Ericast 332 - The Artificial Voice

Wondering if our new robot overlords are going to provide a better future for us? Let's give it a spin. Here's your show.

Website viewers: Click here to listen to this week's episode!

Podcast listeners: Click here for a link to the show notes page!
Everyone: Call 701-645-ERIC (701-645-3742) and comment

As you probably figured out by now, this episode is artificially generated based off short voice sample from the last episode. I made sure to not include any of the parodies of the Marines content or else my presentation would sound a lot more dramatic than it usually does.

However, what I'm finding is instead I sound incredibly depressed. I'm really not. My friend Eeyore says I sound fine to him. But, maybe I shouldn't be listening to a donkey. This also gives me the opportunity to try out some Steven Wright jokes: I spilled spot remover on my dog. Now he's gone.  I was sad because I had no shoes, until I met a man who had no feet. So I said, "Got any shoes you're not using?" It's a small world, but I wouldn't want to have to paint it.

I have no idea if those are really jokes from Steven Wright but a page on the Internet said they were.

I also realized that a simple typo - did you catch the one a minute ago? - makes me say things that I'd never actually say. But, you'll also notice that there isn't a single "um" or "ah" or misspoken thing, which I think is a first for the Ericast in its 18 years. Happy Birthday to us, by the way. As I joked with Candela when she turned 18, "this means that now you can be tried as an adult". Otherwise, age is just a number... which is, like, literally, yeah, that's what it is. Like, by design.  That's how it works.  Anyway, we've been doing this show non-stop (except for extended delays) for 18 year and I never thought that this kind of technology would be a click or two away. Maybe 18 years from now I'll finally have my flying car.

Oops. Another typo. Or maybe I meant to say "18 year". You'll never know. Except you do, because that makes no sense.

I think this is my chance to experience my lifelong dream of being on National Public Radio.  "You’re listening to... The Delicious Dish, on National Public Radio. It’s Christmas season again, our favorite time of the year. I got real freaky this year. I’m asking Kris Kringle for a wooden bowl, some oversized index cards, and a funnel. I like to leave Santa some tap water and rice. If Santa’s anything like me, Christmas foods really reek havoc on the ol’ digestive system."

As you've probably figured out, I'm running these paragraphs in chunks and playing them back as I go to make sure this hasn't gone totally off the rails. And I just realized that's a terrible cliche to use, considering some of the disasters the nation's rail system has experienced so far this year. Anyway, this is a tool from a company called "Eleven Labs" and I'm having fun with it but frankly it's really pretty disturbing how good it is. My cadence is off, sure, but if I threw a statement in here that I recorded myself, I don't think you'd spot the difference. Unless I was REALLY enthusiastic!!! That might do the trick.

But, apparently, if I capitalize something it just becomes dramatic and breathy instead of enthusiastic. I wonder if that would help with my N P R career. "Sitting in for Terry Gross I'm Eric Larson and this is FRESH AIR." Hmmmmm.

Anyway, longtime Ericast listeners will remember the "Ericast from the keyboard" episodes where I would bounce an email, probably one from my Blackberry - remember those? - through a couple blog and email services until it would do a text to MP3 service and then drop into the Ericast feed. This could basically do that, except it's my voice, except that it's not quite "normal". However, this is based on a single training file grabbed from last episode, so maybe I can make it better.

I'm going to put this experiment to an end for now because frankly it's getting a little disturbing to listen to myself say things that I never actually said. I'd love to hear what you think - 701-645-3742 is the listener feedback line as always. Or, go out to Eleven Labs and sign up and generate a fake speech file with some of your feedback and email it to me... but calling is probably simpler.