Can Audacity restore my mutilated vocal recording?

Hi all,

I’d be grateful for your help please.

I recorded a voiceover in Audacity, effectively removed most of the hiss resulting from using a crappy on-board sound card and played the voiceover whilst running a Prezi on my PC, recording it using Expression Encoder Screen Capture.

I recorded in Audacity in 44.1khz at 128 bit, and set the same in Encoder, but it has utterly mutilated my sound. Have a listen! It sounds like system noise has somehow got into the sound recording. Any idea how best to restore in Audacity please? Also any suggestions as to how this might have happened?

I’m loathe to go through the whole Screen Capture process again, but will if I have to.

Thanks in advance!

I can see why you wouldn’t want to do it again. “The party of the first part, hertofore known as the legatee…”

That’s not recoverable and it has some very interesting problems. The most obvious one is normal voices don’t look like that on the timeline. It’s highly compressed and squeezed so all the up and down peaks are at the same level. If that wasn’t intentional, that means you’re overloading something somewhere. I’m hearing popping P sounds which means you’re too close to the microphone or you’re not using a blast filter.

The odd bubbling and weird digital noises came from your noise reduction. You can’t ever actually take hiss out of a show with noise reduction. The best you can do is suppress it between the words and leave it during the words hoping that nobody will notice. I noticed.

You need to not go through your sound card. Most of them are complete trash and are put there just so the ad can say they have one. It doesn’t have to actually work.

I have one of these and it works very well. It’s a speech microphone, it’s directional, and it completely divorces the microphone from the sound card.

This is a sample of what the microphone sounds like and the waveforms. You do need a relatively quiet room to do any kind of vocal work. You can’t do it in your kitchen.

Screen shot 2012-10-14 at 11.47.15 AM.png

The problem seems to be due to a combination of factors:

  1. Excessive noise in the original recording (the root of the problem)
  2. Excessive noise reduction (to try and cope with (1)
  3. Excessive dynamics compression (may have been applied as an effect to “even out” the volume, but has also increased the residual noise at the start and end of words)
  4. MP3 encoding losses (exacerbating the above problems)

You probably do not need to redo the screen capture, but you probably do need to re-record the audio.
The root of the problem is excessive noise in the original recording, so really you need to use a better microphone/pre-amp combination. An inexpensive USB desktop mic may be all that you need (such as the $20 Logitech mic suggested by Koz)

What is your microphone? You may just be using it wrong. Koz