Audio compression can be a confusing concept but it doesn’t have to be. Simply stated, compression is the difference between the loudest and quietest parts of an audio source. It is a vital part of making your voice sound natural and consistent while streaming.
Imagine you’re streaming a game where you are hunting ghosts. You’re whispering to your viewers about what’s going on and then a ghost scares the hell out of you and you scream! If you don’t have compression then the whisper is too quiet, and the scream can be too loud. So, compression reduces the dynamic range, difference between loud and quiet, of your voice. This also prevents your mic from clipping and creating a really terrible audio experience for your viewers. You do have to be careful with compression though. The idea is not to have zero dynamics, or changes in loud/quiet, but to balance them a bit.
Practical Application
So, now that we have a basic understanding of the importance of compression, how do we apply? When you add a compressor filter in software like OBS you’ll have several options. Ratio, Threshold, Attack, Release, and Output gain. Let’s go over what each of these mean.
Ratio
By default the ratio is 10:1. This means that for every 10 decibels (dB) of audio input above the threshold, it will be attenuated to 1 dB. So 10dB over the threshold only outputs a volume gain of 1 dB. It “squashes” the audio down to 1 dB. This can make your audio sound terrible so it’s important to understand this concept. If we hit 20 dB past the threshold, the volume is reduced to 2 dB at that ratio. The ratio is only part of this equation. Let’s talk about threshold.
Threshold
Threshold is the point when compression begins. If your audio doesn’t go over this threshold, the compressor won’t engage. If the threshold is -18dB then any sound louder than that, will engage the compressor. Threshold tells the compressor when to turn on, ratio tells the compressor how much gain to apply.
Attack
Attack tells the compressor how quickly it needs to engage after the input signal passes the threshold. If you set this to 5 ms, then it will take 5 milliseconds for the compressor to chop the volume down based on the ratio that was set. 5 ms is faasstt. This attack time really depends on what you are wanting to accomplish. However, for podcasting, etc., 10-40 ms should be fine.
Release
Release is exactly what you’d expect, the opposite of attack. If attack engages after we pass the threshold by input signal going UP, then release is how quickly the compressor “releases” compression after we pass the threshold by input signal going DOWN. In our example, that threshold is -18 dB. So once our signal drops below that, the compressor releases based on milliseconds. Don’t release the compressor to quickly though. It can have negative effects on the audio. Conversely, too slow and the compressor never releases. 60ms is an okay starting point for this setting.
Output Gain
Output Gain, sometimes referred to as Makeup Gain, happens last. So we’ve hit the threshold, the compressor did it’s job but it lowered the volume of the signal, which is what it’s supposed to do. Depending on your application, you may want to add some gain back to the output, increasing the volume of the final signal. That level fluctuates, it is possible that for your application you won’t need to increase this at all. When I set up my compressor for streaming, I don’t tend to increase the output gain. We use it as a sort of limiter to keep screams, excited yells, etc., to a volume that doesn’t hurt our viewers ears.
Conclusion
This should give you a much better understanding of when, why, and how to use compression. It’s an important tool to have good quality audio, just don’t go overboard because it can cause your audio to sound really terrible. Hope this helps!