If is is NOT Zimmerman, what would be the chances of actually getting a 48% match?
The software compared that audio to Zimmerman's voice. 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.
Ok, if a actual match is better than 90% could we not reasonably expect an elimination by the inverse, that a "no match" example would be 10% or less?
And if an intentional "no match" 10% example was the inverse of a match 90% example how the heck do you get to 48%?
Given the other factors such as background noise, volume, cell signal altering the voice spectrogram, brevity of the part being identified (who screamed?) etc, I would think that would degrade the "no match" 10% example even further to near zero.
Not having an inverse elimination because one DID get a 48% match just might be strong evidence it IS Zimmerman.
One would need a unbiased, disinterested third party analysis and opinion to expect a reasonable analysis.
What we are being presented is a former employee of the >>>New York<<< Public Library, a hand pick of the >>>Chicago Tribune<<< and paid by >>>Martin's<<< family comparing a scant few seconds of high pitched scream to normal volume speech samples.
Nope, no potential for bias there, right?
.
I don't know if a reasonable analysis could be done without recording of Martin screaming assuming there are even recordings of him speaking.
I shudder to think it but if this case is tried it's possible they will have Zimmerman reenact screaming, record it from the same distance of the 911 call it was in the distant background on, and compare.
These experts are really damaging their own credibility making such declarative statements - assuming the press quoted them accurately, which is always in doubt. They could well have said "they can't say with certainty it is Zimmerman," which is a great deal different than saying "it definitely isn't him."