Audio analysis software needed

[Nelson37]

I am looking for a program that can examine an audio clip, and determine that is NOT composed of repeating or looped segments, and show that a 30-minute to 2-hour recording actually occurred in real-time.

Does not need to be FBI quality, at least not yet, but some documentation or expert statements that it can indeed perform this function would be helpful.

And/or any companies which would perform such an analysis, hopefully fairly cheap?

[Cornucopia]

Nelson37, I don't know of any Audio-specific apps that do that, but one thing you might try would be to export the audio file as a headerless, RAW LPCM stream (1 channel only each). Then you will have a continuous stream of bytes, which you could use more general-purpose pattern matching applications on (now it's just a Binary file). Like a "GREP" utility. Find "dupes" etc. You'll have to tell it somehow to parse on sample boundaries though (16bit, 24bit).
Do you already know of a questionable section that you think may be repeated, or are you trying to search for any repetitions?

Scott

edit: or maybe run a "decode" through a steganography app (hidden/masked/encrypted piggybacking)

[Nelson37]

I wish to demonstrate that a lengthy clip does NOT consist of shorter clips that repeat. I will be creating the clip. It may be used in a somewhat formal environment.

I thought of making the recording with a movie playing in the backround, but then it would be easy to just take a 5-minute clip and loop it to overlay. The hour or two time interval required to be covered is for my benefit, and it is known that I could make such an overlay or repeating clip.

I need to demonstrate to a non-technical, very small audience that something they will listen to a very short sample of goes on for hours, as an actual event in real-time and not faked by "special effects".

Sort of like the Seti@Home in reverse, I want to show there is NO repeating pattern.

Thank you for your interest, any and all ideas appreciated. I have one week from tomorrow.

[BJ_M]

if you put the whole thing through a FFT -- you would see certain peaks repeating over time .. though by pitch shifting and other methods -- you could disguise this ....

ask the producers at CSI , they can do anything