Much higher than it was 5 years ago...
Sure they've been used. I listed two major brands of DSP boards in the 80's and 90's. Apparently you know very little about this at all.
Uh? You're dealing with raw audio here. Custom DSPs won't increase the quality of the audio, they simply make processing faster. DSPs have nothing to do with the forensics or audio aspect at all. Simply speed of processing.
Dialogic was a leading brand 20 years or so ago. They provided DSPs boards to split T1s to analog lines... most of them used AT&T DSPs... the DSPs were mostly used for text to speech, automatic speech recognition, etc...
I worked with Dialogic boards in Brazil back in the 90's...
BTW, if you want to attack the forensics angle, you have to go towards signal quality, which would start with an analysis of uLaw, the premier digital compression algo for telecom calls these days...
Much higher than it was 5 years ago...
look what up? I actually know what I'm talking about here...
uLaw
Now explain why that matters with audio forensics...
Yes, I thought that's what I was doing. Attacking the signal quality, starting with the end to end equipment.
Hint: Audio forensics isn't based on frequency...
No Sherlock. Not based, but it still makes a difference, already downgrading the nominal 90%.
How about attempting to put the pieces together, instead of attacking individual points, because the work together. Not singularly.
The frequency range is more than capable to distinguish different voices, timbres, accents, etc.
It's the signal/noise ratio... THAT is the quality that matters for audio forensics... and that's individual to every call...
You're welcome, BTW. You could simply stated you didn't know and wanted to know... but you always have to take the long road and look dumb in the process...
This is my last posting on the topic with you because I am getting pissed that you assume so many things wrong about my reasoning.
The frequency range is more than capable to distinguish different voices, timbres, accents, etc.
It's the signal/noise ratio... THAT is the quality that matters for audio forensics... and that's individual to every call...
you.
Yes, quality matters. Yes, accents, timber, etc. matters.
Quality was ty. Signal to noise was ty, and cannot be completely accounted for.
Accents change... like say a singers often have no accent, but do while talking.
Do you scream with an accent?
Again, too many variables to account for to expect 90%.
lol timber
It's all that matters. That's how a voice fingerprint is created. Obviously, I wouldn't expect you to know that, after the garbage you keep posting.
But it can. Specifically Owens said the 911 call happened to be of great quality. Tell me, since you know all these things, what was the db signal to noise ratio on the call?
That's why not a single factor is used for voice fingerprinting. But you didn't know this, and now you do. You're welcome
I do, actually
I'm not a singer, and my accent is oblivious to me. But it's irrelevant, seeing it's not the only parameter on voice fingerprinting. In my case, it would be an important factor.
I can see how it might look like there's too many variables for an ignoramus of the science like you. Voice research, including forensics, is a 25+ years old field of study. One you apparently have no idea of.
Timbre...
So what.
ElNonothing isn't listening to my argument, and assuming things I have accounted for. Not that I expect you to do any different.
Timbre matters, right...
Now that has to do with harmonics in the voice that exceed 4kz, which is the maximum you can expect to find on these recordings for comparison. Less bandwith reduced the accuracy a little. Each factor reduces the accuracy.
With this expert witness claiming he would expect 90% for a match under these multiple uncontrolled variables with no equipment baseline, that's where I call bull and think he just wants his 15 minutes of fame.
Your argument doesn't hold... if any of the voices were above the frequency limit, they wouldn't be part of the recording...
How dumb are you?
I've done a little reading of various links having this information. It seems the real finding is the 48% is "not enough to make a positive match." The experts never said it means it isn't Zimmerman. That is what the media is making up. Even in the original link, Owen cautions that he didn't have a sample of martin's voice.
"you can say with reasonable scientific certainty that it's not Zimmerman."
Another analyst came to a similar conclusion using different technology.
You're talking to Wild Cobra
Oh ! You hear the ElNono, he's done a little reading. Better throw anything those experts have said out the window!
I actually stayed mostly away from this thread because I thought this story was more the "story of the week" than anything... I don't think anything terribly exceptional happened with this case. I already stated I disagree with the standards set by the state law...
But outright dismissing actual forensic evidence because "too many variables" when you don't know what you're talking about is really a disservice to the rest of the posters here.
I'm just dismissing unsolicited and unauthorized forensic experts that went straight to the media. My "15 minutes of fame" statement still stands. If the police and FBI corroborate this outsiders opinion it obviously will carry more weight.
Couldn't possibly be that someone is taking steps on their own because the Polizei and the DA aren't doing ANYTHING?
And you know this how?
You can't be serious. Really? You REALLY think if they had evidence to charge him they wouldn't?
Stand your ground. How can they investigate something that isn't a case because of the law?And you know this how?
You can't be serious. Really? You REALLY think if they had evidence to charge him they wouldn't?
And no, the police have NEVER sat on evidence before.
The whole ing world is second guessing them, they are one inflammatory speech away from race riots down there and you think thy are hiding evidence?
If they have incriminating evidence they will present it to the grand jury. The prosecution would be stupid to publish it now and give Zimmerman's attorney advance warning to prepare.
There are currently 5 users browsing this thread. (0 members and 5 guests)
Posting Permissions