How Roger Ebert Will Get His Voice Back

Years of battling cancer have left film ubercritic Roger Ebert without a portion of his jaw, and consequentially, his voice. Esquire's superb profile outlined his efforts to regain a voice—his voice—but left us wondering: How will that work?

From Esquire:

Ebert is waiting for a Scottish company called CereProc to give him some of his former voice back. He found it on the Internet, where he spends a lot of his time. CereProc tailors text-to-speech software for voiceless customers so that they don't all have to sound like Stephen Hawking... CereProc is mining Ebert's TV tapes and DVD commentaries for those words, and the words it cannot find, it will piece together syllable by syllable. When CereProc finishes its work, Roger Ebert won't sound exactly like Roger Ebert again, but he will sound more like him than Alex does.

CereProc is headquartered on the sixth floor of an imposingly ugly tower in Edinburgh, Scotland—I'd know, because it was in the five floors below that I spent most of my undergrad career at the University of Edinburgh, which owns the building. By all counts, it's a small operation, and a relatively new one, started about five years ago.

But I checked into CereProc's work online, and their sample voice sets speak for themselves, so to speak (sorry and sorry!): Obama sounds like a slightly more hesitant Obama; Arnie, whose verbal tics are his trademark, sounds like almost exactly like Arnie. (You can listen to both, and others, here and here.) Listening to what they can do with publicly available voice data sets is heartening, so the prospects for a man with such a broad catalog of vocal recordings, from radio broadcasts to his TV show to old podcasts, seem fantastic. He felt the same way when he discovered CereProc back in August:

I have my fingers crossed. I have launched an e-mail to Edinburgh with my appeal. I can see my own voice hosting online or telecast video essays. I am greatly cheered.

CereProc got that email, and answered his appeal. So!

The first step would be a desktop software system, which would dictate text in the same way that Mac OS does natively. While this would be a great bridge, but a mobile solution would really change things. CereProc's software is licensable for just about anything, and has already been incorporated into an iPhone app, albeit for simple news dictation.

In other words, a voiced Ebert is something the he (and we!) can realistically look forward to. And by this, we too are greatly cheered. [CereProc, Esquire]