Amazon Now Lets Developers Create Alexa Skills in 8 New Voices

Rivalling Google Assistant and Apple's Siri, Amazon is upgrading Alexa with the ability to speak in not just one or two but eight different voices. Developers can utilise the new development to easily give distinct voices to their Alexa skills. The new voices are available through text-to-speech service called Amazon Polly that the Seattle, Washington-headquartered company announced back in December 2016 as a part of the Amazon AI platform.

Developers looking to uplift their Alexa skills can use the new voices by applying for the latest developer preview. Unlike the original Amazon Polly portfolio that includes voices in a large number of international languages, Alexa has initially received eight US English voices. But, of course, the new addition brings a variety of male and female voices that you'd soon listen to through Alexa skills developed by third-party developers. Ultimately, the arrival of the new voices is expected to make the experience more engaging for users and will help developers create new models, including adventure stories and games with different voice profiles for each character.

The fresh change doesn't mean that developers weren't able to add new voices to their Alexa skills previously. There were techniques such as adding multiple voices using MP3 files. However, the process was difficult in nature as there weren't any pre-defined models to implement new voices.

But all this has been eased with Amazon Polly, which can determine the differences within various homographs and offers natural-sounding voices. It lets developers choose a different voice by using the Structured Speech Markup Language (SSML) and then specify which voice they would like to pick for their skills using the "voice name" tag. By this way, developers no longer need to use any audio files to add new voices and instead make the changes using the text.

The latest move transforms Alexa into a more compelling voice assistant model than before. It also makes the competition tougher for Google Assistant that was previewed with six new voices that are derived from Google's Wavenet at I/O 2018 earlier this month. Google also showcased the voice of American singer John Legend that is expected to hit Assistant later this year "for certain contexts".