Vocalizer

From: Scott Royall [mailto:royall@conchbbs.com]
Sent: Friday, August 16, 2013 15:17
To: Rachel Elias
Subject: Vocalizer

Rachel,

I have a request: Would you please re-send the Vocalizer SDK to me? Or rather, given the vagaries of how Nuance operates, would you please make arrangements for me to re-acquire it? I obviously no longer have it, and I need a second look at it.

We both know how I feel about Eloquence, I routinely have it doing things that even its creators probably never even considered. But, as you pointed out a while back, Nuance currently owns Eloquence and regards it as a dead technology. Putting aside my personal opinion of that decision, I can’t dispute its logic. There has never been much public interest in high-quality voice synthesis, and Eloquence achieves quality by being very customizable. Few users would take the time to tweak that quality from it, however, preferring simpler solutions. That’s really why Eloquence was allowed to sunset.

Yet, there are a few people like me who need voice synthesis that’s good enough to effectively communicate with total strangers while hopefully being flexible enough to convey at least some of the user’s personality. Given that Eloquence is dying a lingering death, somebody has to be looking at what comes next. It’s a safe bet established vendors of assistive and argumented communication solutions aren’t, because the healthcare and education personnel who distribute such products continue to favor simplicity.

Once more, it falls to reluctant pioneers like me to investigate what remains possible, and that’s the reason for my request. I need to find out what products like Vocalizer can and can’t be made to do, since I’ll probably outlast Eloquence by at least a little while. In fact, I do have a test regime to subject Vocalizer to, and Nuance may benefit from what I find.

My first step will be to construct a bare-bones application with just enough functionality to allow me to exercise Vocalizer. The Eloquence SDK came with such a test app, but the Vocalizer SDK doesn’t as I recall. In any case, the purpose of the app will be to give me enough access and control to test Vocalizer under a variety of conditions.

You may dimly remember that I am also an Amateur Radio Operator so you may have some small appreciation of the conditions I need to test Vocalizer under. It’s fair to call some of the conditions downright unfavorable, but let’s remember that Eloquence has survived in such environments for years so they are certainly survivable. My thinking is that, even in situations where Vocalizer may fall short, valuable information can still be obtained and relayed back to its engineers. That’s one advantage of dealing with a current product, we stand at least a little chance of affecting positive change. Since most users don’t bother to adjust what is handed to them, maximum intelligibility has to come baked in.

Once we get reasonable performance from Vocalizer, I then have to face the onerous task of mating it with Xpress-It. That’s the actual app that provides what limited effectiveness I do have verbally. It does this by learning the word patterns I tend to use and then providing predictive suggestions. While I don’t really expect to ever sell a copy of Xpress-It, you never know, especially after I’m gone.

Scott