Removing hissing noise from audio file from SOV

I’m putting together an audio trailer for a Shot On Video movie from the 80’s. I’ve extracted the audio from the movie into a .wav file, and am using that to create my trailer. The audio, being from an SOV, isn’t that great to begin with, and there’s a bit of what you’d call tape “hiss” that appears in and out of the film. I’m thinking using the noise removal option will work, provided I first start with a section that is just that hiss and no other sounds of the film to confuse it (which I’ve found), and then I can use that profile to apply to the rest of the file.

The question is, is it better to make the audio file I’m working with a greater or lesser amplification? Ultimately, I think it will need to be amplified a bit when I use it to create the trailer, but that seems like the hiss might still be there once I amplify it, even if I use the noise removal option during the audio file’s original level. Is it best to try the noise removal out with the entire file, or break it down into smaller files? Smaller files would work if I tacked on the section of hiss to each file so that Audacity can create a profile, but I don’t know if that’s more of a headache or not.

Just checking to see if anyone has had extensive experience dealing with this kind of audio hiss on their files. Thanks!

Just checking to see if anyone has had extensive experience dealing with this kind of audio hiss on their files.

Well, yes. You can’t actually take hiss out of a show.

Noise Removal works by being given a “taste” of the noise by itself — the profile step — and then it generates notches and filters appropriate to that noise. Run the tool again and it tries to remove that particular noise from the show. Hiss is made up of all tones, so Noise Removal tries to remove the whole show.

The noise removal control panel has sliders and settings that try to steer what the tool does. In particular “Frequency Smoothing” tries to work around voices and performances and process everything else, the idea is you won’t mind slightly noisy voices and music if the rest of the show is pretty quiet. If you have severe enough hiss, human speech will turn into snake voices over a quiet show. If you let Noise Removal have its way with no restrictions, voices will turn into gargling martians.

You can try it and see what happens.

I would Effect > Amplify to a goal of -3dB and then go about applying the Noise Removal tool. Do Not go for studio quiet between words. Stick with a gentle removal of 9dB or 12dB. Having a little noise left over is far better than producing brittle, tinkling speech.

Unless you’re sure all of the shows have exactly the same noise, you should be getting a fresh Profile each time. You can obviously experiment with that, too. The Profile step is insanely important. If you include the slightest voice or music in the profile step, that voice or music will get distorted.


Sounds good, thanks for your response!