How do I remove noise from a set of files?

Hello everybody, how are you doin’?

I want to remove the tape noises from a set of old audiobook recordings, that are already digitized as .mp3 files.
I used one track to experiment a bit; found a 4 second section where the narrator was silent and got a noise profile from that, used 25 - 5 - 0 for the parameters and voila that works great and I am happy with the result. :slight_smile:

Now I want to apply this to a whole battery of audiobooks and to all the files for the tracks.
The recordings were made with the same device in the same circumstances and are from the same series, so I assume the noise sample I have should work for all of them.
How do I do that?

and bonus question:

Is it possible to export the resulting files with the “same” quality. Meaning similar/same audio fidelity (disregarding what might be lost from the effect) and similar/same file size?

Thanks a lot and have a good one.

(using Debian bullseye’s audacity 2.4.2 package.)

While you keep Audacity open and running the latest Noise Profile that you obtained will remain. It will be lost when you close Audacity.

What I would suggest is the the create a small project with just your “noise” and save it. Then in future when you are working on Audacity just open this Noise Profile project and use that to sample the Noise Profile.


No you will get extra MP3 compression degradation when you then export as MP3 - to minimize that export at the highest rate of 320 kbps.


So you can create a macro with two lines:

  1. Noise Reduction
  2. Export As MP3

To do that, Tools > Macros, New (your macro name); Insert > Noise Reduction; Insert > Export As MP3; OK
Now, load up your noise file, select your noise, then Effects > Noise Reduction > Get Profile;
Delete your track, so you now have an empty project, then Tools > Macros > Files, select your files > Open

Your files will be written to the macros-output directory. Don’t forget to get your noise profile if you run this again later.

I hope this helps. :smiley:

And don’t forget Peter’s (waxcylinder’s) suggestion. Do a manual MP3 export first to setup your MP3 parameters.

to minimize that export at the highest rate of 320 kbps.

And just to be clear, that will greatly increase the digital size of the files—a violation of your requirement. You can get noise-free files with compression distortion, or perfect, noisy files. Never do production in MP3. It’s a land mine. It sets you up for explosions later.


Thanks to you both.
The piece that was missing for me was how to tell the macro what noise reduction profile to use.
But the way you are telling it it looks like if will use whatever sample was taken last while the same audacity window was open, even if the original files are closed. The interface doesn’t really expose that information. Thanks.

About the MP3 option. I realize another MP3 encoding pass will always reduce quality further no matter what. What I meant to do was export the file with the same “parameters” the original imported file had.
In Audition for example this is done if you do File → Save . This will do an export, too, but the resulting file will be of (almost) the same size. I assume the program reads the original file’s parameters and then applies that on exporting it. To my ear that results in “good enough” quality compared to the source file (which was lossy mp3 too).
Is there a similar thing in Audacity?

Is there a similar thing in Audacity?

Audacity famously has no knowledge of the original file. There is no Get INFO. It exports using last used parameters. If I export MP3 right now, it will use 192 quality which is the value I used last week.