A couple of times recently, I've run across these xtranormal movies, generated by online software that pretty much automatically converts text to awkward animations of humanoids (or othernoids) speaking and gesturing quasi-expressively. To say I love this sort of thing is a big understatement. In the case of the examples above and below, one might argue that part of my fascination lies in seeing/hearing these movies as "translations" of familiar texts, and we all know I'm a big fan of translations and transcriptions.
I also have been a big fan of synthesized speech, ever since I first discovered my otherwise horrible 1995 Mac Performa (my first real computer) could speak text. I used to like to entertain guests with this party trick, although I suspect I was mostly entertaining myself. If you've been following this blog for awhile, maybe you remember my Virtual Singers covering such classics as "Nessun dorma," "Hey, Jude," and "On Top of the World." I also once wrote (here & here) about an indie pop singer who, to me, sounds like a virtual singer. And, just a few weeks ago, I posted text2speech versions of a few of my #operaplots on Twitter (here & here), as well as someone else's #operaplot rap (here) and a full-scale Virtual Barry Manilow rendition of someone else's #operaplot (here). I'll wait while you sample all those irresistible goodies...
I think my fascination with virtual speech has a lot to do with an ongoing but mostly subconscious interest in the music of speech. I started thinking more consciously about this fascination in the past few months as I've started listening to podcasts of the This American Life radio show. (I'd never listened to the show much before because the one-hour radio documentary format doesn't really fit my lifestyle - but podcasts are perfect for commuting.) I could go on and on about ways in which I love This American Life and about ways in which it drives me crazy, but I'll save that for some other time.
For now, I'll just note that one of the first things to drive me crazy about TAL was the way in which so many of the hosts talk the same way - specifically, they tend to talk like Ira Glass, the founder and guiding voice. He has a distinctively quirky way of speaking that is sort of affectedly unaffected (or maybe unaffectedly affected) and the fact that others on his staff mimic his quirks can be grating. But, I had to admit to myself as I was noting this that I was listening to the podcasts eagerly - because I find them such compelling listening.
And here's the thing - I don't think the subject matter is always what drives the show. I think as much as anything, there's a truly musical quality to the way the shows are presented. Yes, actual music plays a big part in these talkfests, but the pacing and pitching of the speech is also quite well worked out - the fact that the various hosts use a lot of the same inflections and cadences allows for the same kind of stylistic consistency that music lovers learn to appreciate when hearing various works by the same composer. The one-hour shows always seem to fly by, and I think it's partly because they're music to my ears.
All of that is to say that I've started noticing speech patterns more and more, especially the way pitch and cadences are used, even among us supposedly unlyrical Americans. Having a 3-year-old and a 5-year-old walking around the house adds to this fascination, I suppose, as children are quite naturally sing-song in their delivery. So, when I listen to computer-generated voices, it's fascinating to hear what they get right and what doesn't work so well. Of course, a lot of the humor comes from what they don't get right. In the "Who's on First?" sketch above, the bit from 1:11 to 1:21 kills me every time because the Bud Abbott robot keeps saying "Who" on the same low pitch. Just about any human would naturally raise the pitch each time he gave the same answer in succession; the stubbornly monotone responses somehow make the straight man even funnier.
Here are a few other "transcriptions" I've particularly enjoyed. The last two (actually, the last four, including the two updates) are my own creations.
[UPDATE: Only after I'd created the Seinfeld re-enactment above did I discover that someone else had done that same dialogue (one of the more famous bits from the show, I suppose), although I think mine's a little better...]
UPDATE 2: One more from me...
[UPDATE 3: and another one from me...