I Screamed It Couldn't Be Done

The Orlando Sentinel brings on some voice experts to analyze a 911 call in the Trayvon Martin shooting to attempt to establish whether it is Zimmerman or Martin screaming for help.  I have previously expressed my skepticism of this process:

Then we may hear from the FBI audio wizards.  On CSI they would crack this easily (as would Abby on NCIS).  But here in Realityville we are hearing background noise on a 911 call, not hi-fi from a sound studio; I presume we lack past tapes of Trayvon screaming in pain, and we aren't likely to have the authorities beat up George Zimmerman and record his anguished cries.

So, how credible will the FBI wizards be?  I would guess "Not very", which means we are back to reasonable doubt, but folks with relevant expertise are urged to chime in.

[Down in the updates I stack up four experts who agree that the voice quality of the different samples needs to be similar as a starting point for the matching process, so score one for common sense.]

These experts brought in by the Orlando Sentinel overcome the lack of past Martin voice samples by just not worrying about it.  Strong stuff!

Tom Owen, forensic consultant for Owen Forensic Services LLC and chair emeritus for the American Board of Recorded Evidence, used voice identification software to rule out Zimmerman. Another expert contacted by the Sentinel, utilizing different techniques, came to the same conclusion.

Actually, his software [rolled out March 1, 2012, so this version has a track record of roughly zero] failed to "rule-in" Zimmerman; Pressing on, with my emphasis on the non-trust of Martin:

Zimmerman claims self-defense in the shooting and told police he was the one screaming for help. But these experts say the evidence tells a different story.

...

Owen, a court-qualified expert witness and former chief engineer for the New York Public Library's Rodgers and Hammerstein Archives of Recorded Sound, is an authority on biometric voice analysis — a computerized process comparing attributes of voices to determine whether they match.

After the Sentinel contacted Owen, he used software called Easy Voice Biometrics to compare Zimmerman's voice to the 911 call screams.

"I took all of the screams and put those together, and cut out everything else," Owen says.

The software compared that audio to Zimmerman's voice [presumably from the Zimerman 911 call]. It returned a 48 percent match. Owen said to reach a positive match with audio of this quality, he'd expect higher than 90 percent.

"As a result of that, you can say with reasonable scientific certainty that it's not Zimmerman," Owen says, stressing that he cannot confirm the voice as Trayvon's, because he didn't have a sample of the teen's voice to compare.

...

Another benefit of modern biometric analysis, Owen said, is it doesn't require an "in context" comparison. In other words, Owen didn't need a sample of Zimmerman screaming in order to compare his voice to the call.

I am not now and do not aspire to become an expert on the strengths and limitations of voice biometric analysis (although such an effort will surely result in samples of my screaming as a by-product).  However, the Zimmerman defense (if it ever reaches the trial stage) will introduce bafflegab like this:

Therefore, all that the system would learn about the identity of the individual is tainted by the channel characteristics through which the audio had to pass. On the hand, at the time of performing the identification or verification, a completely different channel could be used. For example, this time, the person being identified or verified may call from his/her home number or an office phone. These may either be digital phones going through voice T1 services or may be analog telephony devices going through analog switches and being transferred to digital telephone company switches, on the way.  They would have specific characteristics in terms of dynamics, cut-off frequencies, color, timber, etc. These channel characteristics are basically modulated with the characteristics of the person’s vocal tract. Channel mismatch is the source of most errors in speaker recognition.

Another problem is signal variability. This is by no means specific to speaker recognition. It is a problem that haunts almost all biometrics. In general, an abundance of data is needed to be able to cover all the variations within an individual’s voice. But even then, a person in two different sessions, would possibly have more variation within his/her own voice than if the signal is compared to that of someone else’s voice, who possesses similar vocal traits.  The existence of wide intra-class variations compared with inter-class variations makes it difficult to be able to identify a person accurately. Inter-class variations denote the difference between two different individuals while intra-class variations represent the variation within the same person’s voice in two different sessions.

Hmm, I started the morning believing I was not anti-science, but my view may evolve (if I believe in evolution...).  Let's wave in [from the Updates below] Mr Primeau, the second Orlando Sentinel expert to address this (my emphasis):

4. When conducting voice identification, it is important to create an exemplar of the accused for audio comparison using as exact conditions and equipment as close as possible to the measurements taken from the evidence as outlined above. The speech must be the same as the speech on the evidence in order for the testing to be accurate. As an audio forensic expert, I often have to coach the accused into the same energetic voice tone and inflection as the evidence recording. However, it is still possible to compare speech if the exemplar is not as close to the evidence as I would like.

Oooh, trouble in Paradise!  Don't you hate it when experts disagree? 

What this says to me is that establishing an accurate match of one person (such as for a bank account password) is dependent on the media used, and the quality of the person's voice.  The media differences with the 911 screaming call versus the baseline Zimmerman 911 call is unknown to me.  The Sanford recording equipment should be the same for both, but we may have a Zimmerman cell phone versus the 911 land line call, for example.

I also infer that the voice quality has to be affected by the screaming, if a recognition system can have trouble identifying a person's normal speaking voice from one day to the next.

So I now have learned that, with the 48 percent match, if I call my bank screaming in pain they won't authenticate me; good to know at tax season.  But that doesn't mean it's not me, it just means they can't confirm the match.

I don't think American jurispridence is ready for the notion that, since the whiz-bang science that defies common sense about screaming says the voice is not conclusively Zimmerman's, it must be Martin's.  The second expert sidesteps that neatly:

Not all experts rely on biometrics. Ed Primeau, a Michigan-based audio engineer and forensics expert, is not a believer in the technology's use in courtroom settings.

He relies instead on audio enhancement and human analysis based on forensic experience. After listening closely to the 911 tape on which the screams are heard, Primeau also has a strong opinion.

"I believe that's Trayvon Martin in the background, without a doubt," Primeau says, stressing that the tone of the voice is a giveaway. "That's a young man screaming."

The prospective Zimmerman defense will focus on the fact that this second expert does not believe in the whiz-bang technology and is quite clear that a proper comparison requires similar speaking styles; no shouting, please.  As to his confidence that this voice is Martin, even though he has never heard Martin's voice - I am speechless.  I think I could handle that cross-examination, maybe even without a cup of coffee.

If I were the prosecutor I would be begging for better experts - sending up a guy who doesn't believe in the science won't be helpful.  Sending up a guy who says he can't match it to Zimmerman so it must be Martin will be ripped up by defense experts who will explain the limitations of the techniques (to be fair, the first expert doesn't claim it is Martin's voice, which leaves us wondering, what would he tell us if neither voice matched?).  The defense will uncharitably point out that it might not be possible to match the voice to either person, the judge can then expound on "innocent until proven guilty" and "reasonable doubt", and away we will go.

Now, let's be fair - the Orlando Sentinel actually showed some initiative here by bringing in some experts, so that's a good job by them.  But the story ought to include a lawyer to tell us where the burden of proof lies and what the standards for evidence might be.  As a bonus, I am sure they could find trial lawyers with experience with audio experts and evidence in a courtroom - what do they think?

The story also ought to mention 'John', the witness who told the police on the night of the shoting that he saw Martin beating Zimmerman and heard Zimmerman calling for help.  That sets a high hurdle for the prosecution. 

HMM: Yesterday CNN covered 'John' by shoving him down the memory hole; are we seeing a pattern?

MY HEAD IS SPINNING: Am I anti-science because I don't accept on faith the pronouncements of these two experts in a field far beyond my expertise?  Or am I pro-science because I am trying to reach an independent opinion guided by other experts whose qualifications I am unable to evaluate?  That might seem like a hard question, but since I am a righty, I know libs know the answer.

THAT WOULD TIE IN TO THE 48% MATCH:

From 'myiq2xu', who may also be aided by stronger coffee:

Did anyone consider the possibility that BOTH men were screaming and yelling at the same time?

If Zimmerman did half the screaming they have a 96% match.  Kidding.  The linked article about the compounded difficulties of identifying one voice and separating two would suggest that the defense will make this point:

The existence of wide intra-class variations compared with inter-class variations makes it difficult to be able to identify a person accurately. Inter-class variations denote the difference between two different individuals while intra-class variations represent the variation within the same person’s voice in two different sessions.

If this trial is televised I won't be able to hit the remote and get to baseball quickly enough.

HOW DOES IT WORK:  Excellent intel from Cecil Turner, who musters some demonstration videos of the Easy Voice Biometrics system. Apparently they can come pretty close to spotting Richard Nixon (86% from one speech to another, lending credence to the talk that there really was a New Nixon) and do OK (above 90%) with lightly disguised voices.  A demo of someone matching a scream would seal the deal but I see no such claims on their website.

Interesting - I see from their press release that they just started shipping their product on March 1, 2012.  Well, I guess they welcome the headlines, but that suggests they have zero track record in a court room for this version.  Sadly, the JOM budget does not include $4,995 for their software, so we won't be demoing it at home.

BUT DESPITE MY BUDGETARY CONSTRAINTS:

I am waving in my own expert witness, who claims to have been doing this for decades.  By uncanny coincidence, he is the same guy the Orlando Sentinel used, but whatev.  Let's cut to his bottom line on comparisons based on shouting versus speaking (my emphasis):

4. When conducting voice identification, it is important to create an exemplar of the accused for audio comparison using as exact conditions and equipment as close as possible to the measurements taken from the evidence as outlined above. The speech must be the same as the speech on the evidence in order for the testing to be accurate. As an audio forensic expert, I often have to coach the accused into the same energetic voice tone and inflection as the evidence recording. However, it is still possible to compare speech if the exemplar is not as close to the evidence as I would like.

Well, the Orlando Sentinel won't want to interview him.  Oh, wait - they did!  Well, they won't want to put their two experts in the same room then.  But the hypothetical Zimmerman defense team will. 

Remember, this guy is so good he can identify Martin's shouting without ever hearing it - sign him up! Team JOM gets props, especially narciso.

THE BUDGET LOOKS GOOD:  Let's hire Stutchman Forensic Laboratory, Advocate for Evidence Since 1992:

Collection of Exemplar

It is recommended that the exemplar of the known voice must be collected in as close to the same manor as the recording of the unknown voice was recorded. For example, if the recording of the unknown voice was recorded over the phone, the exemplar of the known voice should be collected over the phone, etc. When the exemplar is collected, the suspect is asked by the examiner to stay the same words in the same way as they were spoken by the unknown person. In other words, in a normal, natural voice.

Or maybe natural, normal screaming, if you need a headline to push your product.  No word yet on whether Stutchman et al are good enough to ID Martin without hearing his voice.

STILL MORE EXPERTS: From p. 91 of a guide to audio forensics, IEEE SIGNAL PROCESSING MAGAZINE [91] MARCH 2009:

AURAL-SPECTROGRAPHIC VOICE IDENTIFICATION

Audio forensic examination of recorded dialog may lead to a legal dispute over the identity of one or more of the conversation participants. A criminal suspect or a party to civil litigation may deny being the individual who uttered the recorded words, especially if the recording was made via telephone without eyewitnesses to identify the talker visually. In these situations, the forensic audio examiner may be asked to identify or to exclude that the suspect was the source of the words in the recording in question.

The aural-spectrographic method for audio forensic voice identification is based on the judgment of a trained examiner who compares the unknown example of speech with one or more known examples [30]–[33]. As the name of the method implies, the task of the examiner is to render a judgment based on both an aural comparison (careful listening) and a visual comparison of speech spectrograms.

In a typical case, the examiner begins by listening critically to the recording of the unknown talker and identifies specific phrases that are distinctive and relatively noise-free.  The examiner then arranges a recording session with the suspect to create exemplars that match the selected phrases of the unknown talker in pace, emphasis, and enunciation. The suspect repeats each example phrase multiple times to produce recordings with as close a match as possible to the timing and speech pattern of the unknown examples.

Or, the target screams hysterically in the background of a 911 call and liberals lap it up like kittens.  Either way.

ONLY SUPPORTIVE BY IMPLICATION: Here is Forensic Science Services telling us how they like it done in Canada (my emphasis is on the last sentence):

Voice Identification

The spectrographic voice identification analysis has two steps. The sound of speech is first transformed into a three dimensional (time - frequency - volume) graphic pictures which do reveal numerous acoustical features of an individual’s voice. The second step involves the pattern comparison of the same phrases/sentences from the unknown sample and the suspect’s sample. The results of analysis are expressed as:

•  
  • Probably the same speaker (high level of confidence).
  • Possibly the same speaker (intermediate level of confidence).
  • Inconclusive(due to the insufficient number of comparison words, poor quality of recordings, too high variability of the voice, possible disguise).
  • Possibly not same speaker (intermediate level of confidence).
  • Probably not the same speaker (high level of confidence).

The results depend on quality of recordings, the total number of comparison words, speakers’ condition, and individual speakers’ voice variability. There is a requirement for a minimum number of 20 comparison words in a ‘connected speech’. The suspect should provide the comparison sample by reading three times the transcript of the unknown voice sample.

Background screamers need not apply.

So, is there any reasonable doubt that experts aren't going to line up behind the comparison of the background screaming to the Zimmerman 911 call and silence from Martin?

If our friends on the left could provide an article or two suggesting that the experts above are wrong that would be helpful.  Bonus points for evidence that this breakthrough method not widely endorsed will be accepted in a court and not crushed by a phalanx of defense experts arguing the conventionl wisdom.

Comments

[Ignatz]

--"That's a young man screaming."--

So 28 is an old geezer now?

[Clarice]

You are an ambulance chaser's nightmare.

[George]

I'm trying to come up with a scenario where it makes sense for Martin to be screaming for help, and coming up empty. Are we to believe that Zimmerman managed to catch up with a running and more nimble Martin, grab him and produce screams of fear from the teenager, by which time Zimmerman decided he'd had enough of his screaming and fired into him at point blank range?

Sure, if you believe that Zimmerman is a racist monster it makes sense, but not otherwise.

[fdcol63]

Can this thing devolve into any more absurdity?

[Ignatz]

--Can this thing devolve into any more absurdity?--

That question has been answered in the affirmative for a month now so I'd have to say, yes, yes it can.

[Captain Hate]

The dead horse flailing continues.

O/T Dr Yeeaarrrgh must've been self-medicating last night with something that doesn't give him that alkie skin-sheen on the morning after for his appearance on FNS; he still insists that El JEFe is in great shape regarding Bammycare and that striking down the mandate will be in his favor.

Btw am I the only person that thinks Santorum looks creepy when he smiles?

[jwest]

The Orlando Sentinel article with the two voice identification “experts” is fascinating. Why in the world would these people go out on a limb like that?

Tom Owen, from Owen Forensic Services LLC, said “"As a result of that, you can say with reasonable scientific certainty that it's not Zimmerman," That is a pretty definitive statement for someone who apparently makes a living doing this. With the contrary evidence (John and the initial statement of the 13 year old dog walker) along with what is actually being screamed (“Help me”, not “Don’t shoot”), it seems likely that Zimmerman will be found to be the screamer.

When this happens, what becomes of every case Owen ever testified in? If I were the lawyer in any case he ever brought his “expertise” bear in, I would be getting my appeal ready.

[Gmax]

A 48% match with lots of background noise and admittedly severe limitations due to the transmission and recording systems?

I dont know what this does for the prosecution, but I would say a jury of peers is not going to hear " the voice does not vett you must convict"

[Janet]

Clarice's Pieces is wonderful today! Capitol Follies

"If the legislative branch really wants to be respected in the courts and public eye, it's time for them to actually start reading what they legislate , quit passing fancy mandarin-generated nonsense that the unelected, barely overseen executive branch fills in as it damned well pleases, and start repealing and revamping legislation which has ceded to the executive branch far too much power to actually do the legislative work that is Congress' right and responsibility."

No Regulation Without Representation!!

[Danube of Thought]

Minus 14 at Raz today.

Leads Romney by 1.

[Threadkiller]

If you are against Martin, you are a Skittle denier.

If you are against Big Z, you are a Zither.

[Janet]

and via Instapundit, I love this post & it's question -
Where Are David Brooks & HuffPo's William Bradley Getting Their Story Ideas?

It seems as if liberal flotsam is in control. They write legislation, they feed stories to the MFM, they are the MFM touted "experts",....

[jwest]

I forgot to add in the evidence that it was Zimmerman on the tape the contemporaneous statement he made to the police that “I was yelling for someone to help me, but no one would help me.”

[jimmyk]

Owen said to reach a positive match with audio of this quality, he'd expect higher than 90 percent.

"Expect" or "require"? "Expect" makes it sound as if this is a high-quality test that should be a slam-dunk. I doubt that's what the guy was saying.

But the story ought to include a lawyer to tell us where the burden of proof lies and what the stadards for evidence might be.

As a statistician-type, I'd put it this way: The analysis should have as the "null hypothesis" (the one being tested) that the voice is Zimmerman's. You reject that hypothesis only if you're, say, 90% sure it's incorrect, since the burden of proof is to show Z is lying. This "expert" seemed to be doing it the other way around, saying it's not Z unless

I couldn't penetrate the jargon, but I were doing this test, I would put Z's voice through the same "filter"--record him shouting through a cell phone 20 yards away, or whatever it was, and then see whether what you get from that is of sufficient quality to learn anything from the biometric test. I have a feeling the answer would be no--surely not beyond a reasonable doubt.

[Jane]

I would guess Trayvon's purported girlfriend has some recordings of is angelic voice that could be analyzed.

[jimmyk]

That should have read "it's not Z unless he can be 90% sure it is."

[myiq2xu]

So let's say they find some clip of Trayvon talking and the same analysis shows a 46% match. Have they now ruled out both men?

[jimmyk]

I would also say that if whoever screamed actually pronounced all the consonants in the words he was shouting, it's probably Zimmerman.

Great pieces, Clarice. I have to say that Clement's reply here:

JUSTICE KAGAN: Wow. Wow. I'm offering you $10 million a year to come work for me, and you are saying that this is anything but a great choice?

MR. CLEMENT: Sure, if I told you, actually, it came from my own bank account. And that's what's really going on here, in part.

was brilliant. It totally exposes the shallowness of the left's thinking on "free stuff" coming from the government. And if he thought of that on the spur of the moment, I'm even more impressed. It's remarkable how few people think about where the money is actually coming from.

[myiq2xu]

Assuming that this case goes to trial and the judge rules that voice recognition evidence is admissible, I would expect the defense to bring it their own credentialed experts to say it was Zimmerman screaming.

There are prostitutes in every profession who perform for pay.

[Chubby]

if Jorge supplied a new tape with a sample of him replicating his scream on it, would that, if put through the gizmos, irrefutably prove it was him?

[rse]

At this point even if the story is absurd upon analysis, most readers do not analyze in the manner TM does. In print means true for most people.

Secondly a false but plausible story keeps the debate from seeming all one-sided now. Otherwise the extent to use this boy's death for political purposes based on race and emotion becomes painfully apparent. And the real story.

clarice-Thanks for covering that kagan exchange. As I noted to myself and alito apparently saw as well and commented on, the healthcare analysis on medicaid is quite analogous to education. You must adopt x policies to get money that the feds were previously providing without those strings. It uses money to bypass federalism.

One comment on the discussion with kathy on a previous thread last night. She accused us of not wanting to consult or learn from liberals. I spend most of my days in the belly of the beast reading what liberals say is the true intent for their recommended policies. I would be the last person to say I do not learn from liberals. In fact I even read their footnotes and bibliographies and then track down used the most relevant supporting documents for their policies.

And then I usually want to take a shower or at least wash my hands. Sometimes I sip wine while I reflect on explicit scheming.

So on another dismissive point she made, if I think many AGW pushers are global statists looking for a post-1989 control vehicle, it is because so many of them say so when they think only the true believers are listening or reading.

I don't wear a tin foil hat. With thick curly hair it wouldn't fit if I tried one. I don't go looking for conspiracies. But if people in a position to benefit say they are conspiring, here's the plan, and some of the participants, we do get to take them at their word.

[myiq2xu]

Did anyone consider the possibility that BOTH men were screaming and yelling at the same time?

[jwest]

Only a personal opinion, but listening to the 911 tape with the screaming, it seems to me that only one person is heard.

[Clarice]

Thanks, Jimmy and rse. Jimmy I agree that the Clement's reply was genius. Right to the heart of the matter. Yes, I am sure this was thought out ahead of time.

All the participants in the arguments spent days in moot courts where smart people anticipated the questioning. The govt did poorly I think, not because the SG wasn't prepared, but because he had no persuasive responses he could possibly have made .

[NK]

JimmyK-- many thanks for that Kagan/Clement quote. Clement is sharp and quick on his feet-- Kagan? not so much.

[Enlightened]

I said earlier that I thought GZ sounded effeminate - I think he would sound like a little girl screaming

And I bet there will never be a comparison voice test done using a TM voice message - his family knew it wasn't him and they can't walk that back now. I bet TM had a rather deep voice - his Dad dies.

OT somewhat - Is Sybilvia Nancy Grace's twin? Funny how closely the crap they spew is remarkably similar

[Chubby]

I enjoyed your Pieces this morning Clarice.

I'd like to see your cite that Jimmy posted above posterized. It would make a great piece of graphic art, big black letters on a white background, and up on millions of walls.

One more time:

JUSTICE KAGAN: Wow. Wow. I'm offering you $10 million a year to come work for me, and you are saying that this is anything but a great choice?

MR. CLEMENT: Sure, if I told you, actually, it came from my own bank account. And that's what's really going on here, in part.

[Threadkiller]

Wonderful Pieces, Clarice. This caught my attention:

"Let them stand in the well of the Congress and argue that the delta smelt means more to the world than the productivity of the Central Valley and those who live there."

Here is McClintock, on a different environmental concern, saying exactly that.(only 5min, watch to the end)

http://www.youtube.com/watch?v=r_iEtH_WqHU

And here he is on smelt regulations:

http://mcclintock.house.gov/2012/02/opening-statement-on-hr-1837-house-committee-on-natural-resources.shtml

http://mcclintock.house.gov/2012/02/statement-on-amendment-in-the-nature-of-a-substitute-hr-1837-house-committee-on-natural-resources.shtml

http://mcclintock.house.gov/2012/02/on-the-rule-for-hr-1837.shtml

[Enlightened]

**his Dad DOES**

[Danube of Thought]

"Owen said to reach a positive match with audio of this quality, he'd expect higher than 90 percent."

I don't understand that statement.

[NK]

Clarice@10:08 -- I agree with you entirely. Appellate advocacy isn't performance art-- it's answering questions persuasively-- conceding points that don't hurt you and hammering home your strong points. The poseur Lefties attacking the SolGen are comparing his and Clement's speaking styles-- OK Clement is especially smooth and verelli was choppy those days. But you know what, starting Tuesday morning the SCOTUS questions were playing out exactly as Clement hoped and verelli dreaded. That affects the lawyer's appearance. Verelli as SolGen can't pound on the table, so what can he do when things are going badly? I refuse to be optimistic about kennedy, but I can wish can't I?

[Threadkiller]

Ken Cuccinelli was on Hewitt this week as they played audio from the hearings. He said that Clement had no notes. He would just walk up and do his thing, without any reference material at all.

All I can say to that is wow, wow!

[NK]

TK-- if that's true-- I'm really impressed by Clement. he's not just smooth, he's incredibly prepared and sharp.

[Threadkiller]

Simple, DoT. The crappier the quality, the more accurate it needs to be.

Wait. What?

[sbw]

Excellent, Clarice.

And--attempting to deflect attention away from the pedantic lint analysis of navel gazing--he asks:

To read Clarice's summary of the genesis of the "law," why isn't Obamacare unconstitutional because it is vague, arbitrary, and subject to ex post facto interpretation. Obamacare is not so much a law as a placeholder for whatever those in power decide "health care" means after the fact.

[Threadkiller]

It is true that I heard Ken Cuccinelli say that, NK.

;-)

[NK]

OK- then when in legal call Paul Clement (wouldn't want to see the bill though)

[NK]

legal= legal trouble

[matt]

and last night, the First Lady of the United States allowed herself to be "slimed" on Nickelodeon. No respect for the dignity of her position.

[centralcal]

matt, matt, matt!

Slimed? That was the most positive thing about her "appearance." What in our dear Lord's name was she wearing?

That has to be one of her vilest outfits ever.

Utterly clueless. Utterly without class, or taste, or even self awareness.

[Ignatz]

--and last night, the First Lady of the United States allowed herself to be "slimed" on Nickelodeon--

Ew. I thought that was a kids network.

[Chubby]

((No respect for the dignity of her position.))

more than than, doesn't the fact that pelting adults with green slime is a high honor say something sad about kids today? We've regressed a long way from Lindley Murray, baby!

[Danube of Thought]

I have a hunch Clarice enjoyed writing that every bit as much as I enjoyed reading it.

Watch David Boies argue if you ever get the chance. He has no notes of any kind, and he never looks down at the podium. He keeps his eyes on the judges the entire time, while saying things like "this circumstance is not covered by Section 762(b); instead we have to look to Section 788(d)(3), which says blah blah blah..." Simply unbelievable.

[centralcal]

Well, Ignatz, she did take Sasha and Malia along for the show.

[Janet]

Obamacare is not so much a law as a placeholder for whatever those in power decide "health care" means after the fact.

Exactly, sbw.

[NK]

Boies-- I've seen Boies argue bland real estate litigation. He's so good, he makes it seem like the other side doesn't even have a case. He's the second best advocate I've seen in person-- the best? hate to say it, it was William Kuntsler.

[Threadkiller]

Has Michelle Obama had an adult life, for any actual "first times" to occur?

[Threadkiller]

He is good at the not so bland too, NK.

http://m.rollingstone.com/entry/view/id/11306/pn/0/p/0/?KSID=204f5d4a46bbb73af8f67bbe34e62bfa

[Ignatz]

--more than than, doesn't the fact that pelting adults with green slime--

Oh, green slime. Never mind.

[Jane]

Clarice,

There was so much in that Pieces to love. I loved all the democrats at the beginning displaying their complete ignorance of the constitution.

That should be a campaign ad somewhere.

[Ann]

Good Morning Patriots

This FLOTUS can't leave fast enough for me:

[Clarice]

Jane, it should be. I wonder if it's all been videotaped.
Ann, you beat me to it.

[Ann]

We can only pray it is this fast:

RUDE

[Enlightened]

My eyes my eyes - WTF is that?

[Ann]

Unbelievable C, I am off to read PIECES.

[Clarice]

DoT, it WAS great fun to write and an opportunity to get something off my chest that has been making me angrier every year. I know the people who drafted those environmental laws and I slam them every time we meet for confusing "species" with large mammals and destruction of the US economy with environmental protection.
And for decades I have worked with lawyers who think that there are absolutely no limits to federal power.

[Enlightened]

And I was just off to grab some Pepto. I think ipecac is more apropos. To wake up and see that....sigh.

[Danube of Thought]

The beauty of it is that, at least until Kennedy lets us down for the umpteenth time, we get to see those arrogant assholes in a state of shellshock. And so far as I am aware, not a one of them--Greenhouse, Lithwick et al.--has yet answered the questions they never bothered to ask themselves before the arguments, and that the Justices put to Clement.

[ThomasD]

I'd agree that Owen just threw any shred of credibility out the window.

While a 48% match may be nowhere near the threshold needed to match a voice to a person, it in no way can exclude a person.

But you don't need to know much about acoustic analysis to figure this out.

Compare it to DNA analysis. DNA, although profoundly complex, is a entirely known, and limited system. The number and type of codons are fixed, so there are simply only so many ways to arrange them. Meaning one can describe with a very high degree of accuracy the limits of any particular match. It is analogous to the limited number of possible numeric combinations of Powerball tickets.

Audio samples collected from an open environment, on the other hand, can contain any number of uncontrolled, and otherwise unidentifiable artifacts. So, in principle the universe of possible alternatives is unlimited. Meaning there is no chance in hell you can this disprove, or otherwise eliminate any one particular person from being present via this form of analysis.

'Not a match' does not equal 'can't be him.'

[Jane]

OMG Ann, what on earth is that outfit?

[Ignatz]

I have a few reservations, not many but some, about talking about the first lady's stature and fashion sense, but that get-up is really something else.
Is that a saddle blanket her trainer tossed over her hips?

[Frau Shikse]

I forgot to add in the evidence that it was Zimmerman on the tape the contemporaneous statement he made to the police that “I was yelling for someone to help me, but no one would help me.”

I want to be on jwest's team.

[ThomasD]

If it is a saddle blanket wouldn't it be better to speak of her hindquarters?

[Danube of Thought]

The insufferable Nina Totenberg reviews the arguments on NPR. Not much there.

[Danube of Thought]

Looks like something out of Mad Max or Rollerball.

[Manuel Transmission]

Clarice, 'cleanup in isle 13.'

Every Sunday, as soon as my eyes can focus, I reach for my phone and dial Pieces to get the morning off to a great start. Unfortunately, there is a bug in the formatting for smart phones that causes the text to be zoomed to a fixed (but random) size requiring the reader to scroll back and forth every line. I think the softies call that a loose variable. I don't know if you have your own copy of a formatted file you load your wisdom into, or whether they do it for you at AT, but it seems to be unique to your column. I emailed AT a couple of weeks ago, but no change.

[Dave Turson]

This reminds me of the brilliant expert who claimed OJ's footprints were left on the sidewalk of the murder scene, but it turned out the footprints in question were just imprints left in wet concrete when the sidewalk was poured.

[PaulL]

Think of the world class experts who testified for O.J. They were eager to mislead the jurors and did not seem to worry about risking their reputations.

[Danube of Thought]

Well, it certainly appears that the science is settled for this babe:

An Oregon University professor has controversially compared skepticism of global warming to racism.

Sociology and environmental studies professor Kari Norgaard wrote a paper criticising non-believers, suggesting that doubters need to be have a ‘sickness’.

The professor, who holds a B.S. in biology and a master's and PhD in sociology, argued that ‘cultural resistance’ to accepting humans as being responsible for climate change ‘must be recognised and treated’ as an aberrant sociological behaviour.
Controversial: Kari Norgaard is a professor of sociology and environmental studies at Oregon University

Resolving skepticism about climate change alarmists, she added, is a challenge equitable to overcoming ‘racism or slavery in the U.S. South’.

Read it all here--and check out her picture.

[jwest]

Here is the comment I just left at Ed Primeau’s website. He’s the other voice recognition expert cited in the Orlando Sentinel’s article.

1. Mr. Primeau,

However this turns out, everyone should admire your courage for putting your professional opinion out in public. Not being familiar with the intricacies of voice analysis, I hope you will go into as much detail as possible so that we can understand the science involved.

Considering that two witnesses at the scene indicated that Zimmerman was on the bottom during the altercation and was the person screaming, that he made a contemporaneous statement directly after the incident stating that “I was yelling for someone to help me, but no one would help me”, that Martin’s father initially denied that the voice on the tape was Trayvon’s and finally, what the words being screamed were – apparently more consistent with someone being beaten as opposed to someone begging for someone not to shoot.

To be able to ascertain with the level of certainty that you have that it was Trayvon Martin’s voice on the tape in the face of all the contrary evidence is something that will, of course, receive a great deal of scrutiny. Personally, even if I possessed your knowledge and experience, I don’t know if I would have courage to place my entire professional reputation at risk.

Please continue posting as much information as possible about the specifics of this case.

[Frau Aprilejeck]

Ann @ 11:20, Pres. Assburgers had just said, "You're my closest ally and you're punchin' above your weight."

[Rick Ballard]

TM mentioned and linked it but if the contemporary statement by 'John',

"The guy on the bottom, who had a red sweater on, was yelling to me, 'Help! Help!' and I told him to stop, and I was calling 911," said the witness, who asked to be identified only by his first name, John.

John said he locked his patio door, ran upstairs and heard at least one gun shot.

"And then, when I got upstairs and looked down, the guy who was on the top beating up the other guy, was the one laying in the grass, and I believe he was dead at that point."

confirming GZ's statement to the police isn't a hallucination induced by an overdose of Skittles, I don't understand the intent of the Orlando Sentinel in publishing this trash. It's just fill without value.

It's almost as if there were no such thing as journalistic standards.

[Threadkiller]

I have learned something valuable after going to DoT's link, despite knowing better.

If he instructs me to jump off a bridge, I probably would.

[PaulL]

Yes, Dave, the so-called expert, Dr. Henry Lee, was a real POS.

But FBI Special Agent William Bodziak, who analyzes footprints for the agency's crime lab, testified that what Lee characterized as impressions on the walkway actually are the tool marks and footprints left by the workmen who originally poured it. As it turns out, the walk--which previous testimony consistently described as tiled--actually is a poured slab of colored concrete inscribed and grouted to create the sort of faux-tile effect frequently used in so-called Southwestern landscaping schemes.

Of course the prosecution went on for hours with this and quickly lost the jurors, whose attention span was probably measured in seconds.

[Clarice]

jwest, very funny.

MT I don't know what can be done about that and I am sorry to hear of your problem. Have you tried hitting the print button at the top of the page to see if the print version ((you needn't print it to read it that way)is better.

[Janet]

from DoT's link -

"Kari Norgaard is a professor of sociology and environmental studies at Oregon University"

[myiq2xu]

Is voice recognition evidence even admissible? Will it pass the Daubert test?

The trial judge performs a “gatekeeping” role in excluding unreliable testimony. The United States Supreme Court first addressed the reliability requirement for experts in the landmark case Daubert v. Merrell Dow Pharmaceuticals, Inc. 509 U.S. 579 (1993). The Court laid out four non-exclusive factors that trial courts may consider when evaluating scientific expert reliability: (1) whether scientific evidence has been tested and the methodology with which it has been tested; (2) whether the evidence has been subjected to peer review or publication; (3) whether a potential rate of error is known; and (4) whether the evidence is generally accepted in the scientific community. Id. at 592-94. Kumho Tire Co., Ltd. v. Carmichael later extended the Daubert analysis to include all expert testimony. 526 U.S. 137 (1999).

[jwest]

I can understand the Orlando Sentinel publishing this trash, but for the life of me, I can’t understand these two voice recognition “experts” putting their reputations and livelihood on the line.

This Ed Primeau, the second expert, says he stumbled on the link to the 911 calls while reading about this in Mother Jones. Naturally, the Mother Jones article makes it pretty clear Zimmerman is your typical White-Hispanic racist gun toting killer.

What could this guy and Thomas Owen, who it appears has even more to lose, possibly be thinking?

[PaulL]

Superb comment, jwest!

[myiq2xu]

and check out her picture.

I just had a flashback to my high school prom.

[sbw]

Isn't reading Mother Jones an automatic DQ?

[bgates]

Not much there.

Look at the quotes from Kennedy that she holds up as emblematic of the two sides (of his mind and the case):

Here, the government is saying that the federal government has a duty to tell the individual citizen that it must act, and that is different from what we have in previous cases. That changes the relationship of the federal government to the individual in a very fundamental way.

vs
In the insurance and health care world...the young person who's uninsured is uniquely, proximately very closer to affecting the rates of insurance and the cost of providing medical care in a way that is not true in other industries.

It boggles the mind that any adult American outside of an asylum considers those to be equally compelling positions. The latter is just a mess. "uniquely, proximately very closer"? Is there any reason to believe that a young person who's uninsured is "closer to affecting the rates of insurance" than in other industries? If so, is there any reason to think that's a problem? If there is reason to think that's a problem, is there any reason to think Obamacare will solve the problem? And if there is a reason to think Obamacare will solve the problem of a young person who's uninsured being uniquely, proximately very closer to affecting the rates of insurance and the cost of providing medical care in a way that is not true in other industries, is there reason to think Obamacare is doing so constitutionally?

By contrast, look at the former quote. If the Constitution defines the relationship between the federal government and the individual, and Obamacare changes that relationship in a very fundamental way, doesn't Obamacare have to be unconstitutional?

[jwest]

I just noticed that my comment on Primeau’s site is still “awaiting moderation”.

Let’s hope I sugar-coated it enough that he lets it through.

[Frau Aprilejeck]

And the prints that happened to match the "ugly-ass" shoes that OJ swore he would never have worn but was photographed wearing? I always liked that touch. The jury was more impressed by Dr.Lee demonstrating blood splots.

In this case, we already have Sharpton and Jackson as the Twin Devils of Deception.

[Rob Crawford]

confirming GZ's statement to the police isn't a hallucination induced by an overdose of Skittles, I don't understand the intent of the Orlando Sentinel in publishing this trash.

Incitement.

[Frau Aprilejeck]

jwest, I'm glad someone else remembers that Mr. Martin said on record that the voice was not his son's. 75k expected in today's march of FL no-peace-no-justicers.

btw - who get's Sharpton's salary from his TV gig? I thought he did not own-own anything due to the large fine he was given for his part in the Tawana Brawley hoax.

Ewww - scary TV of Hillary! speaking in her--choppy--cadence*--while--seated--with Saudi--official-- who decided to sit-- at the extreme-- other end-- of --the table.

*the result of speech lessons to eliminate her constant "uh" and "you know" from her speech when she ran for, uh, you know, president.

[Danube of Thought]

When I say "not much there," I mean she doesn't tell us anything we haven't been made aware of at great length. Those two passages are probably the ones most frequently quoted in the post mortems.

And yes, the first quote certainly suggests to me that the "fundamental change" mandates a holding of unconstitutionality. But this is Anthony Kennedy.

[Ignatz]

--It boggles the mind that any adult American outside of an asylum--

That would appear to exclude Dr. Kari Norgaard, if her picture is any indication.

[myiq2xu]

I would like these experts to pass a test:

Send them five sets, each set containing 1 recording of a person screaming and four recordings of people speaking in a normal tone of voice. In 4 of the 5 sets there there will be a match. In the 5th set there is no match.

Use blind testing methods and see how accurate the results are.

[Ann]

C, that was another masterpiece. If we could only demand the supreme court justices read it.

[matt]

Clarice;

Try this.

Down in Texas some ranchers have tried to make their places more profitable and imported endangered species of antelope like Oryx from Africa and Asia. These species are disappearing in their home ecosystems because of poaching, encroaching development, etc.

The Texas ranchers have nurtured their herds and these herds have grown through breeding by 400% in just the past 5 years. The antelope roam and seem to like the lack of predators and feed.

The ranchers charge hunters ridiculous amounts of money to bag some every year.

So now the animal rights activists are ramming through regulations to "stop the destruction of endangered species"

Wanna try and figure out the logic of that one?

[Jane]

Let’s hope I sugar-coated it enough

Speaking of sugar coating, I want to candy some pecans and I have no eggs or milk. Anyone have any ideas?

[Danube of Thought]

Crikey--Fox News Channel has been taken off DirecTv.

[Danube of Thought]

Why was Congresswoman Frederika Wilson wearing a pink cowboy hat?

[Manuel Transmission]

Clarice, this problem is exclusive to the 'mobile' version that the site automatically reverts to when it gets the request from the user's browser. Not sure whether the folks with iPads have the same problem. Knowing a little bit about software testing, I know that it is a pain to cover all bets. It may, for instance, work fine on a Blackberry, but not on an iPhone.

[sbw]

Jane:

1) look in your cupboard for a can of evaporated or condensed milk.

2) go to the nearest 7-11 and trade in some skittles for milk.

[Cecil Turner]

There's some youtube on Easy Voice Biometrics that shows how the software works. In the first, it matched two audios of Richard Nixon giving speeches (with some audio issues) at 86%. The third showed some disguised voice attempts (but with remarkably similar tone and speech) and matched at 90+ percent. None of them were close to the variation in screamed cries for help versus a relatively calm 911 call.

I have no idea whether the voice match reliability is sufficient to draw a conclusion, or whether the 48% number is consistent with the variation due to the known circumstances (stress, yelling, audio). But so far, color me unimpressed.

[Jane]

Oh I do think I have some evaporated milk. Nice job sbw.

I drove 102 miles this morning for peanut butter. I figure it cost me about $12 in gas. I'm housebound to make up for it.

[henry]

Clarice, the problem MT describes also happens on my iPhone. It started last week, and may correspond to a phone software upgrade. I don't have the details anyone would need to test / fix the problem. (hangs head in shame as software CEO who gets daily problem reports with insufficient documentation).

[centralcal]

Why was Congresswoman Frederika Wilson wearing a pink cowboy hat?

Because she is a rodeo clown? If you check out these Bing images of her, hats in loud colors and humongous flower pins appear to be her fashion statement.