When auditioning, giving a loud, clean sound to your submissions is essential. The level at which you safely record those auditions, being careful not to overdrive the microphone or recorded signal, can leave you with a properly recorded, but weak sounding final product. Here’s how to fix that.
Electronically, you can “turn up the volume” by using a process called normalization to tweak the audition’s overall loudness before exporting to your MP3 file for submission. Normalization is a process available in any sound editing and recording software, including ProTools and Audacity.
Here’s a before-and-after example of normalizing a piece of audio. The top graphic is as recorded, and the bottom graphic is after normalizing the audio to full 100% digital volume. To hear the difference, click on the links below the graphic.
Listen to the audio as recorded
Listen to the audio with 0.0 dB normalization
Two things you’ll notice: the normalized version is louder, and the room noise is more noticeable. That’s because normalizing increases all audio evenly, by the same amount. This means that your words will be louder, but depending upon the amount of room noise in your space, that may be louder as well.
You’ll find the Normalization command under the Effects menu in Audacity. To use it, highlight the entire audition, then go to the Effects menu, and choose Normalize. You’ll be presented with a screen that lets you choose the level of normalization, where 0.0 is the loudest, and for every full dB you choose, you lower the loudness by about 8%. Audacity is usually set at -2.0 dB, and you can change that to -1.0 or 0.0. Leave the other checkbox (zero crossover error correction) checked. Click OK. You’ll see your waveform increase in height. You’ve now normalized the audio, and you can continue on with exporting your work.
Do this with every audition you perform, if needed. It really does act like a simple volume control, and adds to the power and presence of your auditions.
If you have any questions, feel free to comment below.
Great article. David also told me that if you normalize and hear the sound floor, your mic input level needs to be turned up in system preferences. :)
Gentle correction – the term is “noise floor” – the ambient noise in the room that is also part of the overall sound that is normalized. You shouldn’t notice it much, but if your sound is softly recorded, you just might.
What about Audacity’s “Noise Removal” effect? You record a few seconds of ambient noise to get a “noise profile” that is used to set the proper level.
Noise removal has nothing to do with proper levels. That function removes, in a not so predictable way, background noise from a region, based on a sample of noise within that region. It is rarely clean and precise, rather it’s meant to be able to distinguish words spoken in a high noise level recording environment. Noise removal should never be used to enhance an audition, as there’s damage done to the overall quality of your voice and performance.
Waaaaay better than “remove noise!” As we learned, that damages the voice and is usually used to forensic analysis…. no place in VO. :)
Is normalization available in Garage Band? (I have not yet been able to learn to work efficiently in Audacity, but I’m very familiar with Garage Band, so I’ve been using it, and I’ve been very satisfied with it.)
I actually don’t know – I don’t use garageband ever. If anyone does know, please post.
David
I’m using the AT2020 and trying to find the best input level to find a good balance of room noise vs. hot mic sibilance. I’ve been recording at .5 or 50% input but find that there is a LOT of mouth noise and sibilance after normalization and compression. Do you ever eq after all this to take that out or should I use a lower input level and then just deal with the room noise increase?
Mouth noise, I wouldn’t worry so much about. Sibilance, you can EQ with a lo-pass rolloff at between 9 and 10k – you’ll need to set your playback to loop a particularly sibilant passage, then use a parametric EQ to sweep across that space to find where your particular sibilance is, and then notch filter that particular frequency or range of frequencies down to lessen the harshness of your esses. It will also help with the same quality in your fricatives – an “f” sound is often as harsh as an “s” sound, and similar in construction and frequency range.
Hope this helps!
David
Awesome, Dave, thanks!