need help?

I’m about to embark on a major project to convert many of my old (very old!) audio cassettes to MP3 and plan to edit them in Audacity (in Windows 10).

As most of the recordings will be of people talking I was wondering if there were any recommended settings in Audacity to clean up the sound (remove tape hiss, etc.) and generally spruce up the audio so it sounds the best it can be, given the original source.

I’m an Audacity newbie so forgive me if I haven’t discovered any ways to do this myself. Any help would be appreciated.

clean up the sound

That’s a danger flag.

Most of the tools (noise reduction, filters, etc) create “good sounding distortion.” They’re not magic.

Noise Reduction removes part of the sound most of which most of the time is tape hiss, but some of the show goes too. For example, audiobook publishers do not allow Noise Reduction because of the voice distortion it can cause.

You can go really far by matching the player to the tapes. If they were recorded Dolby B then they should be played on a machine that knows what Dolby B is and how to manage it. The tape type should match, too. CrO2, Metal, Plain Oxide.


first, are you getting a good digital recording from the cassette? (Does the digital recording sound as good as the original cassette?)

Every recording is different so there are no automatic presets.

For the hiss you can try [u]Noise Reduction[/u]. But, it’s something you have to experiment with… If the noise is bad, noise reduction can damage the sound and “the cure can be worse then the disease”.

Then you can experiment with [u]Equalization[/u]. For experimenting, the Graphic EQ mode is easier than the Draw Curves mode. If you’re not familiar with graphic equalizers, the bass/low frequencies are on the left and the higher frequencies are on the right. Boosting the higher frequencies (maybe 5-10kHz) can help with intelligibility by bring-out the “T” and “S” sounds. (But of course, boosting the highs will also boost hiss.) You can generally cut the frequencies below ~100Hz because anything below that in a voice recording is likely just noise. The main voice frequencies are around 200-400Hz (but you need to include the higher frequency harmonics).

Whenever you boost with EQ you have to be careful of clipping (distortion). It’s a good idea to run the Amplify (or Normalize) effect before exporting in case the overall level needs to come down.

You may also need to even-out the volume. You can use the Envelope tool, Limiter, Compressor, or there some optional leveling or automatic gain control plug-ins. Compression and the more-automatic effects can be tricky to set-up and these effects also adjust the hiss (and other background noise) and these side-effects can be annoying/distracting, so again it’s something you have to experiment with to make sure the cure isn’t worse than the disease.

If you can’t get good intelligibility, consider typing-up a transcript to go-along with the recording. :wink:

…Sometimes it’s a good idea to come-back the next day with “fresh ears” to make sure you haven’t over-done the effects & processing.