PRESS 5: Forensic audio expert interviewed by MSNBC regarding scientific voice identification
NCAVF forensic audio expert David Notowitz is interviewed about
the scientific validity of identifying an individual through the voice
analysis of a scream.
The George Zimmerman trial of 2013 was a high profile case that discussed using voice recognition software in an attempt to identify the source of screaming on a 911 call that captured the shooting of Trayvon Martin.
David Notowitz, the founder and president of the National Center for Audio & Video Forensics was interviewed by MSNBC regarding his opinion of this evidence by Tamron Hall of NBC's News Nation.
Hall: What does the judge need to weigh with this scientific evidence?
David: It's a very controversial technique that the judge needs to weigh in this case because the audio is extremely hard to hear... you're not hearing actual words that can be compared scientifically.
Voice identification analysis will have the best results when the audio samples are clearly spoken and intelligible words. When the audio is of screams, grunts, or other unintelligible audio, it can be extremely hard, if not impossible, to ascertain a confirmed identification. Mr. Notowitz confirms this point with Ms. Hall's next question:
Mr. Notowitz says based on the evidence for this case it is not
possible to determine the identity of someone based on
Hall: When you say controversial, is that a more gentle way of saying that there's no way to absolutely determine who's voice that is?
David: ...With my knowledge in all the aspects of this audio, I think it is not possible to determine who it is on the tape.
Mr. Notowitz states that for anyone to claim a definitive speaker identification would be very cavalier. He went on to predict that although the use of the evidence in this particular trial would be controversial, the judge would allow both the prosecution and the defense to offer witness testimony identifying who was screaming. In a surprising turn of events, however, the judge did allow the audio to be played in court for the jury, but did not allow expert testimony on the scientific evidence supporting the identity of the screamer. The judge stated, "There is no evidence to establish that their [the audio experts] scientific techniques have been tested and found reliable."
Soon after this interview, the judge on People v Zimmerman
ruled similarly to what Mr. Notowitz claimed, that she would not
allow experts to testify in front of a jury that they could scientifically
identify the screamer on the audio recording.
The judge's comments were in the context that the audio evidence being analyzed consisted of screaming, not dialogue. Not being able to offer "audio analysis experts", both the Prosecution and the Defense offered the mothers of Martin and Zimmerman to identify the scream. Each mother claimed the scream to belong to their respective son. In essence, the credibility of their testimony came down to whomever the jury found more believeable.