I'm New and Need Help

filtering out machinery hum

Hi There! I made a recording of nature sounds and I need to filter out the low hum of a distant air conditioner. How does one do this? I know it’s probably easy, but I am clueless in Davis. Thanks for any help. :)

you prob need to do this in a wave editor rather than n. I do this using Goldwave, but other wave editors should have similar functions. The operation goes something like:

select a short section of the “noise” on the wave file. Go to filter => noise reduction and there is a selection in Goldwave to select what noise you want to reduce. With the whole file selected I choose to remove noise using “clipboard”. This processes the file and removes all noise that is the same as the initial selection. Works very well.



What I often do is use Audacity.

It has a ‘Noise Removal’ tool. Basically, you find a part of the sample which is just noise, and select it. Then do Effect->NoiseRemoval, and hit the ‘Get Noise Profile’.

The app now knows what to look for as noise. Then select the whole sample (ctrl+a will do it), and do Effect->Noise Removal again, this time you’ll do step 2. Set the slider to the left as far as it will go (I have never had to use it any more than the minimum setting), and hit remove noise. If the signal to noise ratio isn’t too good, then you might get some slightly phasey artifacts, but they shouldn’t be too bad.

Once you are happy, use file->export with which ever type of file you are after (you can set the WAV type using File->Preferences, File Formats tab). Audacity has a problem exposting to the same filename you might already have open, so use a different name (I generally put a suffix of _proc {for processed}), then you always have the original!

I have to do this all the time with my vocals, since I have a laptop with no outboard soundcard - I have to use the mic straight into the onboard chip. Provided I sing loud enough, the results of this method are useable.

This is the simplest (and cheapest - it’s free!) way I know of doing it. I’ll be happy to be proved wrong.


Thanks DSP and Maaszy!! I’ll check them out. monksbread