Analisis of voices in a particular room

Question

Hi Guys.
I'm not the best with audio editing software, and with that in mind I would like to pose two questions:
1: How does one prove that two speakers are speaking in the same room (And also discount that say one voice is not a telephone on loudspeaker function)
2: How does one "Fingerprint" a voice in order to match that voice against another recording?
I sure would like some help here because as you have probably guessed by now, I'm not having the most entertaining part of my life.
Help would be very much appreciated.

Answer 1

Okay.  I know there are people on this forum who actually have experience at forensic audio so hopefully one of them may pop in.  However, since you haven't had any replies yet, let me at least try to give you some pointers that occur to me as a person experienced with audio but not the sort of thing you want to do.
First, let me say that, despite what you see them do on NCIS, there aren't any quick, easy, automatic ways to do this.  It'll depend on you using the tools available to you and interpreting what you see.
In terms of seeing if they're in the same room, I'd be looking at the background noises, probably in spectral view.  With a bit of luck you should be able to spot a "standard" background noise (ventilation, outside traffic, just general echo) and see if it's the same on both voices.  If the track is edited, there's a very good chance you'd see changes every time an edit happens.  A good audio editor would try adding a continuous "atmosphere" track over both voices but you should still be able to see minute changes at any edit point.
For the question of a speaker phone, phone lines only reproduce sounds between (roughly, depending where you are in the world) 300Hz and 3300Hz.   If you check a voice and find frequencies outside this range  then it's a darn good chance it was NOT a phone line.  Again, you can check this in spectral view--or with Audition's frequency display.
The "fingerprint" thing would be much more difficult.  I'd be looking to find distinctive words common to both recording and compare the waveforms (zoomed in quite a lot).  However, this is going to take a lot of interpretation since different microphones, different distances from the mic and even different rooms will also be changing the way the waveform looks.
Good luck with whatever is going on.
Edited to Add:  Let me say that SteveG is on of the experienced guys I was mentioning and he posted while I was typing.  He makes the same point I did about being able to deliberately fake the room noise.  It's sort of a case of it's relatively easy to prove it HAS been edited but much harder to prove it hasn't.