Text to Speech and Ham Radio

I use text to speech engines (TTS) in a number of ham radio applications.  When you are dealing with voice communications and repeaters, it is handy to announce the time, IRLP, EchoLink, Allstar connection status etc.

Heck even as things move digital, I use them.  Lets face it, when mobile listening to your radio is a lot safer and convenient than looking down at the display.

Linux comes with the Festival speech synthesis system, but by today's standards it really doesn't sound very professional.  Sadly Festvox, Flite, and Espeak all sound about the same.

Cepstral was founded in 2000 by leading scientists from Carnegie Mellon University, where Festvox/Flite was developed.  And it sounds quite a bit better.

Cepstral isn't open source, but they have Linux builds available.  Early on, the cost for a license of Cepstral was only $30.  In 2013, when version 6 came out, they started using the new licensing scheme which made it impractical for hobbyists.

About this time some people figured out how to use the Google Translate tool to generate instant text to speech.  Many people switched to this.

Sadly just recently Google deprecated free access to the language translate api. (There may be ways around this, but for how long?)


I wish someone would create a kickstarter campaign or something like that to take up a collection to get someone to contribute an updated open source TTS voice as part of the standard Linux distributions. Festival/flite are all pretty rudimentary sounding.

Recently our local National Weather Service Office announced "that for the first time in over 15 years, the NOAA Weather Radio system is getting an upgrade. A new computer system will be used to generate audio recordings of the forecasts, observations, watches, and warnings which are routinely heard over the air. The actual broadcast programming will remain unchanged, and there is nothing you need to change on your radio to continue receiving warnings, watches, and forecasts from the National Weather Service.  But listeners may notice that messages are being read in a different voice."

Wouldn't it be nice that when the government spends money on software, they would put those software projects out to bid and tack on an open source requirement?  That way the taxpayers would benefit in more than once way from how their money is spent.

I read a while back that Russian Prime Minister Vladimir Putin ordered government agencies there to transition to open-source software.  And something about a government funded national repository for open-source software.