Voice technology made headlines recently when a forensic expert determined who was heard on a 911 call made before Treyvon Martin was shot to death by George Zimmerman last month in Florida.
The Orlando Sentinel called on Tom Owen, a forensic consultant for Owen Forensic Services LLC and chair emeritus for the American Board of Recorded Evidence, to analyze a woman’s call to alert police to a dispute. A voice in the background could be asking for help before a shot was fired. Zimmerman, who claims the shooting was in self-defense, said it was his voice.
Using Easy Voice Biometrics software, Owen, however, determined that the voice was not Zimmerman’s.
“I took all of the screams and put those together, and cut out everything else,” Owen told the paper.
The software compared that piece of audio to Zimmerman’s voice, and returned a match of 48 percent. Owen said to reach a positive result with the quality of audio used he would expect a match greater than 90 percent.
“As a result of that, you can say with reasonable scientific certainty that it’s not Zimmerman,” Owen told the Sentinel. However, he stressed that he couldn’t positively confirm the voice as Martin’s, as he didn’t have a voice sample for comparison.
However, Dr. James Wayman, a San Jose State University speech science expert told Web news site, The Daily Caller, that he disputes Owen’s findings, and would go so far as to testify against their admissibility in court.
“There is no history of, or data on, the comparison of a questioned scream to a known speech sample,” Wayman told the Web site. “Even if we were to have Mr. Zimmerman recreate the scream under identical conditions with the same cell phone, it would be difficult to attribute the scream to him without a sample of a similar scream from Mr. Martin under the same conditions. This is clearly not possible.”