In this article we're going to look at how to use compression in podcasting. It's a tool that's often talked about, and even more often misunderstood. But, used right, it can really improve your audio.
So, let's get to know the humble compressor, starting with the obvious question...
What is a Compressor?
Recording voice, instruments and other sounds can result in large differences in the recorded dynamics (volume levels over the course of the performance). This leads to some very loud and some very soft sounds in a performance. These variances are difficult to mix or sometimes inaudible in the track. A compressor helps compensate by limiting the dynamic range.
Dynamic range: the difference in amplitude (dBa) between the loudest and the softest audio in the track.
Let’s take a look at the settings on a compressor and see what makes it tick. I’ll be showing the Dyn3 compressor found inside Pro Tools, but nearly every compressor contains a similar form of these settings (except perhaps Gain Reduction, which varies from compressor to compressor). It should be easily translated into Audacity, Audition or any other DAW.
Note: Threshold and knee are also represented in the graph here and are discussed in the sections below.
My next go-to knob after setting the ratio is the threshold. Threshold is the smallest amplitude the compressor needs to detect before compressing. Any signal under the threshold is given no compression (1:1). Compression is applied to signals above the threshold.
Set the threshold at the minimum level of the desired audio input. That is, set the threshold to the lowest volume of the recorded instrument or sound.. Then adjust by ear until the desired threshold is found.
When setting any setting, your ears are more reliable than the numbers on the screen in determining which value is correct. While there are general numbers that will work for specific situations, no number is perfect for all situations.
Compressing that signal by 4.0:1 squashes that signal by ¼ (invert the ratio). This leaves us with an output dynamic range of about 10 dB (60 ÷ 4).
Any signal above our -15 dB threshold will reach at most an amplitude of -9. You’ll need to boost your output signal gain by 9 dB if you want your loudest levels to reach 0 dB.
Gain Reduction and Auto Gain Correction
Many compressors have a gain reduction meter in the levels or output indicator. They are a visual representation of your gain loss as we described above. In terms of our example, the meter would display its highest value at or around -9 dB. From there, we could adjust the gain knob until the output meter reached the desired level.
Some compressors have a checkbox or setting that allows for auto gain correction. Auto gain correction automatically sets your gain values according to the calculated gain reduction.
Attack and Release
I tend keep my attack settings over 50 ms to reduce pumping and harmonic distortion in the lower frequencies. At minimum, attack times should engage the compressor before the signal completely decays.
Release times should allow the compression to fall off naturally at the end of a sound. Too short release times result in clipped, unnatural sounding ends of words. Too long release time allows audio below the threshold into the compressor. Typically a release of 80 - 150 ms is enough for most dialog.
Soft Knee vs Hard Knee
‘Knee’ refers to the bend in the transfer curve between the compressed and uncompressed audio.
A hard knee will immediately compress the signal after reaching the threshold. A soft knee gradually attenuates the signal before and after the threshold to make a more gradual sweep into the compression.
Soft knee compression is often more pleasing to the ear. Use the soft knee for things like vocals and melodic instruments. Drums and concussive sounds often sound better with a hard knee.
Attenuate: To reduce the level of a signal
What’s the difference between a limiter and a compressor? In actuality, a limiter is a type of compressor. More accurately, a limiter is a compressor set to a very high threshold and a very high ratio. Anything above the threshold compresses to keep the signal from clipping or going over the desired threshold, hence limited.
Any compression over 20:1 is a type of limiting. A brickwall limiter - common for recording loud percussive noises like explosions and gun shots - is a compressor set to a threshold of -1 db with a ratio of 100:1,inf:1 or ∞:1.
To clarify how this works as a type of compression, a ratio of ∞:1 means that for every infinite dB of gain of input (in other words for any gain) the compressor allows only 1 dB of output gain. Used properly, limiting keeps your loud signals from clipping. However, too much limiting can cause an unnatural squashed sound.
Out of all the tools in my audio arsenal, compression and limiting are the ones I use the most often. Properly used compression helps compensate for extreme volume adjustments in recorded tracks, allowing smoother, more consistent listening levels for the listener.