Quote:
Originally Posted by Jalfrezi Right you can try this, it may work, it may not. it worked for me when i wanted to remove a high hat from a recorded track. not totally but it was close enough.
copy the track onto a new channel. filter all the sound out except the frequency range where the fuzz sound is. so there should be no bass on this just some fuzzy mess. then, and this is where the magic happens. get a gain plugin with a phase reversal button on it (i'm talking in logic terms, i hope your DAW has something similar).
This should remove or at least make a significant change to the fuzz. then from there just fiddle with the eq and volume until its acceptable. by the way the tracks must be at exactly the same volume for this work and must have exactly the same processing etc.
Hope it works for you. |
Thats a rather ingenious approach to the problem, and one that got me thinking I'd missed a trick for years!
However, it isn't the most elegant solution to this particular problem, I'm only posting this because it got me thinking, its certainly a valid solution!
The trouble with the method is that phasing in audio is quite complex, and doesn't behave in the way you may expect. Isolating a band and phase reversing it in such a way does not just cut the chosen band, it also boosts in various other areas unless you use a powerful fft style filter with an infinitely steep q (and a lot of time setting it to perfectly mirror the offending sound, a near enough impossible task!). Often while the center of the offending band is cut when phase reversed the frequencies on the LOWER side end up getting a small boost and the HIGHER side a bigger one, so although you are cutting the offending frequencies you are also altering others! Depending on the slope of the eq on the phase reverse track you can get much more radical effects, such as cuts either side of the focus frequency and a little boost in the middle. Try using a freeware spectrum analyser like voxengo span and watch the results of moving the phase reverse track eq around while trying to cancel out an area of pink noise, for example. (if you are as much of a nerd as me and want to see some screengrabs of the results let me know

)
So in the case of cutting unwanted signal from tracks a normal subtractive eq will be a cleaner way of going about things, and is capable of getting exactly the same sound as the phase reverse method, though that means using another band or so to put boosts in the areas the phase reverse method will accentuate.
Its a good suggestion though, as not only is it a great tool in getting to know phase a bit better but the resulting eq curve from this method may actually be more musical than a direct cut. When cutting a specific frequency from a track it can sometimes help to boost a frequency nearby to thicken out the sound again (unless thats not the desired effect). EQing this way can change a sound subtly without it sounding thin/muddy/processed while still achieving the desired result of shifting the emphasis points in a sound. The phase reverse method does this automatically so may sound very good.
Aside from EQing, a similar technique would be to record a bit of just the noise you want to remove (for example amp hiss) on a separate track throughout the songs duration and phase reverse that. This is a more old school method and will not remove the offending noise completely as its not an identical sound (and sometimes won't do much at all), but with a bit of fiddling about it CAN reduce it a bit, and will have a lot less of an impact on the 'desired' sound content. No use when trying to remove a hi hat though, or course!
To the op, the noise is most likely the cheap computer preamp, a little interface will likely be a lot cleaner and shouldn't cost all that much! Though getting used to the methods mentioned in this thread will teach you a lot about audio!