Audiobooks are quickly becoming a favorite method of ingesting the vast amounts of the Internet’s widely available information. As a voice over talent who regularly narrates and edits audiobooks, it is essential to create an extremely efficient process for the creation of long-form audio. This article will cover my personal method of cleaning up audio…especially long audio.
First, I use Adobe Audition CS6 exclusively for my editing. I have found that there is nothing quite like it for removing pops, clicks, breaths and other odd noises that seem to pour forth from your otherwise magical sounding mouth. AA contains the tools necessary to surgically remove every noise from your audio, leaving you left with ultra-clean, high-quality sound (providing you have good sound to begin with). AA’s waveform editing mode is also my preferred method of recording and editing mono files; as it is ultra clean and easy to work with.
I’ll start by saying that it is much better to prevent noise than it is to remove noise. We all know that every extra minute you spend editing is a chunk of cash out of your bottom line. If there was no pop or click to remove, editing would be much smoother; but, unfortunately, mouth noise is simply a part of human blabber, so when it comes through, ya just gotta get rid of it.
After doing everything to avoid noise, it is time to explore the process of cleaning up your audio. I prefer to use the AA Spectral View for this.
Adobe Auditon CS6 has a really great mode called spectral view. I reconfigured my hot keys to allow me to hit the ‘D’ key to toggle this view (and use the Nostromo Razer I mentioned previously for all my editing). When in spectral view, you see the waveform in the top half of the screen, and the spectral view in the lower. The spectral view is a visible representation of the entire aural spectrum of your audio, broken down by frequency and amplitude. It reveals things that are otherwise more difficult to represent…like mouth noise. After using the spectral view for a while, you will quickly start to see mouth noise before you even hear it.
So, on to the process:
Start by opening a file you’d like to clean. Any audio file will do, but a longer one will make it easier to learn the process. This needs to be a raw audio file. You can remove noise from a processed file, but it’s better to start with raw audio.
Next, listen to your audio for noise while watching the spectral view go by. For examples sake, just listen and watch for now…it will help you develop an understanding of what I’m talking about. As a side note: It can be easier to do this process in headphones. Using headphones can reveal noise much clearer…which is good for editing, but not so good for the final listening experience (if the audio isn’t well edited). Said basically, if you can hear a noise in headphones, it is likely the listener will too (though the conversion to a lower quality mp3 and mastering will mask some of it). After you start to become familiar with what the spectral audio looks like, head back to the beginning of the file.
Now, let’s clean this puppy up! As you listened back, you probably heard some mouth noise. Depending on your recording process (and hydration level), you may have heard a lot of mouth noise. Play the audio until you hear that first offending pop or click. There are a few ways to remove that noise. The traditional way (NOT in spectral display mode) is to zoom in really close, look at the waveform, find the pop or click in the audio, and cut it out (even in the middle of a word). This method works, however, it can mess with a word’s timing and sound. There is a better way.
Once you have found the noise, take a look at it in the spectral view. Can you see an obvious visual imperfection in the audio? Now, there are two ways to remove it in the spectral view without cutting anything at all. First, you can use the autoheal brush to simple paint the click out, or, second, you can highlight the offending segment of audio, and click autoheal to remove it. I typically use the latter. The algorithm in AA will, most often, remove the noise without changing the sound of the audio at all. Both methods take some understanding of how AA “heals” audio. It can mess up the audio if you don’t do it right, so play with it a while, and you will start to see how to remove noise without removing precious audio. Try starting with noise that is out in the open (in an area of what should be silence), and then move on to noise within a word.
Sometimes, the noise is just too prevalent to remove. In this cases, you simply need to re-record that section.
Some setup will make this process even faster.
I have my keyboard (actually the Nostormo Razer) setup so that I don’t have to move my hand at all to switch views, autoheal or even delete sections. I have the F key programmed as autoheal, the D key programmed as toggle spectral view, and the G key programmed for delete. This will remove some of the default functionality, so you can choose if you’d like to follow suit. I found that I personally didn’t need the default hot keys.
After becoming comfortable with this process, it is extremely easy to develop speed in editing, and as we all know, the faster you get the audio cleaned up, the less time you spend editing.
Thanks for taking the time to learn this process. It can be used very effectively in both long-form editing and short. If you have any questions, feel free to contact me directly. I’d be happy to point you in the right direction if you’re stuck.
Check out the video on how to get this done![youtube=http://www.youtube.com/watch?v=WZi49M-qjfk&w=560&h=315]