February 12, 2014 at 6:03 pm

CereProc’s ‘Human’-Voiced Music Player Read Personalized News, Tweets Between Songs

It’s easy to make fun of FM radio: no personalization, over-saturated with audio ads, prone to over-caffeinated “morning zoo” antics, and barely customizable at all. Still, it can be nice to get those news and weather snippets in between songs someone and/or something decided you should hear that morning, in that particular city or region.

What would the Web 3.0 (or whatever we’re on at this point) version of that look like?

Surprisingly, we haven’t seen anything like that take off, even though APIs and feed technologies make it super simple to grab news items, weather reports, and even your Facebook and Twitter updates, and read them in between the songs of whatever radio station you’re listening to. (An advanced feature might include news stories that go with the mood or subject matter of the song that just played.)

That’s why we were excited to hear from the UK-based CereProc, makers of the free MyMyRadio iOS app, currently available in public beta form. CereProc models human speech to do text-to-speech, including company did previously modeled the voices of film critic Roger Ebert and former NFL player Steve Gleason, when they lost their abilities to speak.

MyMyRadio plays any of the music locally stored on your iPhone, using an algorithmically-generated Irish woman’s voice to read news snippets from whatever categories you select, as well as your Facebook and Twitter feeds. There’s no local weather (or even better, weather alerts), but there could be.

The app’s Twitter integration was a bit buggy, during our testing — it made me sign in with my Twitter credentials three times before the first song ended, then the app froze, and when it restarted, I had to enter my Twitter credentials again. Also, we didn’t hear our Twitter and Facebook updates or news summaries read aloud in between songs, even when we upped the frequency to five feed items per song.

Also, it cuts in rather abruptly, when you manually play a news item, because your music doesn’t fade out (again, this is supposed to be a public beta). And then, we kept hearing the voice read the same NYT item about Libya, ringing out from our pockets, even though no songs were playing in the app.

All of that stuff can be fixed. So, how does the voice sound? Here’s MyMyRadio reading a summary of a New York Times article:

Again, this app only plays local music — it would be nice if it streamed radio and on-demand music, too, or integrated with something that did. There’s no reason CereProc or some other company couldn’t slap this or something like it on top of a Spotify or Rdio app, or a streaming radio service, and come up with something compelling. After all, the human voice is the only thing many of us miss about college radio.

The right kind of API or widely-available text to speech technology could do amazing things for music and other audio apps (podcast players and navigational systems, in addition to, of course, just about any kind of music app.

We asked CereProc whether it plans to license the technology, and CereProc’s Francesca Shaw responded, “Licensing the app is definitely something we’ve been thinking about.”

Follow Evolver.fm’s story feed on Facebook or Twitter