Attenuating Signal Noise

[DaleVO]

This may just be only a pipe dream. But I had a thought about using a track of reversed, room and signal noise/silence to attenuate any S/N in a vocal track. When I record a clip, I usually run 5+ seconds of silence at the end, after my read. My noise floor is usually around -57 to -55dB. Occasionally while recording, the fridge compressor kicks on, in an adjacent room, passing noise through the wall, which is barely audible in the recording. But it is there, nonetheless. So I gate it.

If I were to copy that trailing room/fridge/SN/silence, stretch it to match the time-length of the vocal track, reverse it, and then mix down the two; would that attenuate the underlying noise?
Although I have not experimented with this yet (I will), I want to know your thoughts or experience.

Thanks in advance,
Dale

 

[Armistice]

I think it would have a deleterious effect on what was actually recorded, but feel free to try it and report back. I don't know - I'm just surmising... others with more knowledge of sound may give you different/better advice on this.

Perhaps you should consider instead (a) turning off the fridge, and (b) using the "rumble" filter on the mic / preamp / interface / DAW to cut down on the low frequency rumbling...

I do both of these things and don't notice much general room noise in anything I record, and that usually includes softish vocals and acoustic guitars...

 

[Mo Facta]

While I'm not exactly clear on your method here, these type of tricks seldom work. Are you talking about polarity reversal?

Cheers +

 

[Bobbsy]

If the background noise was an absolute constant your idea might work. Noise cancelling headphones do exactly what you're talking about in real time.

However, the noises you're talking about are highly variable--for example your fridge cutting in and out. Unless the noise is 100% identical in both tracks, inverting the polarity won't help you a bit.

I'd be looking at doing more boring, less technical things:

-use the cardioid pattern of your mic (I'm assuming cardioid) by positioning yourself in the room so the null (back) of the mic is pointed at the noise source.

-hang soft things behind you when recording to minimise reflected noise off the wall that the mic is facing.

-if possible, make use of the inverse square law by getting nice an close to the mic. If you're presently a foot from the mic capsule, moving to six inches will make you an extra 4 times louder compared to the noise (or, put another way, will make the noise seem only a quarter as loud as it was).

-inverse square law part two: Move to the far end of the room away from the noise source. Doubling the distance from the fridge will quarter the noise picked up from it.

-finally, yeah, consider unplugging the fridge while tracking. If the door stays closed things inside will stay nicely cold.

 

[snus]

This might actually work a lot better than some of the other feedback suggests. There will be some by-products on any "wanted" audio content that shares the same frequencies, but this may not be a case of the solution creating unacceptable problems.

Give it a try. I would be interested in hearing the results.

 

[Ethan Winer]

If I were to copy that trailing room/fridge/SN/silence, stretch it to match the time-length of the vocal track, reverse it, and then mix down the two; would that attenuate the underlying noise?

As Bobbsy explained, this will not work. Aside from the fridge, room noise and preamp hiss are random. That means the noise now is different than what it was five seconds before. In order for polarity reversal cancelling to work, the two signals must be truly identical, not just have a similar spectrum. However, you could use software noise reduction to reduce the noise based on that five seconds of silence.

--Ethan

 

[mjbphotos]

Can anyone tell me what frequencies to cut so that the fish tank bubbling is filtered out of my recordings?

Oh never mind, I unplugged the tank and gave the fish to the cats!

 

[Armistice]

Can anyone tell me what frequencies to cut so that the fish tank bubbling is filtered out of my recordings?

Oh never mind, I unplugged the tank and gave the fish to the cats!

Yeah, but now they'll be expecting more fish and will start miaowing randomly...

 

[DaleVO]

Folks thanks for the discussions. I understand that this is like herding cats, and won't solve the problem that actually calls for good acoustic treatment. I was just pondering out loud and will experiment with the idea.

I do apply roll off on all of my mics that can, the 286 processor, and lightly gate, which gives me usable track. As Ethan suggested, if I have to, I have used occasional noise reduction to clean up.

I just had one of those "what if" moments, that crossed my mind, as I was watching my inverted guppies, roll over and play dead ). try 440Hz mj ), and Bobbsy- I have been known to shut the fridge down for a minute or two, when I had to get the track recorded.

Thanks guys for taking the time to suggest and diagnose. I always learn a lot from your discussions. Like I initially said- I felt I was probably just chasing a pipe dream.
Dale