In last week’s article, I mentioned the Audient EVO interface series, which effectively handles the task of setting input levels for recording. If you have been through one of my voiceover recording classes, we covered the concept of setting those input levels correctly from a few different angles. However, I’ve received enough follow up questions about setting those levels to make me realize that step remains a bit mysterious in our voiceover studio recording workflow.

Input is not Playback

We are used to setting “Playback” volume. When you turn the knob up on your speakers, things get louder. We know that intuitively. We often incorrectly engage in that same process when we record: turning the volume knob up or down so that the audio plays back at a comfortable level. While that does protect our ears (highly recommended, by the way), that’s the wrong place to change things for recording. 

To capture the audio in our software we need to be aware of the INPUT level – how loudly the software “hears” our performance. This is called the Input Gain. It is set through a separate step which takes place first. 

Devices such as our phones actually set this input level for us – often in a dynamic way.  It’s why things might sound good when someone is talking, but then get noisy when they stop speaking. A simple automatic Gain system might be actively trying to compensate for lower input levels. This type of ever-changing auto-gain won’t work for voiceover because background noise will increase when we get quieter. 

For voiceover recording we use a “fixed” Input Gain level. We want to set things at a level where nothing gets distorted while still capturing enough signal to work with. In other words, there’s a sweet spot.

Peak is Key

The simplest way to control our recordings is to be mindful of the decibel (dB) “Peak” value. Remember, we work in a negative scale when recording. As recordings get louder, the numbers move “up” the negative scale towards zero. For example, -48 dB is quieter than -24 dB. 

The Peak value is the loudest single moment of our performance. If that loudest part goes above -0 dB, it records with distortion. That causes a buzzy, harsh, overly crisp kind of sound. If we distort when we record, there’s generally no way to fix it. Recording too “hot” means we’ve got to start over and be brilliant again.

We control our Peak through the Input Gain control on our audio interface (or possibly through a software control in our computer if we are using a USB-Direct-Connected microphone. When we increase our Input Gain by turning the knob, moving the slider, or changing settings in software, the Peak value will increase in our recordings. 

A good target value for the Peak in our recordings is no higher than -12 dB. If our loudest bits hit that level, we have plenty of signal without distortion. In many recording software metering systems, this is where the level bar changes from green to yellow. To be clear, this is a rough target – if you go slightly above that, no worries. Aiming at -12 dB provides “headroom” above our recording. That gives us a margin of error to compensate for a little extra energy in a performance when we’re behind the mic.

Identify The Variables

The biggest variable in this equation is our performance level. The human voice has a tremendously wide range of volume. You’ll also notice that the Peak moves around all the time. That’s to be expected. We just want to make sure it doesn’t go much above -12 dB. 

We can think about setting the Input level as an equation where the Energy of Our Performance is added to the Input Gain Setting to equal Peaks Hitting Around -12 dB. Simply put, when we are going to be loud, we need to reduce the Input Gain on our interface. Conversely, in a quieter performance we might nudge the Input Gain up a bit. 

(As I’ve mentioned before, recording a 24 bit WAV at 48 kHz* provides both the precision we need and enough dynamic range to cover us for almost anything we do in voiceover recording).

Run the Equation – Once!

We are not going to change things “on the fly” during our performance. As long as the loudest bits are hitting in that -12 dB-ish range, we’re covered. 

With a little trial and error, you’ll likely end up with two or three basic Input settings, depending upon which performance level you’ll be giving. For example, I’ve found that about 2 o’clock on the input gain is my starting setting for “Average” type performances. Louder video game characters mean I’m down around 10:30 or so. In both cases, the recordings are reaching Peak values right around -12 dB. I simply set that before I step behind the mic.

Delivery volume is not (necessarily) recording volume

All of the above steps are for capturing the performance we give at the microphone. Delivery volume is a separate process. Where we deliver the audio determines how loudly we might adjust things. That will vary for audiobooks, agency/online auditions, or the delivery specs your client has requested.

Setting the Right Recording Level: Key Points – 

• Set levels conservatively – Peaks hitting around -12 dB is a good target.
• Use an effective sample rate and bit depth in your recording file – my go-to is 24 Bit / 48 kHz (44.1 kHz for audiobook work). 
• Delivery volume is a separate step – it varies by genre and project.

Next: How Loud do They Want It?

This information recently went out directly to my email community. Each week I send out a new article about using technology in the home voiceover studio, and how to balance those challenges with the need to be creative behind the microphone.
If you would like to join in to receive those emails the day they publish, please take a moment to share your contact information through this sign up form.
Thank you!