Skip to Content

How to Clean Up a Distorted Voiceover

How to Clean Up a Distorted Voiceover

Welcome to this month’s Home Recording Trade Secret, my fellow six-string mavens. I originally had a couple of different ideas to share with you. Instead, I received an urgent email that required my services. So, let me give some context before we get into the nitty-gritty.

As some of you know, my home recording business is “a hobby that pays for itself.” I don’t advertise and like using my home studio for creative projects. But I still say yes to outside work.

Sidebar: I asked the news company if I could use a portion of the audio and create a Youtube video for this article to demonstrate the process used for this article. First they said yes. But when I asked for something in writing, they said no. So, I can’t provide the actual sound file and demo. I apologize for this. Nonetheless, I felt this is a good tip to share.

How Did I Get Into This Mess

In addition to being a fabulous singer, my wife also has a side hustle as a distorted voiceover artist. And it’s her that I have to thank for getting this gig—I’m not being facetious as this is a good opportunity. Here are my notes, cleaned up because only I can decipher the hieroglyphics of my scribbles in my notebook.

  • Client: Local News Station
  • Project: Podcast
  • Description: Assemble the pre-recorded parts, adjust the volumes, mix, and master.
  • Deliverables: 3 Audio Files (MP3, WAV, and 2-minute MP3 promo)
  • Assets: 4 Audio Files (Host with Interview 1 and 2, Folly, Bumper, and Music)
  • Comments: The Host audio is distorted, “Can I fix it?” She’s on another assignment and can’t re-record.
  • Delivery Date: ASAP

2 Things They Didn’t Tell Me

  1. The audio is recorded on a cell phone
  2. All of the audio parts (host plus 2 interviewees) are on one file

Problem 1: The audio is recorded with a cell phone

The host recorded the audio with her cell phone, a medium-quality Android phone. My wife is also a journalist, and she uses Zoom H6 for planned interviews. For impromptu interviews, she uses her iPhone. I’ve cleaned up iPhone audio, and it’s suitable for a last-minute recording substitute. This audio is not iPhone-quality audio.

The Voice Memos app that comes with the iPhone has two settings, Compressed and Lossless. The Compressed works for field recordings, and you can get good-enough quality audio (yes, with background noise, etc.), but the news channels feel that adds to the authenticity. The Lossless audio is much better, and I’m impressed with this technology.

Also, she recorded with her the phone pressed to her lips. Have you ever received that ear-drum-smashing voice message from a loved one or colleague?

Problem 2: All of the voices are in one file

The host assembled her and the two guests’ voices (separate interviews) in one audio file. Since she’s somewhere in the Colombian rainforest on assignment, we’re stuck with what we’ve got, and the episode needs to air ASAP.

The quality of the audio is the real issue here. The levels were an easy fix using automation. It was just more tedious than it needed to be.

How Do We Fix The Distorted Narration

This is where tools like EQ and compression can help. But, in this case, the audio is already compressed, and more compression will only make it worse. There are also noise reduction, normalizing, and limiting tools, but I rarely deal with field recordings and want to avoid paying a king’s ransom for the proper plugins that will clean up the audio with a tweak or two of a preset. So, who’s up for a challenge?

Problem 1: The Overall Clipping

When the host brought the cell phone close to her lips, her sound level pressure was too much for the internal microphone creating an unpleasant sound. The sound wave gets destroyed at the wave peaks. These peaks get cut off, producing a nasty distortion. So, we need to smooth out those sound waves.

selection-based processing

  • Solution: Use Selection-Based Processing

Selection-Based Processing allows you to add effects to a portion of the audio on the track. In my case, I need to “soften the attack” to round out the peaks that got cut off. Fortunately, in the dropdown menu, under the Channel Strip Settings, there are four categories at the bottom. Dynamics, Effects, Utilities, and Vocal Tools. Choose Dynamics and select “Soften Attack.” This took care of that problem.

Problem 2: Remaining Audio Anomalies

Some ‘audio dirt’ on the track still needs to be diminished.

remaining audio anomalies

  • Solution: Waves X-Crackle and X-Noise plugins.

The X-Crackle plugin is mainly used to clean up vinyl that has been digitally recorded. Since some of the ‘audio dirt’ reminded me of listening to a scratchy old album, I used this tool, which got the job done.

x crackle x noise plugins

Finally, I looped the background noise in the gaps between the host’s voice. I then added the Waves X-Noise to learn the noise print and set the threshold and reduction. This helped the voice stand out. I also added the music, creating my own loops to extend the track to last the 10 minutes of the podcast segment.

Problem 3: Plosives and Sibilance

This is the tedious part of the job. I’ve discovered that these are easier to deal with when working with raw, uncompressed files. In this case, the plosives and sibilance were absolutely nasty. I had to deal with these one syllable at a time.

Again, the Selection-Based Processing saved the day. There are separate presets for plosives and de-essing. This didn’t clean up every syllable. But it made the audio listenable and worked despite all of the issues. And I made the client happy.

Closing Thoughts

Nowadays there are producers of many forms of content. And they can only be experts in some mediums and not all. This leaves room for some of us, with experience, opportunities to flex our chops and make a few bucks. 

I enjoy the equal parts of creativity and technology in troubleshooting something like this. I’m also a podcast fan and appreciate the medium. This was an excellent gig that took a few hours to complete, and what seemed like a big problem at the beginning of the project turned into an opportunity. We’ll see what the future holds.

Find me
Latest posts by Ed Lozano, Managing Editor, Guitar Space (see all)