The Pulitzer prize-winning movie critic, Roger Ebert, lost his voice to cancer several years ago. He is one among many thousands of people a year who lose their ability to speak from disease or injury. There are some technology fixes for replacing the physical reproduction capability. (See SciTechStory: Replacing the larynx with a palatometer) However, for Ebert and many others, physical repair is impossible. His hope for a voice lies almost entirely with digital technology – artificial voice production. Although Ebert knew about computers producing voice, like most of us, his recollection of these voices is of the odd off-human sounds heard on automated telephone menus. Many years ago the physicist Steven Hawking began using a computer driven voice producer, which (if you’ve heard any interviews with him) is the typical squawk-box. Not for Ebert; he reasoned that Hawking was stuck with ‘ancient’ technology – there must be something better and he began ‘moseying’ (his word) around the web to see what he could see.

His searches brought him to the web site of CereProc, a company in Edinburgh, Scotland that specialized in computer generated voices that sound like the person (used to sound). Ebert’s account of what happened after that is in his online journal. Here’s a sample, but it’s well worth reading the entire piece:

Anyway, CereProc didn’t need to hear me speaking a specific word in order for my “voice” to say it. They needed lots of words to determine the general idea of how I might say a word. They transcribed and programmed and tweaked and fiddled, and early this February, sent me the files for a beta version of my voice. I played it for Chaz, and she said, yes, she could tell it was me. For one thing it knew exactly how I said “I.”
…
CereProc is now blending in my audio snippets for “Casablanca,” where I sound enthusiastic, and “Floating Weeds,” where I sound calm and respectful. It’s nice to think of all these great movies sloshing around and coming out as my voice.

What will I use this voice for? I could talk with Chaz and our grandchildren — and it would be me, not Alex. I could do audio for Webcasts, talking under clips from movies I’m describing. I could do radio. I could tell jokes. Chaz and I are producing a new movie review program for TV, but I won’t be one of the two critics.

[Source: Roger Ebert – Hello, this is me speaking]

This does not happen to everyone. Not that only people the stature of Roger Ebert can get help like this, but leading edge technology does not come cheap. Even Roger Ebert probably couldn’t have afforded the R&D that went into his voice production. But in some ways that’s not the point, not at the initial use of a technology. The point is to show what can be done, and when Roger Ebert steps out onto the set with Oprah (Tuesday March 2) it will show millions of people just that. In short, at this point the technology is inspirational.

Next, it needs to become commonplace. Unfortunately, there is a lot of technology – equipment and programming custom made to solve the problem of a particular person (or a few people), but hopelessly beyond scaling (that is, manufacturing on a large scale). Some of the time the dead-end is a matter of economics. If it’s not profitable, it never becomes commercial and widely available. In some cases, the basic technology gets the idea across, but is too big, clumsy, or expensive for anything except a demonstration. New, smaller, lighter, better, cheaper technology has to come along before some of these ideas become commonplace.

Will the type of voice production used by Roger Ebert become commonplace? It has a better chance with him as unofficial spokesperson (pun intended). In any case, now Roger Ebert has both his voices: The inimitable one of his written word, and an authentic spoken voice to let him speak for himself.