1

Context:

I have a huge amount of audio files (tape recordings that were digitalized into mp3) that I've been asked to clean up for someone. It's only voice (conferences).

Why:

I guess it would be a lot of work even for a professional and there is no budget for one, but I'm willing to learn and do my best.

Starting point:

The sound is overall muffled, as if through an aquarium, with coughing, loud clicks, and the voice sometimes disappearing, and the speaker goes from almost whispering to almost shouting.

What I've done so far:

  • converted into .wav files to use in Audacity.
  • Noise reduction
  • Normalized (should have done it later?)
  • boost in the 2000 Hz and above and cut below the 100 Hz.
  • Limiter (which didn't make much of a difference on the very loud parts)
  • Envelope tool on very loud parts (works great but it requires again to do it almost sound by sound)

Result:

  • the sound is a bit more clear (as far as my very untrained ears can tell) but there is still as if there was a blanket over the voice. I actually did that operation twice (though with less boost in the higher Hz the second time) so it was even better, but I still don't find the voice clear and close, and now I have a lot of sibilance. If I apply a de-sibilance thingy (in EQ) I'm worried it will dull the whole thing even more so I'm busy doing it whenever it pops up...

Also if I turn down the volume the sibilance goes away but that doesn't help the rest of the sound.

And:

I'm willing to learn and try and all, but to be honest when playing with EQ I had to apply very big differences in dB before I could hear it.

SO:

Is there any sort of advice you could give me to make the voice come up better?

I suppose that having the audio without the wind, coughing and munching is already a great improvement but I would really like to bring the voice to its best...

oh, and I've tried fiddling around in the low mids but with no satisfying result (that I could hear).

So my question is:

  • any advice? even just a direction you could point me ?
  • am I condemned to adjusting bits by bits of every audio?

Thanks a lot for your advice and my apologies if I misused the jargon, I only found out three days ago about frequencies in the voice...

P.S.: I had no idea it would be so interesting!!

1

Wav files are better quality but unfortunately converting a compressed .mp3 to a .wav doesn't improve quality.

I would interpret the "blanket" over the voice to sound like it's muffled which means the high-mid to highs may be lost but you can try to balance it out by lowering the low as well as raising the higher frequencies.

Also using a limiter or compressor is a good way to raise the whisper and suppress the yelling, but you have to make sure that the threshold is in an appropriate place to take effect. Then raising the overall volume so the quieter parts are loud enough to understand.

Honestly, there is only so much that you can do to enhance a bad recording before you start going backwards. It's kind of a "it's as good as it's gonna get" thing.

  • I can't upvote as I'm a newcomer but thanks. For those interested : a multiband compressor seems to make a huge difference ! – timotheap Apr 11 '18 at 21:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.