can LEVELING help reduce steady background noise?

[HopeForTheVoiceless]

i've used all else in audacity, according to instructions, and now i'm grasping, i know.
also, there is freeware called The Levelator. has anyone used it successfully for this purpose?
thanks . . .

[steve]

I doubt that it will help. but if it is free there's no harm trying it.
Here's what we have on noise reduction: http://wiki.audacityteam.org/wiki/Noise_Removal

[kozikowski]

All the leveling, volume control, compression and processing tools can actually make noise removal worse or more difficult. The only two that actually do anything valuable are Noise Gate and Noise Removal.

http://wiki.audacityteam.org/wiki/Nyqui ... Noise_Gate

Noise Gate tries to suppress any sound lower than a set threshold. Hopefully, the voices are louder than the noise. Noise removal recognizes the noise because you told it what it was in the profile step, and then tries to subtract that from the show.

In neither case will they remove noise during words. Just between them. Any effort to directly affect spoken words will lead to robot or gargling voices.

The other standing problem is by the time most people decide they need either of these tools, it's too late.

Koz

[HopeForTheVoiceless]

Koz, thank you for responding.

so there's no way in the world the removal can be done? not through adobe audition? not through more sophisticated removal processes than that such as those found in recording studios/radio stations?

too bad the track can't simply be split between the background sound and the speaker. what the speaker's saying could help not only at-risk animals but the so many of the people serving them.

[steve]

All forms of Noise Removal need to be able to detect a definable difference between "what is noise" and "what is the sound to keep".
Some "identifiers" make separation easier than others, for example, "noise on the left, sound to keep on the right" is very easy to deal with.

Humans have incredibly sophisticated mechanisms for separating sound from noise in what they hear that takes into account context, spacial awareness, visual clues, an enormous database of sound patterns (memories) and many other factors that allow us to clearly hear someone talking in a noisy environment, hear a baby crying in another room (or building) over the sound of the TV, or someone putting the kettle on from 50 m. Computers can not yet achieve that level of sophistication.

so there's no way in the world the removal can be done?

We don't know what you are dealing with but if the noise level is greater than the sounds to keep level then the chances are not good, especially if the sounds and noises are both "broad band" (a mixture of low and high frequencies).

[HopeForTheVoiceless]

Koz, the background sound of the room air conditioner is as loud/louder than the voice, i think.
i wonder if there's anyone "out there" who, for a good cause, will have a look. the clip's about a half minute long.
i'm not sure about etiquette, regarding this. please forgive if i'm imposing. i imagine if i'm breaking forum rules you or someone else will let me know.

[steve]

Go ahead and post a sample.
See here for how to do so: http://forum.audacityteam.org/viewtopic ... 49&t=64936

[kozikowski]

You may need two posts. One should be "clean" nose. Noise with nothing else going on like talking, music or barking. The second post should be a portion that you want to try and clean -- with the interview or talking.

Koz