audio subtract process? (to create multi track master from stereo sources) -Sound Forge - PC

Subject:audio subtract process? (to create multi track master from stereo sources) Posted by: madsci Date:7/28/2003 11:19:13 AM I'm looking to take a sound file and take a second file, and isolate what the two files have in common, and what is different about them. For instance if you have two WAV files that are different mixes of the same recording. Say one has a guitar solo in it, and another has a horn solo instead, but in all other aspects they are the same recording, this program would compare the two files, and save a new WAV file with the guitar solo only, and another with the horn solo only, and another with the rhythm tracks (what the two files had in common). The result would be somewhat a multitrack master. In my case I have a bunch of recordings that were not recorded multitrack, but instead were generationally overdubbed - playing along with a 2-track tape, and recording onto a 2nd tape. Then taking that tape, and playing along with it on another instrument, onto a 3rd tape, etc. I would like to take these different tapes and process them to isolate the overdubs, and thus have a multitrack master. Does Sound Forge have any plugins that do this? Or Active-X plugins by a 3rd party? If not, Sonic Foundry please make one. This would be very useful for restoring old recordings (esp rare performances where multiple tapes exist, but none are particularly great, but put together, would yield better quality.) It seems like it should be possible, if the two sound recordings have common data, then you just compare. Of course, two different mixdowns (esp from an analog source) would need to be edited to start/end at the exact same second (millisecond?) and be exactly the same length (maybe one tape source was slower, so one file would have to be stretched/skewed to be the right speed -anyone know how to do this in SF?), and the function would probably need some kind of fuzzy logic to analyze the sounds, because if they aren't exactly digitally identical, the program will need to compare them with some intelligence.
Subject:RE: audio subtract process? (to create multi track master from stereo sources) Reply by: Geoff_Wood Date:7/28/2003 11:21:02 PM You could do this relatively easily if the audio files are sample-accurate synched - with careful level matching, phase-inversion, and mixing. If not exactly matchable, results would be trial-and-error only. geoff.
Subject:RE: audio subtract process? (to create multi track master from stereo sources) Reply by: madsci Date:7/29/2003 9:22:09 AM Can you give an example? I'm not familiar with level-matching & phase inversion... Thanks

I'm looking to take a sound file and take a second file, and isolate what the two files
have in common, and what is different about them.

For instance if you have two WAV files that are different mixes of the same recording. Say one has a guitar solo in it, and another has a horn solo instead, but in all other aspects they are the same recording, this program would compare the two files, and save a new WAV file with the guitar solo only, and another with the horn solo only, and another with the rhythm tracks (what the two files had in common). The result would be somewhat a multitrack master.

In my case I have a bunch of recordings that were not recorded multitrack, but instead were generationally overdubbed - playing along with a 2-track tape, and recording onto a 2nd tape. Then taking that tape, and playing along with it on another instrument, onto a 3rd tape, etc. I would like to take these different tapes and process them to isolate the overdubs, and thus have a multitrack master.

Does Sound Forge have any plugins that do this? Or Active-X plugins by a 3rd party?

If not, Sonic Foundry please make one. This would be very useful for restoring old recordings (esp rare performances where multiple tapes exist, but none are particularly great, but put together, would yield better quality.)

It seems like it should be possible, if the two sound recordings have common data, then you just compare. Of course, two different mixdowns (esp from an analog source) would need to be edited to start/end at the exact same second (millisecond?) and be exactly the same length (maybe one tape source was slower, so one file would have to be stretched/skewed to be the right speed -anyone know how to do this in SF?), and the function would probably need some kind of fuzzy logic to analyze the sounds, because if they aren't exactly digitally identical, the program will need to compare them with some intelligence.