If it is truly white noise, you will not be able to remove it, because by definition white noise is random and broad-spectrum.

You may be able to reduce some objectionable areas by selective high freq filtering or a noise filter as the pp suggests, but at the expense of some of your program material, as well. Sorry, it's not going to go away.

by definition white noise is random and broad-spectrum

There seems to be much disagreement over what white noise is. My understanding of it has always been equal power at all frequencies. Or, in our case, equal power at all audible frequencies.
Every now and then, I read a definition that combines "random" with that concept. Doesn't make sense to me -- seems like white noise is anything but random.
musicvid, would you please explain it?

It is random in that there is no pattern that repeats, such as you have in an air conditioner noise. It's also random in that any given small section of it may have an uneven distribution of frequencies that has no relation to sections before or after it. The equal power at all frequencies is evident over larger periods of time.

Imagine rolling a die. Each roll is random since you never know what number will come up next. However, roll it 6000 times and you're going to have pretty close to 1000 of each number.

So yes, white noise is random and this is what prevents noise reduction plugins from removing it. Noise reduction attempts to predict the pattern of noise both in repeating patterns and frequency distribution. White noise doesn't have either of these patterns so noise reduction cannot 'follow' it from moment to moment and distinguish it from other sound.

Thanks, chienworks.
I experimented with audio files that are presented as "white noise." They supposedly are the types of sounds which are used to soothe babies, mask unpleasant sounds made by office buildings, etc. They're the sounds of water running from a faucet, hair dryer, vacuum cleaner. Although at any given moment, the freq spectrum is not flat, over time they're pretty close. And NR software does a great job of attenuating them, with no audible artifacts.
Do I still fail to understand?

Thanks for the information, I downloaded the demo of both tools and decided to go with iZotope RX. It worked great for removing the background noise!

Joe

"There seems to be much disagreement over what white noise is. My understanding of it has always been equal power at all frequencies."

You are thinking of pink noise...what spectrum analyzers output to calibrate a sound system to a room.

**They supposedly are the types of sounds which are used to soothe babies, mask unpleasant sounds made by office buildings, etc. They're the sounds of water running from a faucet, hair dryer, vacuum cleaner.**

Wow, didn't mean to get this deep on this, but since you asked:

* None of the things you mentioned are white noise, which chien described quite accurately, as he always does.

* SOME of the things you mentioned either qualify as "pink noise" (which is contoured to the human hearing spectrum), or "patterned noise," either of which are identifiable by countours or repeating patterns quantified by interval or frequency, like air conditioners. Although I could, I will not go into the advanced trigonometric representations of these, but i will tell you that they are readily available if you are truly that interested in them.

* The fact that true white noise can not be subjectively reduced without a detrimental effect on the program material; or that filtering effects ARE somewhat more effective (without quite as much impact on the source material) in cases of identifiable contours and / or frequency patterns is not disputed here, or anywhere else, as far as I am aware.

Hope this addresses your questions.

EDIT: If you are interested in some really good waveform generators, as well as excellent white / pink noise generators, and a bunch of other useful audio utilities, check out the excellent stuff at http://www.nch.com.au/tonegen/index.html
Used their stuff for a decade or more, and have never been disappointed.
musicvid

I will not go into the advanced trigonometric representations of these,

Heh. I've looked at that. The definition of white noise is quite specific. So in reality, much of what is called white noise is not white noise.

Heck, if the definition were as loose as people seem to think, a symphony would qualify as white noise.

Thanks.