Friday, September 21, 2007

Simulating Osama

How does the CIA identify a recorded voice as Osama's?

It uses a program that digitally compares the recording to a voiceprint taken from a sample of his voice considered to be authentic. There has been a lot of research into voiceprints, with numerous proprietary methods invented by telecommunications companies. For example, Cellmax says its voiceprint technology is so good it can discriminate between close relatives with similar voices and can identify a speaker whose voice is altered by a cold.

We must expect that the CIA's technical services division is using the latest, state-of-the-art classified voiceprint technology.

So, does that mean their identification of bin Laden's voice on various internet videos since 9/11 is reliable? Sure, it might be. But advanced voiceprint technology also implies the possibility of near-perfect voice counterfeiting. In fact, implementation of speech authentication security schemes have been slowed because crooks are becoming increasingly sophisticated at fooling such technology, according to some reports.

Another thriving industry is voice simulation. For example, AV Voice Changer Software Gold boasts that "unlike other voice changers, Voice Changer Software Gold changes your voice over the internet in real time and provides an unlimited number of new voices. You can modify your voice by changing voice pitch, voice timbre, applying effects, adjusting advanced tuner and setting equalizer."

So then, how hard would it be to create a program that inputs the voiceprint numbers into a simulator, which then makes you sound like Osama and -- since it is based on Osama's voiceprint -- fools voiceprint analyzers?

We do know that the Osama tapes all seem to have something wrong with the visuals. Either old footage or use of possible lookalikes. But the feds say they rely on the voiceprints. And the press reports that "Osama bin Laden said in his latest video..."
In one case, there was the possibility of lip-synching by a guy in a black beard.

My point is that the news media should be a bit more cognizant of the problem of authenticity. This morning's New York Times kept the Osama tape to four graphs and the headline qualified the bin Laden link with a "said to be." That's progress, maybe.

Bush jumped at the opportunity to say of the "blackbeard" tape that Iraq was mentioned and that al Qaeda's targeting of Iraq was a serious concern. But Bush's judicious phrases showed that he thought it best to avoid saying flatly that the video was an authentic bin Laden broadcast. However, he was happy to play along with those broadcast media that did flatly assert that bin Laden had spoken on the tape.

Another concern: how about the possibility of framing people with phony tapes of voices that sound just like them? The courts need to take notice.

See articles by David Jastrow, June 1, 2007, Speech Technology, Olga Kharif, April 20, 2005, Business Week, and Greg V. Hulme, Oct. 28, 2002, Information Week.

Scientist tackles 'simulated' election
The problem of high tech rigging of results for political ends is pointed up by Steven F. Freeman in his new book Was the 2004 election stolen? Freeman is on tour promoting the book and sounding the alarm over the apparent ease of election fraud and systemwide collusion in coverup.

Freeman, who has a PhD in behavioral science from MIT, charges that the weight of evidence from exit polls points to massive fraud and that claims of pollsters protecting the system are not viable. Everyone is agreed that statistically the results point to a highly improbable outcome but are not agreed on the likelihood of theft. Freeman, a former professor of management, is not a professional statistician but uses statistics as a matter of course in his lines of work.

Freeman, who wrote the book with journalist Joel Bleifuss, gives a good synopsis of his views at To subscribe to his newsletter, visit

Robert F. Kennedy Jr. of Pace University mentions Freeman's work in a Rolling Stone article titled Was the 2004 election stolen?

John Allen Paulos, who is a professional probabilist, is also skeptical of the election result. Search for his home page and read his article.

Additionally, a group of professional statisticians signed a letter expressing skepticism over the election claims.

Rep. John Conyers wrote a foreward to Freeman's book. Conyers, who probed the situation in Ohio, wrote his own book, What went wrong in Ohio? The Conyers report on the 2004 presidential election. Author and commentator Gore Vidal, who wrote an introduction to Conyers' book, told an interviewer that neither the New York Times nor the Washington Post deigned to review the book and that the Boston Globe gave it only a brief, cursory mention.

Vidal believes the 9/11 attacks were allowed to happen as part of a coup by the Bush bunch in order to transform America into a police state that is compliant with a grandiose international agenda. Conyers' views on 9/11 treason are difficult to find via Google but I would guess that he views the subject as politically intractable at this point.

No comments: