The forensic audio ‘experts’ reports of who was screaming

I was surprised to see Tom Owen and Ed Primeau’s names on the prosecutions preliminary witness list. These are the two ‘forensic audio experts’ hired by the Orlando Sentinal to offer an opinion regarding who was screaming in the background of the first 911 call. Both men said they were quite sure the voice did NOT belong to George Zimmerman. When these guys were interviewed on one of the cable shows (I forget, and I’m too lazy to look it up right now), the host asked them if they expected to appear at a trial, and Owen replied that he expected both the prosecution and defense to hire their own experts. Well, the prosecution seems not to have done that,  just listed the two guys engaged by the Sentinal. We’ve since learned that the prosecution passed the call along to an FBI audio lab, which reported back that they couldn’t tell anything, either about who was yelling or the “mystery word.” I find that odd, since the mystery word, at least, is pretty easy to parse. It’s also odd that they asked the FBI lab only for their own independent opinion, and not to comment on Owen’s methodology.

I don’t think Primeau’s testimony would be worth a hoot, since it seems to be completely subjective, based solely on “critical listening skills.” Unless he can explain exactly what “critical listening skills” are, and demonstrate precisely step-by-step how they lead to valid conclusions, I remain completely unpersuaded. Owen employs a more sophisticated technical analysis, based on computerized comparisons of audio waveforms. But this still has the same issue: can you demonstrate in general that the method works? What Owen did for the Sentinal does not even remotely follow the protocol for a valid scientific experiment. The issue is NOT, as many people have assumed, that Owen did not have a sample of Trayvon Martin’s voice to compare to George Zimmerman’s. It’s that he doesn’t have any ‘controls’ at all, no other comparisons to establish a baseline for his Zimmerman-to-scream comparison. At a minimum, in order to establish some scientific validity to his claim, he would need to do the following:
1. Repeat the test comparing the screams against several other samples of Zimmerman’s voice, for example the earlier 911 calls that have been released.
2. Repeat the test comparing the screams against several other voices that are NOT George Zimmerman’s, but in a similar vocal range.
3. Conduct a new experiment, in which several test subjects (I would guess a minimum of 5, again with voices similar to Zimmerman’s tonally) would re-enact his police call from transcripts, recorded via the type of cellphone Mr. Zimmerman owned, and would also scream for help, recorded via the type of cellphone used by the 911 caller at a similar distance and in a similar environment.

Number 3 would test if the method had any reliability. If Owen says Zimmerman’s vocal print has only a 48% match with the screams, and he would have expected a 90% match under the circumstances, that remains just his assertion. He should be able to demonstrate it by showing he can get something like a 90% match on a majority of comparisons recorded under similar conditions. By the same token, different samples of Zimmerman’s voice should produce roughly similar match levels to the screams, and the selection of voices that are NOT George Zimmerman should also produce roughly similar match levels with the screams. (If the non-GZ voices match up at either 5% or 75%, something’s wrong with the method.

All of this testing would have to be witnessed, step-by-step, by both an independent expert (say a professor of audio engineering, not one of Owen’s ‘forensic expert’ colleagues) and a qualified expert hired by the defense. If you did all that, and Owen’s conclusions stood, you’d have a something. Unless you do all that, if Owen goes on the stand, all Mark O’Mara has to do is ask him whether he applied any measure of scientific checks on his conclusions, and Owen goes “uhhhh…” and you’ve got a nothing.

So at this point, I’m guessing Owen and Primeau are on the witness list as a bluff, Corey’s taking the FBI expert to be definitive (perhaps unwisely), and she has no actual plans to call either Owen or Primeau. IMHO, she ought to be following up with Owen to see if he can substantiate his findings along the lines I’ve suggested. If he can’t, as an officer of the court she would be obliged to disclose that to the defense. So what? Prosecutors are supposed to seek the truth, not just convictions, and even if Owen can’t back up his finding, we’re still where we are now, with no one able to say with any objective evidence, who exactly is screaming.