Removing hiss from an old TV recording

I have an recording of an old TV program that was recorded by placing a cassette player in front of the TV. Unfortunately, there is a constant hiss. I’ve attached the first minute of the original BEFORE any adjustments.

I did try the noise removal and the preview I listened to sounded like it removed most of the constant hiss, but it now sounds like there is choppy occasional static in the recording. I initially tried a noise reduction setting of Audacity 2.0.2 at 15db and that seemed ok but perhaps tweaking the other settings like sensitivity, Frequency smoothing or attack/delay time would remove some of the choppy static sound? What would you do for this recording?

You can’t actually take hiss out of a recording.The best the tools can do is take hiss out of most of the show and try to stop reduction during human speech and music. If the show is bad enough, the tool can’t tell when to kick in and out.

If you don’t do that, then the hiss removal will succeed, but it will take voices with it. It will make voices gargly, bubbly and honky. Try reducing the “smoothing” setting to zero. That should turn off the switching and masking.

We don’t do forensics. We can’t bring back a recording from near-dead. That segment in particular has a significant noise change about a third of the way through. Technically, you would need to get a different noise reduction profile before and after. It’s two jobs. One profile isn’t going to do it.


Given that it’s a very old cassette tape, you could try the Effect > Low Pass Filter set to 5000 or 7000. 24dB. Experiment.

One more. Effect > Notch Filter at 180. That should get rid of that hum.



Should I try the Low Pass Filter and notch filter before or after trying to remove the hiss?

There’s nothing but noise above 4000 Hz, so “Lowpass: 4000Hz 48 dB per octave” will remove a lot of the high end hiss.

There’s also a peak at 60 Hz, so run it twice, once at 60 Hz and once at 180 Hz.

Also, use a high pass filter, 12 dB per octave at around 100 Hz to get rid of much of the rumble.

Finally you could amplify it a bit (Effect menu > Amplify) and convert it to mono (Tracks > Stereo Track to Mono)

Thanks Steve.

Do those steps in the order listed?

Then it sound like you are not recommending using the noise removal feature, correct?

The order is not particularly important in this case.

I did it:

  1. Stereo to mono
  2. Low pass filter
  3. High pass filter
  4. Notch filter 180 Hz
    5)Notch filter 60 Hz.
    which is basically the order of “what is most bad” about the recording.

I don’t think that the Noise Removal tool will make much improvement without also damaging the sound that you want to keep. There is just too much noise for that effect to cope with, though feel free to try it :slight_smile:

Noise Removal has really good publicity (Removes!! Noise!!) when in fact it’s mostly Hollywood. It never removes most common noises and we talked about renaming it “Noise Reduction.” Sometimes it’s debatable how well it does that. By the time you realize you need it, it’s already too late. As above, we can’t do forensics; we can’t turn a bushel basket of barely recognizable garbage into a clear, clean performance and we can’t even get remotely close.

Noise Removal is a festival of trade-offs and work-arounds and it’s more difficult to use than you think. In order to clean voice performances, the tool is adjusted so it reduces noise everywhere but voices. In extreme cases, this gives you a velvety clean background with hissy voices in it. You can’t reduce noises, particularly hiss during a voice without destroying the quality of the voice. And so on.

The colorful advert on the side of the box is not an indication of contents quality.

I know of one (1) person that got good results with Noise Reduction without prompting or hand-holding. They did a shoot in a room with a gentle air-conditioner rumble and you could barely hear it during the show, but they applied the tool anyway and it produced a perfectly presentable, nose-free show. It was years ago and I remember it clearly because it was so successful. And alone.

In your case, you’re far better slicing off portions of the performance (Low Pass, etc.) that don’t carry valuable work and in fact just add tape hiss or other noises.


Thanks. It does sound better. In the future if I have other voice only recordings with just the hiss I should use the low pass and high pass filters, correct?