Zimmerman case: Out of the Frye-ing pan, into the spectagraph

Zimmerman case: Out of the Frye-ing pan, into the spectagraph

The Zimmerman court is engaged in a pretrial Frye hearing that will decide, in part, whether the prosecution’s expert witnesses can testify about their analysis of an audio recording of the 911 call made by Witness #11, which ended up accidentally capturing the last moments of the Martin/Zimmerman confrontation.  The dominant voices on the call were the witness and the 911 operator, but  you can hear some difficult to distinguish cries and shouts in the background

The relevant part of the recording–up to the moment the shot is heard–can be listened to here.

The recording suffers from the distance between the struggle and the phone, the more proximal voices of the caller and operator, the low quality of the telephone microphone, and the stress-induced distortion of the background cries.  In an effort to better understand whether it was Martin or Zimmerman, or some combination of both, who was making the distant shouts, as well as to determine exactly what was being said by either of the two men, the recording was provided to experts at the Federal Bureau of Investigation and several independent evaluators.

The results of those findings can only be admitted into court if the methods used to make those determinations are generally accepted within the applicable scientific community—a threshold for admission known as the Frye standard, named after the Federal court case, Frye v. US, 293 F. 1013 (DC Cir. 1923)  that established the rule.

The Experts’ Claims

Federal Bureau of Investigation & State Law Enforcement Experts

The FBI in particular maintains world-class voice recognition capability through the Digital Forensics unit of its Operational Technology Division in Quantico, VA and satellite offices distributed around the United States.  Although these resources are most commonly applied to matters of national significance, such as terrorism, the FBI was engaged in the Zimmerman trial to determine if Martin’s death was racially motivated.  (After several months of investigation they determined it was not.)  State law enforcement expertise was also brought to bear in analyzing the recording.

On May 18, 2012, the prosecution released an FBI analysis stating that the Witness #11 911 recording is inconclusive in terms of identifying the background voices.  Yesterday Dr. Hirotaka Nakasone, who conducted the FBI’s analysis of the witness 911 recording provided devastating testimony at the Frye hearing, stating if any expert witness claimed confident conclusions from the recording he would find it ‘disturbing’.  We’ll write more on Dr. Nakasone’s testimony in a soon-to-follow post.

Tom Owens, Easy Voice Biometrics

Tom Owens conducts voice recognition analysis using his Easy Voice Biometrics, a commercial product he also offers for sale.  This technology is proprietary, and therefore its methodology is not public.  There is scarce, if any, evidence that the system has been peer-reviewed or independently validated with regard to quality and reliability of its output.

On his web site Mr. Owens  notes that:

“As in any other form of identification analysis, as the quality of the evidence with which the examiner has to work declines, the greater the amount of evidence and time necessary to complete the analysis, and the less likely the chance for a positive conclusion.

In this case, of course, the 911 recording contains only very brief snippets of shouting, poorly recorded and deep in the background.  It is not obvious to this bloggers ear that the people in the background are even speaking distinct words.   Mr. Owens’ own criteria requires that a degraded recording contain more data than normal, making it difficult to imagine how he could reasonably extract useful findings from the witness #11 911 recording.

Even if such findings were possible, , because Mr. Owen’s methodology is proprietary there’s been no peer review of his system, making it almost certainly inadmissible in court.

Dr. Alan R. Reich, Speech Scientist

Alan R. Reich is a former professor from the University of Washington who received his PhD in “speech science.”  Dr. Reich has sporadically published scientific articles on various aspects of human acoustics from the mid-1970s to the early-1990s.

Dr. Reich was retained by the Washington Post newspaper to analyze the recording.  The Post reports that he has been involved in hundreds of criminal and civil cases for more than 35 years.

In this case he used Sony Sound Forge Pro and KayPentax Multi-Speech software to apply spectrographic methods, creating colorful visual representations of the audio recording.

The analysis he produced of the witness 911 recording is amazingly detailed. He says that he hears the younger of the two men yell, “I’m begging you,” 26 seconds later, “Help me,” and in the last second before the gunshot, “Stop!”   Analyzing the pitch of that last word he concluded that it was “highly appropriate for a 17-year-old” Trayvon rather than the 28-year-old Zimmerman.  He concludes:  “It was Trayvon who felt threatened.  The help cries are all Trayvon.”

Since the Washington Post article was published Reich has also been retained by the prosecution.  He now claims to have heard even more detail in the recording, in the form of Zimmerman uttering the pseudo-religious proclamation,  “These shall be.”

Awkwardly, the same Washington Post article also features James J. Ryan, the retired head of the FBI’s forensic audio analysis unit, who states that it’s not possible to determine the age of the person screaming—and, by extension, whether it was Martin or Zimmerman—by measuring the frequency or pitch of the scream.  He states:

“To my knowledge, there are no scientific studies of pitch as an indicator or anything else in a scream that would give someone confidence to say how old somebody was,” he said.

Dr. Harry Hollien & Dr. James Harnsberger

Drs. Hollien and Harnsberger apply an “aural perceptual” approach to speaker identification that relies on carefully listening to recordings for differences and similarities in speech.  Key voice features used for analysis are dialect, ‘sociolect’ (dialects distinguishing social class), speech defects, and voice quality.

Drs. Hollien and Harnsberger were unable to match the screams to either Martin or Zimmerman, even at their “weakest match” level.  Their most favorable conclusion for the prosecution’s narrative was that some of the scream may “tend” towards Martin.

It is notable that Dr. Harnsberger agreed to provide evidence produced from his analysis, but despite repeated reminders this data had not been received by the defense as of May 22, more than four weeks later.

Wrap-Up

The court got through only one expert witness in today’s Frye hearing, which was substantially delayed due to the lengthy hearing regarding judicial sanctions against the prosecutors for misconduct in discovery.  That witness was Dr. Hirotaka Nakasone, who conducted the FBI’s evaluation of the witness 911 recording—and there’s no telling yet which speaker recognition experts Judge Nelson will ultimately allow to testify at trial, if any.

One thing that’s clear, though, is that Dr. Reich is wildly an outlier among this pool of experts, claiming to hear things that cannot be heard by any of the prosecution’s other expert witnesses, nor even the Federal Bureau of Investigation’s top expert in speaker recognition—nor, for that matter, your humble author.

---

Andrew F. Branca is a MA lawyer with a long-standing interest in the law of self defense.  He authored the seminal book “The Law of Self Defense” (second edition shipping June 22–save 30% and pre-order TODAY!), and manages the Law of Self Defense web site and blog.  Many thanks to the Professor for the invitation to guest-blog on the Zimmerman trial here on Legal Insurrection!

DONATE

Donations tax deductible
to the full extent allowed by law.