Text-to-speech (TTS) On Fedora

Scott has been setting up our OpenHAB server, and the latest project was controlling our network speakers. You can play Internet radio stations to the speakers, you can stream music from the NAS … but we also want to be able to play announcements. For that, we needed a text to speech engine.

Festival is in Fedora’s yum repository, but everything I’ve read about Festival says the output is robotic. Which is likely fun at first, but tiring after the first three or four times. Even if you have it say “beep, boop” at the end.

SVox (Nuance, which a long LONG time ago was spun off from Stanford Research Labs) has an open-source version of their text to speech product. Not in convenient package form, but close. Someone maintains a shell install script. Download the script:

wget https://raw.githubusercontent.com/stevenmirabito/asterisk-picotts/master/picotts-install.sh

Then read through it and make sure it’s not doing anything untoward. It wasn’t. Ran the script and a minute later, we can use “pico2wave -w /tmp/ljr.wav “I am your TTS engine”

Quick. Easy. And now we’ve got a wave file to send to the speaker (and remove when we’re done!)

Leave a Reply

Your email address will not be published. Required fields are marked *