Almost statistics

March 23, 2008 2:57 PM

realscoop.com "utilizes proven voice analysis technology to analyze statements made by public figures." The statements being on videos. They're all over Bill Clinton and Reagan. Bush and current candidates - not so much.

How the technology works

RealScoop uses the latest in emotion-based voice analysis technology, as well as some other cool stuff to provide you the best information about what is really going on when people make statements. Sure, applications of the core technology are used around the world for airport security, insurance fraud prevention, intelligence gathering and law enforcement investigations. But what fun is that??? We combine the best technology with the coolest celebrity, sports and political content to provide a one-of-a kind experience for our community.

Unlike other antiquated voice technologies, RealScoop measures and analyzes patterns in voice frequencies, which provide an in-depth view into the person’s range of emotions. During the voice analysis process, RealScoop ignores the actual content or spoken words and focuses entirely on the changes in the voice frequency patterns caused by different emotions. Our technology uses sophisticated algorithms to identify, measure, and analyze patterns embedded in the actual waveform of human speech. The results are then streamed through the Believability Meter in synch with the video and audio.

posted by Kirth Gerson (23 comments total)

Note: This is just voice stress analysis.

For some reason I'd been under the impression that it was actually illegal to run voice stress analysis stuff against certain gov't officials, but I can't seem to find any supporting documents about this now.
posted by c0nsumer at 3:34 PM on March 23

Also, here is some old VSA software. It comes with source, but it's old DOS pascal stuff.
posted by c0nsumer at 3:36 PM on March 23

Is there a site with the phrenology goods on the politicians? I've noticed many of them have "the sloping brow and cranial bumpage of the career criminal."
posted by dances_with_sneetches at 3:55 PM on March 23 [3 favorites]

I'd be much more interested in what astrologers have to tell us about Hillary's star sign.
posted by Pope Guilty at 4:02 PM on March 23

Note: This is just voice stress analysis.

Note: This is just *claimed* to be voice stress analysis.

Given that they're picking and choosing the videos, the meter doesn't appear to be consistent with similar tones of speech, and an overall lack of detail about their process - I would discount this even more than it already should be. Also note the multi-speaker clips, where only the targeted speaker is "tested", showing that there's at least some manual manipulation of the "results".
posted by o0o0o at 4:13 PM on March 23

Even voice stress analysis is of highly questionable effectiveness at best. This site is about as "proven" as horoscopes.
posted by digaman at 4:26 PM on March 23

Just wait until they add the earlobe-analysis module.
posted by Kirth Gerson at 4:41 PM on March 23

From digaman's link: Voice-stress analysis, an alternative to the polygraph as a method for lie detection, is already widely used in police and insurance fraud investigations. Now, however, it is being touted as a powerful and effective tool for an array of new applications -- everything from the screening of potential terrorists in the nation's airports to catching wayward spouses in messy marital disputes.

And here's a cheery story on the guy who sold more than a million dollars worth of his voice stress analyzers to the U.S. government, which have been used in Iraq and Guantanimo Bay, as well as by police departments domestically. His voice analyzer was used on Michael Crowe, for instance.
posted by the other side at 4:43 PM on March 23

i love it when people say "uh" and realscoop reveals they're lying their asses off

don't say "uh" unless you really, really mean it, folks
posted by pyramid termite at 4:53 PM on March 23

Well, I mean, if you didn't have anything to hide, you wouldn't, like, need to use filled pauses.
posted by cortex at 4:55 PM on March 23

Just watching a few, it seems to be a volume meter more than anything. The louder a sentence, the more in the red it goes. Meh.
posted by zardoz at 5:00 PM on March 23

Very very cool. If the algorithm is accurate or not is kinda beside the point - no one is going to put up a really good quality algorithm for the world to use free - not yet. In the future the technology will probably improve and it shows what kind of tools may become available to the masses. Remember the "Hitler Speaks" film from yesterday when it was revealed Hitler took acting classes and his persona "on stage" was much different from real life. What if everyone had a "bullshit meter" in 1933.
posted by stbalbach at 5:20 PM on March 23

I'm a bit torn by posts like this. I mean, if someone were to ask me what I thought about it, I wouldn't really know what to say.

On the one hand, the technology to pull real analysis out of a speaker's intonation, pauses, etc. is clearly, well bullshit. It's simply not reliable enough to trust with anything more important than "did you take the last cookie?"

On the other hand, anyone who's ever done primary qualitative research knows you can train yourself to understand when someone's being less than honest (although not perhaps being outright deceptive). I find it difficult to accept that this demonstrable ability is some kind of confirmation bias, nor am I prepared to throw up my hands and call it spooky action at a distance.

Eventually, we may be able to turn what is now entirely qualitative data into quantitative data, but we're not there yet. However, it would be as equally foolish to embrace this technology in its present incarnation as it would be to dismiss what it will inevitable become. In the either case, the consequences of both use and misuse are not something most people are prepared to contemplate.
posted by digitalprimate at 5:42 PM on March 23 [1 favorite]

but you gotta admit, the meter was right on when Spitzer said he would like to annex the Sudetenland.
posted by subgear at 6:38 PM on March 23 [1 favorite]

Whenever I hear of stuff like this I always remember the CNN chosen psychologist telling the world that Richard Jewel was obviously guilty because he stuttered a few times while being interviewed.
posted by srboisvert at 6:46 PM on March 23 [1 favorite]

Welcome to the World of Reverse Speech, in all likelihood it will change your life.
posted by Sailormom at 7:33 PM on March 23

advanced emotion-based voice analysis technology

so the flatter the voice, the more believable the speaker.
uh... on a scale from one to ten, where a fingerprint is a ten and as such ultimate proof and a lie detector perhaps a four, I'd give this a negative six. it's snake oil salesmanship.
posted by krautland at 7:33 PM on March 23

If the algorithm is accurate or not is kinda beside the point

You're insane, that is EXACTLY the point. There are no other points. The only question worth asking is "Does it give reliable results?", and if the answer is "No.", then there is nothing else even vaguely worth asking.
posted by Pope Guilty at 7:35 PM on March 23

Wait, it's not generally assumed that politicians consistently regurgitate bullshit?

This is my surprised face.


A scale from believable to highly questionable? What does highly questionable mean?
posted by Freen at 7:49 PM on March 23

Pope Guilty, BEEP! BEEP! I'm picking up some stress BEEP! BEEP!
posted by StickyCarpet at 8:33 PM on March 23

What if everyone had a "bullshit meter" in 1933.
posted by stbalbach at 10:20 PM on March 23 [+] [!]

We do have a "bullshit meter". It's that grey squishy stuff between our ears. Most people leave it switched off most of the time, though.
posted by signal at 4:52 AM on March 24

This thing has a BELIEVABILITY METER™! HELLOOOO! This thing is for real!

I went to contest a parking ticket one time and I noticed the judge intently looking at his computer as he talked to me about the ticket. When I got up to leave I sneaked a peak and saw that he was using some kind of voice analyses software. I don't know how much he relied on the software, but he let me off from paying the ticket after talking eye to eye about the specifics.
posted by P.o.B. at 2:46 PM on March 24

The whole idea actually seems--at least to me--pretty cool. You have to start somewhere with new concepts--even if it isn't totally 100% in the bag. So...maybe this is the start to something that will actually do what the voice analyzer is trying to do....
posted by Todd Lokken at 8:05 AM on April 8