Y'all heard of any new controversial murders? I'm tired of this one already.
Nasal cavity surgery? Really? Strawman much?
Y'all heard of any new controversial murders? I'm tired of this one already.
Unlike you, I'm not invested nor interested in the case. And I know the science behind the technology. I studied it, I use it, and I had to write software that uses it. The part where I wonder if there's anything wrong with the science or the technology is long in the past. I know what the caveats are for using any particular method.
Which brings me to the question I asked specifically to test your knowledge in the area. Are you going to answer that?
I had to ask.
I forget the right terminology, but everything the software relies on is insufficient between the two samples it had to be tested with. I don't care if ElNono acknowledges it or not. The chest cavity is part of what makes our voices unique from one another. Sitting or laying down makes subtle difference. Being pinned down, even more.
From the start of the "owens" thing, I said that 90% is too high to expect under the cir stances and samples. Unless someone can prove otherwise, I will maintain that.
Not that it matters (or does it,) but I hear Owens is being paid for this as a witness.
I see. So since this software has been used in court apparently and the mother/father said it was their son that would probably be enough in court right?
If you can't grok the terminology, then allow me to completely dismiss any conclusions you draw from your faulty understanding of accoustic modeling.
This is like talking to a caveman. People might sound different after such surgeries, but it's very unlikely that they'll change the way they talk.
You still think this kind of pattern analysis is only done on pitch, etc. You're dead wrong.
I'll acknowledge you don't know what you're talking about.
I'm sure the Police will use similar technology to determine who says what on that call. Unlike the bull you've been posting, there's nothing wrong with the underlying technology.
Are you going to answer the question?
"The pitch"
So...
Yelling help loudly and under high stress, laying down, probably pinned down should yield a 90%+ result when compared to normal talking and sitting down or standing? Then on two different voice delivery systems as well?
Really now.
Is this portion directed at me?
Thanks for the info by the way.
Absolutely. *Especially* when people are under severe stress. Because it's even more unlikely that a person will try to fake the way they talk.
No, that was for Cobra. Completely missed it was your post. Sorry.
I figured.No problem.
BTW, the correct answer is that you would never not use HMM.
Even with a neural network, which is less susceptible to acquire statistical noise from the input and thus better at detecting discrete phonemes, you can't do temporal dependencies with it. So it's good to be used as a pre-process to HMM, but you still have to run HMM over the entire sample.
I for one would love to hear WC discuss stochastics.
Stochastics....That's my favorite cheap Vodka.
I wouldn't.
I really like Stolichnaya.
I can't put down o's from Austin. That's some smooth Vodka.
Good choice if i may say so myself.
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions