Step into the world of Qfrency text-to-speech (TTS), a symphony of technology orchestrated by the CSIR Voice Computing Research Group. Qfrency TTS is the first commercial text-to-speech engine that can synthesise speech in all eleven officially spoken languages of South Africa, with South African Sign language being the twelfth.  

Its ability to generate dependable and cost-effective human-like speech from digital text in the national languages at lightning speeds sets Qfrency TTS apart from other innovations in the marketplace. It silently empowers applications like screen readers, chatbots and interactive voice response systems. 

The story behind Qfrency TTS unravels amidst the intricate complexities of localising technology. The team embarked on an exploration of multiple open-source TTS engines primarily designed for the English language, only to encounter numerous impracticalities while attempting to synthesise text-to-speech in other languages. The stumbling blocks ignited the team’s inspiration to develop a TTS engine tailor-made for multilingual environments. 

The team discovered that software developed solely for research often falls short of meeting commercial objectives. The researchers adopted an unwavering commitment to creating top-notch commercial-grade software. 

The future chapters of this technological innovation include advances in voice cloning and cross-lingual voices. Ethical considerations will take centre stage when securing the ethical usage of voice recordings for machine learning models. 

According to CSIR principal researcher Aby Louw, one of the biggest lessons from the introduction of South Africa’s first TTS system for all national languages, is that procrastination has no place in the pursuit of progress. “Seizing opportunities, taking charge and relying less on external support is key,” says Louw. 

At the heart of this narrative lies teamwork – a collective effort by team members, contributing their unique talents to Qfrency TTS and creating a profound impact on visually impaired individuals. The future of the product now ventures into uncharted territory, exploring new applications and unlocking the vast potential lying dormant within African languages.