HTXT.AFRICA

SpeechKit and the rise of text-to-speech

When most people think of synthetic speech their mind usually goes to one of two places: as a tool for those with a sight impairment or the voice of your local AI assistant. The former doesn’t apply to most people and the latter is more of novelty, and so this interesting technology is ignored by the masses.

But outside of the spotlight things have been moving along rather swimmingly, with SpeechKit at the forefront with an innovative solution for speech-to-text. SpeechKit achieves naturalistic voice that sounds more human than machine thanks to Natural Language Programming (NLP) modules and Speech Synthesis Markup Language (SSML).

One of the best sources to illustrate this is Amazon and its Alexa smart home assistant being one of the most recognisable synthetic voices on the planet. Amazon reports a massive uptick in interest for this field with one of its researchers call it a “gold rush”.

Amazon also points to its Polly system, which allows individuals and companies to create voices (or “personas”) which fit their brands. We also love that Amazon chose “Polly” as the name here with it being the stereotypical name for parrots that talk.

Speechkit makes use of Amazon Polly and augments it to improve on the work and make speech-to-text that is more localised and better suited for individual countries and territories. Are you tired of hearing vague US and UK accents when listening to synthetic voice? SpeechKit is looking to change that with a focus on local pronunciation to nail certain names. In its explanation of Polly augments SpeechKit uses South Africa’s president Cyril Ramaphosa as a name that may be difficult for international software to properly deal with.

This makes sense as, while SpeechKit is a UK-based company, its co-founder and CEO Patrick O’Flaherty is from South Africa. More than that local companies such as News24 and Daily Maverick are already making use of SpeechKit to turn regular written articles into spoken audio.

Sticking with the UK, SpeechKit reports that 24.8 million people listened to online audio in 2018. It really seems that, going into 2021 and the future, sitting down to read the news, or a blog or newsletter is the old fashioned way of doing things.

We’re continually living in an age where you need to offer more to the reader than your competitor, and options such as text-to-speech seem like the perfect way to do it, allowing the automatic retrofitting of site libraries into new audio content. This opens up a vast new reader base with minimal investment as well as access to those who may not have access your content otherwise. And, maybe the best thing about it is that SpeechKit has already done all the hard work.

It’s clear that text-to-speech and its various applications is on the move and growing every day. If you want to catch this wave and work with SpeechKit, you can try it for free thanks to a 7-day trial. For those in the enterprise space who want extra features, 24/7 support, API integration and more, contact the company to arrange a demo.

Exit mobile version