Trayvon Martin, Vigilante Justice

[sgtbaker]

The Tribune asked two different experts using two different methods. They are experts in the field and have been qualified to testify at trials. I think that's pretty strong evidence that it isn't Zimmerman screaming. You can keep an open mind, but not so open your brain falls out.

This is the nail in the coffin. Listen to that tape again. If it wasn't Zimmerman, then Trayvon is the one screaming in terror. And Zimmerman lied about it. If Trayvon is screaming for help like this, that makes Zimmerman a monster. Not a manslaughter case, but a murder case.

It was the Orlando Sentinal that did this, which is why I am not so quick to discount it. They seemed to be the least biased in their reporting and the most factual. The only thing I find annoying is they post a video showing the voices being analyzed but the video doesn't contain the score.

http://articles.orlandosentinel.com...cation-expert-reasonable-scientific-certainty

 

[RandFan]

Well that revelation (voice analysis claim) was a stick in a hornets nest.

 

[sgtbaker]

What am I discounting? The video? The leaked police report?

You said you saw the incident reports but act like no such evidence of Zimmerman's bleeding head exists.

 

[TheL8Elvis]

Zimmerman acted in contradiction to what the police dispatcher advised him, and in violation of the rules of the neighborhood watch program with which he claims to be affiliated.

You don't have any obligations to answer any particular questions at all.


re:Zimmerman acted in contradiction to what the police dispatcher advised him, and in violation of the rules of the neighborhood watch program with which he claims to be affiliated.
As pertaining to the watch program:

http://www.npr.org/2012/03/21/149086792/shooting-renews-debate-on-vigilance-vs-action

"Although Mr. Zimmerman apparently was not part of any official neighborhood watch organization, even if he had been, these folks don't have any more power than ordinary citizens," says Paul Butler, a law professor at George Washington University. "They are not law enforcement officials."

About 25,000 watch groups are registered with the sheriffs' association, which started its neighborhood watch program in 1972. Many other groups are registered with local law enforcement.

In a statement this week, the sheriffs' association referred to Zimmerman as a "self-appointed neighborhood watchman," and said it had no record that his community had ever officially registered a neighborhood watch. So it is unclear whether he would have been familiar with the NSA's guidelines.

Which claim was that? Please provide a quote.

The hilited claim ... that I already addressed

 

[WildCat]

Forensics experts say the voice screaming for help on the 911 call does not belong to Zimmerman.

Uh, yeah... sounds like woo to me. I'm sure it's as reliable as polygraphs.

 

[TheL8Elvis]

The Tribune asked two different experts using two different methods. They are experts in the field and have been qualified to testify at trials. I think that's pretty strong evidence that it isn't Zimmerman screaming. You can keep an open mind, but not so open your brain falls out.

http://articles.orlandosentinel.com...cation-expert-reasonable-scientific-certainty
A leading expert in the field of forensic voice identification sought to answer that question by analyzing the recordings for the Orlando Sentinel.

No, they were workign for the media, and we have no reason to question the medias motive in any of this

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

Science, FTW. Please pardon me if I remain skeptical of the hard science of software called Easy Voice Biometrics.

I will say, if the screaming is clearly identified as Martin at trial, that probably changes my thoughts on events. (I assume the parents have videos and even his cell phone voice mail to get samples.)

Funny but possibly irrelevant: The Easy Voice Biometrics domain is registered by Owen LLC. Which happens to be the company of Owens's daughter where he happens to work.

So - Owens is pimping his own software, that may never have actually been used in court. And the paper does not disclose the relationship between Owen and the software.

Can I be a little skeptical of this ?

 

[WildCat]

It tells us GZ was lying.

It tells us the woo peddlers have another totally scientific machine to play with.

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.

A total BS claim. You simply cannot take a mono recording of such poor quality and do what he claims he has done. And it's also a lossy recording, much data has been lost. Complete woo.

 

[AvalonXQ]

Uh, yeah... sounds like woo to me. I'm sure it's as reliable as polygraphs.

Here's a paper I came across that discusses the current problems in Speech Recognition research.
It seems that this situation has all the hallmarks of what the paper discusses as causing problems with the accuracy of current technology.

 

[totoro]

You said you saw the incident reports but act like no such evidence of Zimmerman's bleeding head exists.

The incident report doesn't seem like evidence to me without some kind of real visual evidence, like photos, or testimony from someone outside the police department, like the EMS people.

Right now, one possible hypothesis (which is most likely wrong, but it is still viable) is that we have a bunch of cops trying to protect the son of a retired judge. Given that this is at least not out of the realm of reason, I would hope that the police would have aimed for a bit more transparency, and would have at least leaked some evidence that their own claims were true when they leaked their report. They chose not to.

Maybe there was ill intent here, maybe there wasn't, and maybe they did their jobs exactly correctly. But they sure haven't done any kind of job of making their behavior transparent to the public.

This whole thing smells like week old fish.

 

[TheL8Elvis]

Right now, one possible hypothesis (which is most likely wrong, but it is still viable) is that we have a bunch of cops trying to protect the son of a retired judge. Given that this is at least not out of the realm of reason,

His father is a retired magistrate - not judge - in VA. So, does that make it more out of the realm ?

 

[sgtbaker]

It tells us the woo peddlers have another totally scientific machine to play with.



A total BS claim. You simply cannot take a mono recording of such poor quality and do what he claims he has done. And it's also a lossy recording, much data has been lost. Complete woo.

I sit out on this aspect because I know very little about the biometric voice recognition software. After I mentioned using voice recognition software a while back I looked into it. The method the second guy used was very problematic in that they don't have two voice recordings of similar environments. Attempting manipulations and eliminating background noise can degrade the quality of the recording, thus contaminating the results. However, two independant analysts coming to similar conclusions is hard to ignore.

As an aside, your freakin cat winked at me. I never noticed that thing winks and it felt so... personal.

 

[TheL8Elvis]

Well that revelation (voice analysis claim) was a stick in a hornets nest.

Are you suggesting the media stirred up a hornets nest regarding this case ? Never !

 

[WildCat]

As an aside, your freakin cat winked at me. I never noticed that thing winks and it felt so... personal.

 

[sgtbaker]

The incident report doesn't seem like evidence to me without some kind of real visual evidence, like photos, or testimony from someone outside the police department, like the EMS people.

Right now, one possible hypothesis (which is most likely wrong, but it is still viable) is that we have a bunch of cops trying to protect the son of a retired judge. Given that this is at least not out of the realm of reason, I would hope that the police would have aimed for a bit more transparency, and would have at least leaked some evidence that their own claims were true when they leaked their report. They chose not to.

Maybe there was ill intent here, maybe there wasn't, and maybe they did their jobs exactly correctly. But they sure haven't done any kind of job of making their behavior transparent to the public.

This whole thing smells like week old fish.

You are entitled to believe whatever you want. So, did you see the incident report before or after you asked me? It's just weird that you would ask me as if I made it up out of the blue, all the while knowing that there is an incident report that contains the description of Zimmerman's bleeding head.

 

[totoro]

URL_DELETED_HERE a paper I came across that discusses the current problems in Speech Recognition research.
It seems that this situation has all the hallmarks of what the paper discusses as causing problems with the accuracy of current technology.

Not sure what makes you say that. I don't do Speaker Recognition stuff for a living, but I do machine learning stuff for a living, and understand at least the basic models they are using (although the supervector stuff is new to me, and looks kind of interesting), and don't see anything in that paper that specifically addresses anything relevant here (other than general stuff about feature selection and data normalization).

As far as I can tell, what they are talking about is building a classifier like an SVM or Gaussian Mixture Model from examples of the target speech, and then running examples through the resulting model, much as you would for any other classifier (after a whole bunch of feature selection and normalization preprocessing).

Generally, when you build a classifier, you can tell when you build it roughly how accurate it is, and you get a score from running an instance through it that tells you how close to being "in class" it is, so they can at least give a confidence estimate.

Without seeing the actual data and models, I couldn't tell how good these models are. But this general field is not pure ******** on the order of polygraphs.

 

[totoro]

You are entitled to believe whatever you want. So, did you see the incident report before or after you asked me? It's just weird that you would ask me as if I made it up out of the blue, all the while knowing that there is an incident report that contains the description of Zimmerman's bleeding head.

Sorry, I wasn't trying to be a dick. I had only seen media reports of the leaked police report. Since you seem to have been following this pretty closely, I thought that my information might have been outdated, and you had seen an official police report.

Nothing more sinister than that.

 

[crimresearch]

The incident report doesn't seem like evidence to me without some kind of real visual evidence, like photos, or testimony from someone outside the police department, like the EMS people.

Right now, one possible hypothesis (which is most likely wrong, but it is still viable) is that we have a bunch of cops trying to protect the son of a retired judge. Given that this is at least not out of the realm of reason, I would hope that the police would have aimed for a bit more transparency, and would have at least leaked some evidence that their own claims were true when they leaked their report. They chose not to.

Maybe there was ill intent here, maybe there wasn't, and maybe they did their jobs exactly correctly. But they sure haven't done any kind of job of making their behavior transparent to the public.

This whole thing smells like week old fish.

Please don't take this the wrong way, but this thread is well over 4,000 posts and links to the little bit of factual information that is available have been posted over and over.

Zimmerman's father wasn't a judge, he held a position in another state that most places would call Justice of the Peace, there is no good reason to think that would give his son any clout with a dead body laying there.
When Zimmerman was arrested for pushing a cop, the diversion deal he got is the same kind that thousands of kids get every spring break, maybe even more severe. No real evidence of conspiracy level strings being pulled.

The police might have dropped the ball on evidence collection, but we don't know, because they turned it over to the state who wants it kept quiet.

We do know that they filed charges against Zimmerman contrary to early reports that they 'just let him go'.

There seem to have been minor injuries to Zimmerman's head and nose, but we don't know more than 'EMS did something' at the scene.

The timeline seems to have been pinned down to within a few minutes, and from start to finish, it happened pretty fast.

The eyewitnesses *as reported* are not in total agreement about who was on top of whom.

The 2 gunshots have turned into 1 and the bullet count speculation isn't inconsistent with normal mechanisms of that particuar weapon. Likewise for the holster speculation.

 

[Skeptic Ginger]

Uh, yeah... sounds like woo to me. I'm sure it's as reliable as polygraphs.

Doubling down when you should be reevaluating, yep.

Spectrographic voice identification: A forensic survey; Bruce E. Koenig; Federal Bureau of Investigation

A survey of 2000 voice identification comparisons made by Federal Bureau of Investigation (FBI) examiners was used to determine the observed error rate of the spectrographic voice identification technique under actual forensic conditions. The qualifications of the examiners and the comparison procedures are set forth. The survey revealed that decisions were made in 34.8% of the comparisons with a 0.31% false identification error rate and a 0.5 3% false elimination error rate. These error rates are expected to represent the minimum error rates under actual forensic conditions.

Pay close attention to those decimal points.

 

[RandFan]

Not sure what makes you say that. I don't do Speaker Recognition stuff for a living, but I do machine learning stuff for a living, and understand at least the basic models they are using (although the supervector stuff is new to me, and looks kind of interesting), and don't see anything in that paper that specifically addresses anything relevant here (other than general stuff about feature selection and data normalization).

As far as I can tell, what they are talking about is building a classifier like an SVM or Gaussian Mixture Model from examples of the target speech, and then running examples through the resulting model, much as you would for any other classifier (after a whole bunch of feature selection and normalization preprocessing).

Generally, when you build a classifier, you can tell when you build it roughly how accurate it is, and you get a score from running an instance through it that tells you how close to being "in class" it is, so they can at least give a confidence estimate.

Without seeing the actual data and models, I couldn't tell how good these models are. But this general field is not pure ******** on the order of polygraphs.

It would be great to do a double blind test of the technology.

 

[totoro]

Please don't take this the wrong way, but this thread is well over 4,000 posts and links to the little bit of factual information that is available have been posted over and over.

Zimmerman's father wasn't a judge, he held a position in another state that most places would call Justice of the Peace, there is no good reason to think that would give his son any clout with a dead body laying there.
When Zimmerman was arrested for pushing a cop, the diversion deal he got is the same kind that thousands of kids get every spring break, maybe even more severe. No real evidence of conspiracy level strings being pulled.

The police might have dropped the ball on evidence collection, but we don't know, because they turned it over to the state who wants it kept quiet.

We do know that they filed charges against Zimmerman contrary to early reports that they 'just let him go'.

There seem to have been minor injuries to Zimmerman's head and nose, but we don't know more than 'EMS did something' at the scene.

The timeline seems to have been pinned down to within a few minutes, and from start to finish, it happened pretty fast.

The eyewitnesses *as reported* are not in total agreement about who was on top of whom.

The 2 gunshots have turned into 1 and the bullet count speculation isn't inconsistent with normal mechanisms of that particuar weapon. Likewise for the holster speculation.

Mea culpa. It got hard to weed out the actual data in all the name calling etc, and I didn't read the thread carefully enough.

Honestly, the post you just made contains more facts that I can at least trust as verifiable/falsifiable than pretty much everything I have seen in the media so far.