Trayvon Martin, Vigilante Justice

[totoro]

Here is a link to a youtube video giving a demonstration on how Easy Voice works.

pt 1


pt 2


pt 3


ETA: apparently I can't count because it's three videos.

I saw those. They don't actually tell me how it works. I want to know what kind of models they used, how feature selection was done, what kind of standardization was done, how one evaluates the quality of a model once built, etc.

I suppose one could do the equivalent of generating a confusion matrix at very small scale by running a small number of samples through the software, and seeing how the numbers work out.

But, for instance, when he says in the video "anything over 85% is very good", I don't really know what that means without seeing what the accuracy of the thing over some larger corpus of examples was.

If what they did was build statistical models, then they had to have already done this, so it would be easy enough to just say.

Perhaps what they did was nothing like what was in that paper from Finland, and that isn't relevant at all, but then it would be nice to know what exactly the software does.

All of this isn't to say the software doesn't work. But it would be a lot easier to say if they simply made some of this information available.

 

[RandFan]

I'm an expert in two categories, one being common sense. Common sense tells me that someone can't be an expert in an area where they claim no expertise in and where no standards for expertise have been established.

I get this a lot from creationists and theists. It's a soft solipsistic response. It's just an appeal to your own incredulity and assumes that others are not capable of common sense.

I don't think everyone should just accept my assertions. Where are you getting that from? Everyone can, and will, think whatever they want to. I give nothing by my opinion.

It would be helpful if you wouldn't assert categorically what is true. You assert that the experts must be dismissed because you are an expert in common sense.

 

[Alt+F4]

What it does is stated on the web page I linked to above.So basically all the hand waving dismissals are also ignoring the fact a thorough evaluation is made in addition to using this software.

Not sure what you mean by a "through evaluation" considering the fact that we don't know the source of Mr. Owen's copy of the 911 calls (YouTube? police?) and we don't know how Mr. Owen treats a scream differently from a normal speaking voice.

What we do know for certain is that Mr. Owen testified as a prosecution witness in the recent murder trial of Sheila Davalloo. He testified that it was Ms. Davalloo who made a 911 call to the police. He testified that he made this determination based on comparing the 911 call with subsequent multiple samples of Ms. Davalloo's voice, two years after the phone call.

We have zero evidence that Mr. Owen or anyone else on planet earth can determine with scientific certainty who is and who isn't screaming in a 911 call.

 

[Alt+F4]

I get this a lot from creationists and theists. It's a soft solipsistic response. It's just an appeal to your own incredulity and assumes that others are not capable of common sense.

Well Rand, I'm neither a creationist or a theist.

It would be helpful if you wouldn't assert categorically what is true. You assert that the experts must be dismissed because you are an expert in common sense.

I assert that the experts should be dismissed because they aren't experts. It's like saying that a dentist is an expert in heart disease just because both have dr. in front of their name.

And yes, I am an expert in common sense.

 

[RandFan]

I assert that the experts should be dismissed because they aren't experts.

I assert this is nonsense.

It's like saying that a dentist is an expert in heart disease just because both have dr. in front of their name.

Fatuous.

And yes, I am an expert in common sense.

And solipsistic.

 

[Alt+F4]

Frankly I'm getting tired of wasting my time.

Then don't.

He is one of the top audio forensics analysts in the country. He's published, he's involved in accreditation of other audio experts, he clearly believes in strict scientific principles. But go ahead, dismiss him as a quack. Frankly it makes you, WC and AvalonXQ look bad, not me.

Please link to where I called him a quack. Oh that's right, you can't.

Again, can you cite even one case where Mr. Owen or anyone else on planet earth was ever able make a voice determination comparing screaming in a 911 call with a person's normal speaking voice in another 911 call?

Oh and by the way, I think Zimmerman is guilty of murder.

 

[totoro]

Here's another company that does this work. As far as I can tell, this branch of forensics is routinely used in court cases and is widely accepted.

http://www.audioforensics.com/index.htm

How about someone email them and ask these questions?

As far as I'm concerned, this sudden skepticism of this particular brand of forensics is just convenient hand waving. If the two experts had concluded the voice to be Zimmerman I have no doubt WildCat would be touting this as a major break in the case.

I agree with you. I just wanted to point that out given that I've expressed my annoyance that the Easy Sound people chose not to tell anyone how exactly their software works.

Having spent a couple hours last night skimming papers, it seems to me that the field at present is pretty much an application of well-known signal processing and machine learning techniques. No woo, no polygraphy or other "science" founded on total nonsense. Just a bunch of EE stuff (that I don't know too well) mixed with a bunch of statistics/applied math stuff (that I do know well).

That being said, it wouldn't kill them to at least tell us how exactly their software works. Maybe they just don't want a bunch if CS/EE people from the other side niggling about specifics of their models in cases to create FUD.

 

[crimresearch]

Meaningless platitude; it's not evidence of presence either. It is exactly what it is: absence of evidence.



Except when they admit to the actual act that would be illegal, but assert there's a mitigating circumstance.



The facts of the state's case as I've given them are supported by evidence which would be introducable.



Fine? I don't think I mentioned Zimmerman's past history.



We were talking Manslaughter, weren't we?



Again, in my post outlining my reasoning I don't believe I mentioned any of these things.



Again, not considered.



See above.



Anybody has the right to assert anything.



I'm not sure that's strictly true. I think if the prosecution can convince the jury that Zimmerman dove into the pond with his eyes fully open, as it were, the jury might not care as much that he didn't know how to swim.

No, this isn't about manslaughter at all.


And that kind of pretzel logic should be sold at the mall, with dipping sauce.

In this case, the act alone is not illegal, the crime is... and as pointed out to you over, and over, and over, the law requires specific elements to prove the crime.
No matter how badly you want it to be so, an affirmative defense isn't an admission of guilt, and self defense isn't 'mitigating circumstances'.

Either this is about the law or it isn't.

I simply don't see the utility of making claims about what the law is, and when it is pointed out that they are wrong, simply switching to 'I'm not talking about in a courtroom, I'm talking about 'objective evidence', and then proceeding to roll that right back to talking about guilt.

Guilt will be determined in the future, not here.

 

[Alt+F4]

That being said, it wouldn't kill them to at least tell us how exactly their software works. Maybe they just don't want a bunch if CS/EE people from the other side niggling about specifics of their models in cases to create FUD.

Yeah, I don't get that either. Obviously they have a patent on the software and I would think they would have to explain how it works to a jury.

 

[Alt+F4]

Guilt will be determined in the future, not here.

Will it though? I just don't see an indictment happening here. It comes down to law=rules, vs. justice=fairness. The law will prevail, justice will fail.

 

[totoro]

What it does is stated on the web page I linked to above.So basically all the hand waving dismissals are also ignoring the fact a thorough evaluation is made in addition to using this software.

Sorry, I should have been clearer. I understand what the purpose of the software is, which is what it does at a very high level.

I'm not even disputing that it works. But as someone who works in a related field and knows the math involved, it frustrates me that they don't at least say what the models are, and how accurate they are.

There may be good reasons why they do this. But I still find it frustrating.

 

[crimresearch]

Will it though? I just don't see an indictment happening here. It comes down to law=rules, vs. justice=fairness. The law will prevail, justice will fail.

You don't see enough to make a ham sandwich (grand jury finding)?

Or do you think some other factor will prevent an indictiment?

 

[WildCat]

Yeah, I don't get that either. Obviously they have a patent on the software and I would think they would have to explain how it works to a jury.

I'd be shocked if that nonsense is allowed anywhere near the jury of any Zimmerman trial.

The test result (no match) is indistinguishable from the result if the software/system didn't work at all.

 

[RandFan]

I agree with you. I just wanted to point that out given that I've expressed my annoyance that the Easy Sound people chose not to tell anyone how exactly their software works.

Having spent a couple hours last night skimming papers, it seems to me that the field at present is pretty much an application of well-known signal processing and machine learning techniques. No woo, no polygraphy or other "science" founded on total nonsense. Just a bunch of EE stuff (that I don't know too well) mixed with a bunch of statistics/applied math stuff (that I do know well).

That being said, it wouldn't kill them to at least tell us how exactly their software works. Maybe they just don't want a bunch if CS/EE people from the other side niggling about specifics of their models in cases to create FUD.

I think that's fair. BTW: I found an excellent article, Voiceprint Identification.

Court Admissibility

Court testimony involving aural- spectrographic voice comparison essentially started having an impact on the courts after the Tosi Study in December 1970. Since then there have been between 150 and 200 trials in local, state or federal courts. Because of a difference based on evidentiary philosophical reasons, some courts have admitted aural-spectrographic voice evidence and others have not.

There are two general "rules" or "standards" by which scientific evidence is accepted in courts of law in the United States. The first, commonly referred to as the Frye "rule" or "test," is based on a 1923 District of Columbia case and basically requires "general acceptance in the particular field in which it belongs." See Frye v. United States, 54 App. D.C. 46, 293 F. 1013 (1923). The second is based on the argument of McCormick (See "McCormick on Evidence," 3rd Ed., 203 at 608.) McCormick states: "General scientific acceptance is a proper condition for taking judicial notice of scientific facts, but it is not a suitable criterion for the admissibility of scientific evidence. Any relevant conclusion supported by a qualified expert witness should be received unless there are distinct reasons for exclusion." See Rule 702 of the Federal Rules of Evidence.

 

[crimresearch]

I think that's fair. BTW: I found an excellent article, Voiceprint Identification.

There are two general "rules" or "standards" by which scientific evidence is accepted in courts of law in the United States. The first, commonly referred to as the Frye "rule" or "test," is based on a 1923 District of Columbia case and basically requires "general acceptance in the particular field in which it belongs."

Hmmm... wonder if anyone here would be able to find out what standard Florida uses?

 

[totoro]

I think that's fair. BTW: I found an excellent article, Voiceprint Identification.

Someone posted this link last night:

http://cs.joensuu.fi/pages/tkinnu/webpage/pdf/speaker_recognition_overview.pdf

It has a pretty nice overview of the field, plus a good bibliography. It's somewhat technical, but pretty interesting.

 

[totoro]

Yeah, I don't get that either. Obviously they have a patent on the software and I would think they would have to explain how it works to a jury.

Most software doesn't have any patents associated with it. I wouldn't assume right off the bat that they have any.

Often, you can get really good results in fields that have lots of prior art, if you combine previous results in a new way or just package things up better than your competitors.

 

[Skeptic Ginger]

Please link to where I called him a quack. Oh that's right, you can't.

WC did. I am conflating the three of you since you are all making the same argument.

I think it would be incredibly damaging to Zimmerman's case, but this assclown with his magical software sure as hell didn't prove anything except Barnum's Theorem.

I don't need to cite evidence X just because you dismiss evidence Y. Evidence Y supports my conclusion.

 

[Kevin_Lowe]

I'd be shocked if that nonsense is allowed anywhere near the jury of any Zimmerman trial.

The test result (no match) is indistinguishable from the result if the software/system didn't work at all.

That makes no sense at all. If a DNA test rules out a match between two samples do you think that is indistinguishable from the result if the system doesn't work at all?

I've got no strong opinion either way as to whether this voice analysis business is sound science or not, by the way.

 

[Skeptic Ginger]

....

That being said, it wouldn't kill them to at least tell us how exactly their software works. Maybe they just don't want a bunch if CS/EE people from the other side niggling about specifics of their models in cases to create FUD.

Yes because they owe all these people on Internet forums their time.

I'm sure if you were interested in purchasing this software the company would be happy to provide you with whatever evidence you asked for.