By Stephanie Ciccarelli
April 18, 2011
After losing the use of his voice following surgery for jaw cancer several years ago, media personality and film critic Roger Ebert has found new ways to communicate and be heard.Through journal entries online and a TED Talk, Roger Ebert has been able to share more about what happened to him and how he has embraced technological advances to regain his 'voice.'
Learn more in today's VOX Daily.
Most people take their voices for granted until they've temporarily lost use of their instrument or cease to speak entirely. April 16th marked World Voice Day, and in light of that, I'd like to take a look at the human voice this week from a number of angles.
When famous American film critic Roger Ebert lost his voice after a tracheostomy almost four years ago he was determined to still communicate audibly, even if it meant having another voice speak for him.
A couple of avenues he pursued were the designing of a text-to-speech version of his own voice and the use of existing text-to-speech personalities for purchase on the web.
CereProc Limited, a company based in Edinburgh, Scotland, leads in the synthetic speech development field. The company was able to craft the voice for Ebert by sampling speech from old recordings over the years ranging from debates with Gene Siskel to more recent programming. From what was shared in Ebert's TED appearance released earlier this month, the development and finessing of "Roger Jr." is still a work in progress.
The video below is a snippet filmed when Ebert and his wife, Chaz, appeared on Oprah. One can only imagine what it would have been like for her to hear her husband's voice for the first time through the text-to-speech medium.
A British TTS persona named "Lawrence," served as Roger Ebert's temporary voice. On his blog, Ebert boiled his casting process down to the following, "I tested the demos, and settled on Lawrence, not because he has a British accent but because he was the easiest to understand, and sounded intelligent." After a while, his wife began to refer to him as "Sir Lawrence."
Apple introduced a new text-to-speech voice named "Alex" when releasing Mac OS X 10.5 Leopard in October of 2007. Ebert decided to change TTS voices at this time and cited Alex as being his favorite voice and also recognized Apple as the source of the best text-to-speech solution meeting his needs.
To sum up, Apple says that Alex:
- delivers natural intonation in English at extraordinarily fast speaking rates
- analyzes text a paragraph at a time and deciphers content with greater accuracy
- closely matches the nuances of human speech for better comprehension of the text
- breathes between long passages
Since Alex's initial release, Apple has released other voices and according to AppleInsider.com plans on releasing dozens more in a variety of languages in the near future for the Mac OS X 10.7 "Lion."
Text-To-Speech has come a long way. You can even use AliasKeys or other text insertion shortcuts to add emotional intonations to TTS voices that program them to say things in more meaningful ways.
While it may appear that TTS is taking jobs away from voice over talent, the technology is not yet so advanced as to be considered on par with a custom read or inspired delivery. From what I've heard, some of these jobs take 50+ hours of voice over work so at least someone out there is getting paid well when giving voice to a persona such as the Alex voice by Apple or the Lawrence voice from Cepstral.
I welcome your comments and invite you to join the conversation now.
StephanieRelated Topics: Accent, Alex, Apple, Chaz Ebert, Roger Ebert, SAG, text-to-speech, TTS, voice, YouTube
Growth is the new business imperative, and that means seeking new markets wherever they may be. Before you cross the border, you'll need to prepare. Learn more about translation.
Vox Daily offers a daily dose of voice acting news, articles, tutorials, interviews, intelligent conversation and business ideas for voice talent and voice actors.
Our feed & social options update you with special offers and news as it happens.