How to improve old tape recorder audio?

I am using Audacity 3.1.3. I have some MP3s of an old tape recording of my Grandpa telling his life story back in the mid-90s. In only 15 minutes and with only novice skills I was able to dramatically improve the audio quality using Audacity. But the quality still isn’t great, and I am appealing to the Audacity community for a bit of advice on how I can improve it some more.

Here is what I did:

  • Ran Noise Reduction from a sample
  • Amplification: 32
  • Noise reduction (again)
  • Bass & Treble: Bass: 0, Treble: 13 (leave volume alone)
  • Noise reduction (again)

I am a novice noob with Audacity, so I don’t know what I don’t know, and I’m not sure what specific question to ask other than “how can I make this audio clip a little more clear and crisp?”

Any help would be greatly appreciated. I have attached 2 clips: (1) A very short clip of the original audio, straight from tape recorder; (2) that same clip run through the Audacity steps shown above. If you think my steps were wrong I can easily try a different approach. Or if I should begin with my cleaned sample and do some additional steps, that is fine too.

Thanks in advance.

Best to perform Noise Reduction in one pass, rather than multiple doses, as each pass adds processing artefacts.

Stereo hiss is more distracting than mono,
so I would split the stereo tracks and delete one to make it mono …
https://manual.audacityteam.org/man/splitting_and_joining_stereo_tracks.html

Audacity’s equalizer (or an equalizer plugin) gives finer control than bass&treble,
e.g. cut-out everything above ~6000Hz, as it’s only noise …
Grandpa_B-A.png

Trebor this seems very promising, but my Graphic EQ looks totally different from yours and I’m not sure what to do:

GraphicEQ2.png
I tried pulling all the sliders above 6K down to the bottom and previewed it, but that seemed to have no effect:

GraphicEQ.png
And should I do what you’re suggesting in place of my Bass/Treble mod, and do my other steps the same as before?

Your original signal is very quiet, you’ll need to amplify it before you can hear it properly.
Before applying any effects, (e.g. equalizer), increase the level of the track, to, say, -20dB RMS.


https manual.audacityteam.org/man/loudness_normalization.html
Then you will be better able to hear what an effect is doing.

Apologies, but Loudness Normalization doesn’t exist in my Audacity. Normalize doesn’t seem to be the same thing, and I drew arrows where I would expect to find Loudness Normalization. Why don’t I have that? How do I get it?


loudness.png

Then just use amplify until the waveform is approximately the same size is this …

The loudest peak in the pale blue areas is approximately 0.5 on the linear scale (on the far left).

Trebor,

Any chance you could summarize the steps? Two of the steps you gave cannot be done in latest Audacity so now I’m puzzling over how to go about this. Taking a stab, I think you’re saying to do this, but I’m probably off:

1- Split the stereo tracks and delete one to make it mono
2- Amplify
3- Remove noise

What am I missing? What else should I do? Is my sequence correct?

Thanks very much for your assistance.

Attenuate those frequencies above 6000Hz, and below 100Hz either with an equalizer,
or high-pass (100Hz) and low-pass (6000z) filters.

IMO if there is dead silence in your processed audio you’ve used too much Noise Reduction,
(you should use as little Noise Reduction as possible, as it will damage the speech to some degree).