ReadSpeaker’s VoiceLab linguistic team has been essential in support for us throughout this project, from the development phase to the launch.”