By Mark Esposito, Guest Blogger
The chair emeritus for the American Board of Recorded Evidence, Tom Owen, and Ed Primeau, a Michigan-based audio engineer and forensics expert, have independently concluded that the furtive pleas for help clearly heard on the 911 tapes are not George Zimmerman’s. Both acknowledged experts used voice enhancing software, but different techniques, to rate the probability of the voice being Zimmerman’s at no more than 48%. A 90% match is considered scientifically reliable.
In a report published by the Orlando Sentinel, Owen said he derived his conclusions based on biometric analysis. “It basically just means using personal characteristics for identification. A fingerprint scanner is an example of a biometric device. Much as the ridges of a human hand produce a fingerprint, each human voice has unique, distinguishable traits, Owen says. ‘They’re all particular to the individual.'” The expert recently used the technique to identify the accused killer of Sheila Davalloo in a 911 call made almost a decade ago.
Owen, who also served as the chief engineer for the New York Public Library’s Rodgers and Hammerstein Archives of Recorded Sound, said that, “as a result of [the testing], you can say with reasonable scientific certainty that it’s not Zimmerman.”
Ed Primeau agreed but went further, saying that, under the known circumstances, ” I believe that’s Trayvon Martin in the background, without a doubt. That’s a young man screaming.” Primeau used the technique of voice enhancement to reach his conclusion. Unlike biometric analysis, his method does require an in-context sample of the voice for testing.
An article in the Cleveland Plain Dealer published in 2010 describes the 66-year-old Owen and his cohort, Stuart Allen, this way:
[The pair have] more than six decades of experience between them in the forensic audio profession. They’ve worked with the FBI and other federal agencies, police departments, private detectives, prosecutors, defense attorneys, and news organizations. Many courts have designated them as expert witnesses. They’re good friends who sometimes are on opposing sides, but respect each other’s abilities. “Both of us are known as sort of contrarians,” Owen said.
Primeau is a former sound engineer in the movie industry who worked with
pop stars Anita Baker, Bob Seger, and Barry Manilow. Primeau has over thirty years of experience in voice identification and is a registered investigator for the American College of Forensic Examiners. He describes his work in voice identification as:
There can sometimes be differences in speech patterns that can help identify clues in your identification puzzle. I look for several similarities as well as differences, nasal resonance differences, voice tone with regard to inflection both similarities and differences.
The test results seem to present another blow to Zimmerman’s credibility who claimed that it was his voice on the tape — and not the African-American teen — heard crying out for help mere seconds before the fatal gun shot.
It is likely that similar audio testing is being conducted by the FBI’s Digital Evidence Laboratory’s Forensic Audio, Video, and Image Analysis Unit, based in Quantico, Virginia. Should they reach the same conclusions as Owen and Primeau, Zimmerman would almost certainly face charges in the death. An opposite result would go a long way in substantiating his claim of self-defense.
Source; msnbc; Orland Sentinel; Cleveland Plain Dealer
~Mark Esposito, Guest Blogger
OK Otteray Scribe Esq: I’ll do it in pieces…
@Randy – Here is the Seinfeld episode in question: http://www.youtube.com/watch?v=NZygCA4Samg I don’t think it was a voice-over but maybe it was. How do you know? It was called “The Pledge Drive” episode.
How do I know the old panicky woman 911 tape was actually a woman? Because not only did she have a female name (your version was blanked out) she also had a behavioral demeanor typical (or peculiar) to HER demographic.
Auto “timing” and Otteray Scribe Esq: Randy I think your understanding of “timing” is pre-ECU/PCM era. The o-scope he refers to can be used to do the new-age “timing” (i.e. a ubiquitous word). Albeit, OS is an attorney (highly-paid no doubt? LOL) he also has probably brought his late model Beamer, Benzo, Lexus, etc. into those ‘bandits’ and paid extra for that darn o-scope technique. Am I right OS? If not then you have basis for civil litigation counselor! (LOL)
Sonofthunder, WordPress will only let you insert two links into a comment. Three or more gets caught in the spam filter.
Look at my web montage of a killer. Who looks scarier? http://i39.tinypic.com/208bbrn.jpg
You be the judge (pun intended)…
@Randy – Here is the Seinfeld episode in question:
For some strange reason my latest posting (kinda’ long) says “Your comment is awaiting moderation.”. It’s been a few hours now and it has not been approved yet. I have no idea what caused that to happen. However, if you want to peek at what I posted got to:
1.) http://notepub.com
2.) Log on as me: sonofthunder (I don’t care what you do there its a freebie throwaway account).
3.) Password is: 123456
I think you’ll really find it interesting. Especially Randy, Otteray Scribe, and Malisha please check it out…
Eventually I guess it will show up here someday (I hope)…
Zimmerman and the corrupt cops who have intentionally or unintentionally botched this case have a strong need for prison time.
I believe that we HAVE to hold our police to a higher standard and when they screw things up this badly, they need to be punished…again whether they intentionally (string ’em up!) and unintentionally screw up.
Same goes for politicians and others we put in control of our society’s well-being.
What would a REAL head trauma look like on somebody that is like George’s ethnicity? What would George look like if you morphed his only two known photos together? Stop wondering:
http://i39.tinypic.com/21j93j9.jpg
anon:
I thought you had a grasp of the minutae but not the big picture. My mistake. We can debate the esoterics of probability but the degree of certainty can be described in terms of a numerical measure and this number, between 0 and 1, with zero being no chance and one being a 100% chance. The audio experts match of voice similarities compares to the percentage of reliability which is translated into a number. It is exactly equal to the probability of the outcome. Thus a 48% match is exactly equal to a 48% probability where a 100% match of characrteristics would equate to a 100% probability. As the article stated a 90% match would be scientifically reliable as it would equate to a 90% probability.
But the big picture is that two experts who know quite a bit more than both of us and working independently, arrived at the same conclusion: It’s not Zimmerman — thus shooting his down the poor, poor, pitiful me defense. Quibble away.
Malisha, the apologists for Zimmerman are out in force and are grasping at straws. They have no interest in the truth, but in smokescreens. I am not wasting my time on this drivel.
There’s nothing bizarre about being able to say “that’s not Zimmerman’s voice” when you can’t say exactly whose (else) it might be. I understand that perfectly. If there is not a better than 50% match-up on the various measurements, he can’t be the one who made that particular sound; there’s nothing mysterious or difficult about that.
Even if nobody has Martin’s voice for a comparison, here’s another thing that the “NOT ZIMMERMAN” tends to prove:
1. Zimmerman lied about what happened, during his interview with the police;
2. Zimmerman was fitting his statement to the police that night to exonerate himself, not to provide details about what had taken place.
3. Zimmerman was pretty glib with coming up with excuses for what took place. “I called out help help help but nobody would help me.” That’s a very planned out, transparently foolish statement, and the fact that he thought it was believable says something very negative about HIM or it says that he already KNEW the police were prepared to take whatever foolish story he told and accept it as true. (Say, perhaps, Daddy already assured him things would be all right?)
4. He knew he could remain silent and ask for an attorney before that interview but he chose not to do that. After being beaten half to death and being shaken and now — really? To me it means he was pretty secure that his fairy tale would be good enough for the interrogators. So, “what about the cries for help?” That was ME, guys, I was yelling help help.
SO THIS GOES TO CREDIBILITY, not just to whether or not he was being attacked.
Had he really been caught off guard and attacked and battered half to death, he probably would not have remembered yellilng, or remembered hearing any yelling, or anything like that. 35 minutes later? He’d still be a trembling jittery freaked-out basket case.
BTW, besides lying about the screams for help and getting found out, I think there should be a little investigation into what the cop was checking out on his jacket and in his pockets right after stowing something in the trunk of the squad car.
So I think Zimmerman’s father, Zimmerman’s friends (who testify that they know that’s his voice) and those others all have a bit of a credibility problem now.
Yeah, and how ’bout that cop who stowed the stuff in the trunk — is he gonna be yelling “help help” any time soon?
It’s tiresome to be constantly tasked with finding mespo’s boneheaded mistakes.
So it goes.
Randy, I knew I’d heard that “cry for help” before. Thank you for the chuckle.
Dredd, thank you for that cite, that was useful and readable. An interesting aspect that I’d still like to know more about is this:
Voice analysis thus rests on the non-likelihood that two individuals would have identical vocal cavities and identical dynamic patterns of articulator manipulation, and on the inability of an individual to change or disguise the particular voice characteristics created by his unique combination of cavities and articulator manipulative patterns. Spectrographic voice analysis involves the reflection of voice characteristics in a “spectrogram” produced by a “spectrograph.
I believe this was a problem for DNA analysis with the first DNA tests saying, “the likelihood of two matches” is astronomically small, and then the science refined showing it was actually quite likely depending on the demogrphics of the local region, which is one reason there is a push on to collect DNA from everyone to create DNA databases.
Similarly, what is the likelihood two people’s voice samples can have identical patterns? What research has gone into determining that number, or are we left with hocus pocus three breasted suit types telling the court and looking very serious that in their professional opinion it is highly unlikely?
Actually mespo and Gene, you need to reread what mespo originally wrote, because it shows how mespo is incompetent and couldn’t understand what he was reading.
“Both acknowledged experts used voice enhancing software, but different techniques, to rate the probability of the voice being Zimmerman’s at no more than 48%. A 90% match is considered scientifically reliable.”
Mespo made the original error of calling a % match a probability and Gene, gene, gene, and you too OS, y’all forgot to catch him on that one.
Whereas my questions, on point, to the point, and asking questions about how this shit is determined, went directly to the probability issue.
But you’re right, it took me until now where I realized I had gone astray and that was in my initial assumption, as usual incorrect, that mespo was competent in the least.
Article said:
Owen told the newspaper that the software compared the screams to Zimmerman’s voice and returned a 48 percent match. He said he would expect a match of higher than 90 percent, considering the quality of the audio.
Mespo crapped out this blog post:
Both acknowledged experts used voice enhancing software, but different techniques, to rate the probability of the voice being Zimmerman’s at no more than 48%. A 90% match is considered scientifically reliable.
Damn it’s happened again, this blog has gone off track before when we assumed mespo was competent.
Gene will you go over slowly for mespo how he got it wrong? (Again)
idealist707 –
I’m surely no apologist for Zimmerman. And I share anyone else’s skepticism re: a neighborhood watch fellow carrying a weapon in the first place. By carrying it, I’d have to conclude that you are establishing in your own mind that – under the “right” circumstances – you’re willing to use it. So the gun got used, and a young teen is gone forever.
Zimmerman may well have been trigger-happy. But we simply don’t know that. He may well have been looking for a confrontation. We don’t know that, either. He does not strike many as a sympathetic figure in this case. But after thousands of field cases investigated, I’ve assumed to the point of looking foolish before. What we think we know ain’t always the truth.
A great example is a case in my area where another highly-suspect fellow found himself in deep doo-doo. ‘Everybody’ figured he was guilty as sin. Trouble was, he was telling the truth all along:
http://blogs.laweekly.com/informer/2011/07/giovanni_ramirez_exonerated.php
In the above case, the media was deplorable in their coverage, and my take is, much of the reporting in this case is just as shameful. Referring to the dead teen as “Trayvon” while calling the shooter “Zimmerman” is a not so subtle example of manipulating public emotion. Showing photos of the teen as a 14-year old – instead of a 6’3″ man-child – is another.
So it wouldn’t surprise me at all if the shooter in this case is lying through his teeth.
But neither would it surprise me if he indeed got jumped by a teenager who was highly irritated at being followed.
There’s a reason we wait for the fact lady to sing.
TIME has a good article ‘The Law Heard Round The World’, in the april issue. “Trayvon Martin;s death raises the question. Can “stand your ground’ be defended.
Sonofthunder,
In the Seinfeld episode they used the same person for both voices. The did a voice-over on the man using the woman’s voice. It was a spoof, not a demonstration.
I never said the use of voice was 100% accurate, only that it is very reliable.
“we’ll see it’s one of the neighbors on 911 who was panicky like that old woman who kept interrupting the 911 dispatcher”
Do you mean the panicky old woman that you identified to be such by her voice alone? 🙂
I know what Otteray Scribe was talking about when he identifies an ocilloscope in an auto repair shop. The problem is that those shops don’t use that tool to set timing.
Ummm – I think you guys are WAY behind in the speaker-independent voice recognition technology. Computers are very accurate today with voice-recog and if you play that 911 recording into Google Chrome at the Google Voice Search site you will see the system decode it pretty close. So can computers distinguish George’s or Trayvon’s NORMAL voice from a SCREAMING voice? I think the technology is there today. Albeit only available to people like DARPA, NSA, CIA, Etc *. but still available. Does FBI have it yet? Maybe. Not sure if Quantico is up to that level yet. I think they’re still using IBM and Microsoft products. :->
*C.T.U. folks use this software daily to identify terrorist’s voices against a library of voices on captured telephone and radio calls (ELINT/SIGINT).
BTW – With new DNA techniques you can use polymerase chain reaction (PCR) on a original oily fingerprint(s) and get some DNA. Once you have that you may be able to tell gender and maybe even ethnicity with the new Genographics database at Harvard. Not sure it’s up to the level of local police, FBI, court, etc. For now kinda’ only used in *certain* circles and communities… The others will catch up soon I hope…
@Rany – OS was talking about this:
http://images.picoauto.com/diagnostic-scope.jpg
Read up on it here: http://www.picoauto.com/automotive-oscilloscope-guide/
@Randy – There was an entire Seinfeld episode about how people mistake a disembodied voice (telephone or recording) for the WRONG gender. I have been amazed at hearing a woman on a Burger King microphone and then seeing a MAN at the window – still speaking in that female-like tone. Some men sound female and visa-versa. You can do voice matching with samples from George and Trayvon but we only have George’s voice. I think if we compare it to the other 911 voices we’ll see it’s one of the neighbors on 911 who was panicky like that old woman who kept interrupting the 911 dispatcher.