Thus we get drawn in further to the bright and amazing future that computers can provide us.  There was a blog post, by a friend of mine;  Dave Courvoisier… discussing the newest technology  which is computer generated voices (not really that new) that had emotions and sounded pretty real (that’s new)… Digital Voice Actors.

Here’s that blog entry http://bit.ly/4hZcn3

I’ve read the entry, followed the links and heard the samples of the computer “generated” voices.  I’ve been in the computer industry for many years, I’ve actually spoke with some professors at UCSD’s computer lab many years ago about computers learning to talk.  It was really complicated, took a lot of compute power, and was just plain difficult.

In listening to these ‘CGV’ it seems to me like they have recorded a voice actor and cut up the words, or phonetic chunks and piece them back together again based on the script.  They are touting how the computer can sound sad, or happy… express real emotions. I disagree.  The sounds coming out of the computer are human generated, the computer just quickly pastes together what is needed and spits it out.

But let’s say.. yes, computers are going to wind up doing some of the work of Voice Actors, just like animated characters do the work of on-camera actors.  Actually computers already do some of the work of Voice Actors, but by using the same method; record a person and mash it up to be what is needed at the moment spliced together on the fly by the computer.

And if they ever really create a computer generated voice with emotional range, it won’t be cheap. 

Just look at the movies from pixar… or disney that are computer generated.  They take just as long to make as a ‘normal’ movie, and have similar budgets.  Instead of having a bunch of actors, they have programmers and systems administrators and computer operators, as well as artists.  Lots of jobs.. lots of work.. just a different kind.

For computers to even be able to compete with human Voice Actors would require lots of programmers, computer scientists, linguists and so on.  Yes, you could purchase a program to do the acting, but is it just one kind of voice?  Is it one kind of emotion? How much will it cost to get the “I’m rather upset” or the “Sweetly generous” sounding modules?  Can you switch emotions on the fly?  Will it take direction?

And a late addition to the post; How will digital voices handle newly created words?  I mean there are new words introduced every day into the english language (not sure about other languages, just on top of this with english), especially words that are in the high tech realm.  Or a new disease?  Will there be update packs for these Digital Voices?   Fantastic, just one more thing you’ll need to upgrade/update/patch.

 If computers are going to be edging more into the Voice Acting arena with digital voices, I figured I’d get a jump on it and create a language to help the computer understand what the required delivery of the script is. The language I have just made up to tell computers how to read is VEIML, which stands for Vocal Emotive Inflection Markup Language, taking it’s roots from HTML, or HyperText Markup Language.

The words in the <> symbols are what are refered to as tags, it identifies the information that is in between the <tag> (starting tag) and the </tag> (ending tag).

<VEIML>

<Backstory>
<VoiceType>MALE</VoiceType>
<Age>35</Age>
<Feel>Friendly but edgy</Feel>
<SocialStatus>BlueCollar</SocialStatus>
<Marital>Single</Marital>
<HeroStatus>False</HeroStatus>
</BackStory>

<Script>

<HighEnergy>Now is the time for all good men to come to the aid of their country.</HighEnergy> <Sarcasm emotion=”none” type=”question”>Isn’t that what people always type on a keyboard to test it out?</Sarcasm>

</Script>

</VEIML>

That’s what kind of stuff we may be looking at in the future.. or not.  One of the most important things with Narration or even commercials with Voice Acting is that you need to CONNECT with the audience.  If the audience were all robots, or cyborgs, or even computers, then a computer generated voice (which doesn’t exist yet) would work just fine.  But people need to feel like they can relate to whom is talking to or with them.  No matter how good the Digital Voice Actor gets, it will always just be a pale shadow to the real thing.