FINDING AUDIOS

[pepelotas]

Hi,
Actually I'm new to Audacity and I don't really know if it´s got what I´m looking for. For me it seems a simple feature but I don´t really know. The situation is the following:
i've got 2 tracks, Track 1 and Track 2. Track 1 is long and Track 2 is a short fragment of Track 1. However, I know FOR SURE that Track 2 IS a fragment of Track 1 but what i dont know is WHERE that fragment is in Track 1. What I need is a finder, or something like that, that can look for Track 2 in Track 1, and tell me its location. I don´t know if Audacity can make this or any other program that you could recommend me, please.
Thanks for your help.

[kozikowski]

No, and for reasons that may affect other programs as well. It's not enough for the sounds to be the same. For a simple comparison program, the actual data has to be the same, too. This is where MP3 kills you and Audacity isn't far behind. MP3 scrambles the data each time you make a new file, so from a data point of view, each song is different, even if they have the same music.

Audacity exports sound files through a dithering algorithm, so again, no two songs are the same, even if they sound it.

So the program you're looking for has to "know" what music is, to find the one place where muted B-flat follows A-natural. The audio data is irrelevant. That's much harder.

We're on a newer version of Audacity and it has a comparison tool, but I'm reasonably sure it's not going to do what you want. It just tells you the volume difference between two clips.

Koz

[steve]

I know FOR SURE that Track 2 IS a fragment of Track 1 but what i dont know is WHERE that fragment is in Track 1

And you can't listen to Track 1 to find the part because... ?

[RS_RS]

Aligning in time two signals differing by a moderate amount of noise (including the noise arising from digital transformations such as re-encoding) is in principle a pretty easy statistical problem, with some connections to the much more difficult problems of aligning tree-ring sequences in dendrochronology and of stitching digital images together. Contrary to what kozikowski says, you do not need to go back to the musical structure, you can do it on any signal, musical or otherwise. That does not mean that you will find the tools to deal with it in a normal audio editing package. As steve suggests, you should be able to make an approximate match by ear, and once you have done that you can load them one above the other into Audacity and slide the fragment around until you get an accurate match.

[PGA]

As steve suggests, you should be able to make an approximate match by ear, and once you have done that you can load them one above the other into Audacity and slide the fragment around until you get an accurate match.

I would forget about the ears and go with the eyes. The human brain is especially good at "pattern recognition". Import both tracks into Audacity, select the shorter one and then use the Timeshift tool to slide the short sample past the full works. Let the eyes and the brain spot where the waveforms seem similar - and THEN do the audio check with the ears.