You can try some [u]Noise Reduction[/u] to reduce tape hiss. But if the noise is bad, “the cure can be worse than the disease” so it’s something you have to try to see if it makes the overall sound better or worse.
Usually the best tool is the [u]Graphic Equalizer[/u] and it’s a matter of experimenting with the preview to bring-out the voice frequencies.
The sliders on the left are the low frequencies (bass) and the highs (treble) is on the right.
You can start by minimizing the frequencies below around 100Hz and above 10kHz. Anything in those extremes is beyond the voice range so anything there is just noise. The “T” and “S” sounds are in the remaining upper frequencies and boosting these can bring-out intelligibility. but that will also boost the tape hiss. And otherwise it just requires playing with the sliders and experimenting.
The [u]Limiter[/u] can make some “grit” and “intensity” in the voice and with make-up gain it can make it louder. But limiting is dynamic range compression which brings the loud and quiet parts closer together. That reduces the signal-to-noise ratio, making the noise more noticeable.
Oh… There is also a [u]Noise Gate[/u]. A noise gate is a kind of crude noise reduction that reduces or completely kills quiet sounds. It’s actually downward dynamic expansion, the opposite of compression and it’s working on the quiet parts instead of the loud parts. But it can sometimes be distracting or annoying if you can hear the tape hiss cutting in-and-out.
Any boosting, including boosting with EQ can push the levels into [u]clipping[/u] (distortion). If that happens you can run the Amplify effect with a negative dB value (attenuation instead of amplification) before exporting, the exported file won’t be clipped. Audacity is actually showing potential clipping but if you export the file it will be permanently clipped so you have to lower the level first.
In any case it’s a good idea to run the Amplify effect (or the “similar” Normalize effect) before exporting because the default setting will either attenuate or amplify as-necessary to “maximize” the volume.
There’s a (currently free) online AI-audio-enhance service from Adobe … https://podcast.adobe.com/enhance
It seems worth a try for cassette tape clean-up.
NB: voice only, Adobe-enhance will ruin music by turning it into speech.
It can apply noise-reduction dynamically, (Native Audacity cannot do that).
Adobe-enhance has zero controls: drag-n-drop.