Question

Is it possible to set the tone or emphasis of a specific word or phrase?

Eg: Good Morning would be in a relaxed voice, whereas The nuclear plant is about to melt down would be far more urgent.

I'm aware I can change the speed of playback but is there something more clever I'm not aware of? Do other TTS engines support this (and are any of them open-source)?

Thanks

Was it helpful?

Solution

Are you using VoiceXML for the application? Regardless, if the TTS engine support the SSML standard you should be able to use the emphasis tag which will indicate that the TTS engine should put emphasis on the phrase within the tag. You are probably controlling the speed with the prosody tag which is also part of SSML.

Check out this online demo for Loquendo TTS. It provides more expression based on punctuation and allows you to insert items like yawns, sneezes and laughing. I think it is one of the better TTS engines out there, and Nuance must have thought so also because they just acquired them. It is not open source though.

Voxeo provides a free TTS engine on their Prophecy IVR which supports the emphasis tag. There are several open source TTS engines available but they will not be as rich or sound nearly as good as something like Loquendo. Just see if they support SSML.

Licensed under: CC-BY-SA with attribution
Not affiliated with StackOverflow
scroll top