Just your friendly neighborhood audio engineer back to continue what we started in Part 1 where we covered some gear philosophy, microphones, and miking techniques. But a microphone is only half your sound. The other half – the key to tightening up your sound, keeping your voice steadily on top of your mix, and preventing clipping from volumes that are too hot – is the compressor.
As I mentioned in the first installment of this guide series, my plan is to also include some gear recommendations so that these guides are informational as well as practical. However, if you have questions about specific gear, please PM me rather than comment below. I’d like to keep the comments section devoted to questions about the guide and gear-related inquiries can quickly take over with a plethora of specific questions and nuances.
Before getting to some gear recommendations, I’m going to run through what a compressor actually does and how it works. That way, when you get yours, you’ll know exactly where to turn the knobs and what all the meters do!
Dynamics, Peaking, and Clipping
The dynamic range of an audio signal is the range between the softest and loudest parts (technically, between the highest and lowest amplitude). Using a compressor, you can affect the dynamic range of a mix to increase the perceived loudness and highlight the most important parts making sure softer sounds are not lost. The way this translates for a streamer is keeping their voice above everything else. Dynamic effects include compressors, limiters, and noise gates.
When the volume of an audio input exceeds 0db (also called unity) – at least in the digital realm – the signal is cut and the result is a painful, distorted clipping. While streaming, many people will increase the volume of their mics in order to be audible over the game or because they simply talk quietly and would like their voice heard. However, we naturally speak with a wide dynamic range and whether we sneeze, cough, laugh, or shout the volume is going to fluctuate. In order to tame those peaks and bring them back down so they don’t hurt anyone’s ears we apply compression!
A compressor works like an automatic volume control, lowering the volume whenever it rises above a certain level (called the threshold). Why would you want to reduce the dynamic range? By attenuating the highest parts of the signal (peaks), the compressor lets you raise the overall level of the signal thus increasing the perceived volume. This way, whether you’re shouting during the stream or quietly talking, everything you say is perfectly audible at a reasonable volume. No more RIP HEADPHONE USERS. Instead, when your volume crosses the threshold you’ve determined, the compressor will lower the volume according the set ratio (discussed later). Besides that benefit, preventing nasty clipping when your input volume exceeds 0db is a necessity to not sounding amateurish.
Compression can also make audio sound better when it is played back in different environments. Not every viewer of your stream is listening through the same type of device. Some are listening on cheap headphones, others through their TV or phone speakers, laptops, tablets, or even professional studio monitors. These various speakers produce audio in different ways and with emphasis on varying frequencies. Compressing the microphone signal can help make the sound fuller and clearer in a lower-fidelity playback situation and ensure your voice’s clarity regardless of your viewer’s device or environment.
Limiters (also called peak limiters) work in a similar way as compressors in that they reduce the audio signal when it exceeds a threshold. The difference is that while a compressor gradually lowers the signal level when it goes above the threshold, a limiter quickly reduces any signal louder than the threshold back down to the threshold level. The main use of a limiter is to prevent clipping while preserving the maximum overall signal strength. Depending on which model you go with, some outboard compressors have a limiter built in while others won’t. It’s not necessary so don’t feel the need to buy a model with a limiter.
Noise gates alter the signal in the opposite way that compressors or limiters do. While a compressor lowers the level when the signal passes above a threshold, a noise gate lowers the signal wherever it is below a certain threshold. Louder sounds pass through unimpeded but softer sounds, such as ambient noise, are cut off. Noise gates can be used mainly to eliminate low-level noise or hum from an audio signal.
The noise gate, for streamers, is commonly used to suppress unwanted sound that is audible when the input signal is quiet. For streamers this is an absolute necessity and you’ve most likely already utilized it if you’ve dug into your broadcasting software. You can utilize it to remove background noise such as computer fans, keyboard strokes, mouse clicks, lip smacks, tongue clicks, or whatever other weird noises you make with your mouth.
Not every unit will come with a DeEsser so I’m putting this whole section in brackets. If it’s relevant for you, then keep reading but a DeEsser is not necessary for a compressor, it’s an added bell and whistle so feel free to forgo it. In fact, none of the models I’ve recommended below have this feature.
Certain letters like the “s” (for which this feature is named) or “f” have lots of sibilance and produce harsh or painful frequencies. The DeEsser is a frequency-specific compressor designed to compress only a particular frequency band within a complex audio signal. It is used to eliminate hiss (also called sibilance) from the signal. The advantage to using the DeEsser instead of an EQ effect to cut high frequencies is that it compresses the signal dynamically rather than statically. This prevents the sound from becoming darker when no sibilance is present in the signal. The DeEsser features extremely fast attack and release times.
Setting the Knobs
Alright! You’ve got your compressor; now let’s go over where to set those knobs so you can get the most out of your new equipment. Keep in mind: these are guidelines not rules. Adjust the parameters of your compressor so that you’re happy with how your audio sounds. Regardless of your model, there will be some form of these knobs for you to adjust or engage. After explaining what each knob does, I’ll give a recommended starting point for you to work with on your new unit.
Attack and Release
The attack refers to how fast the compressor kicks in to attenuate the audio peaks exceeding the threshold. Attack times vary from faster (in microseconds) to slower (in milliseconds). Depending on the unit, sometimes the increments are notched other times they’re simply “faster” and “slower” with vague demarcations. In instances like those, you’ll have to trust your ear because: if it sounds good, it is good.
Release is the time taken for the compressed signal to revert back to its original state. These tend to be considerably longer than attack times so as to sound the most natural in equalizing the disparity between compressed and uncompressed audio. These can be 80ms to over 150ms depending on the signal.
Recommended starting point: I’d start with a relatively fast attack and slower release. If your unit has the time increments marked, begin with an attack of 14ms and release of 25ms. Remember, you don’t want to make your release so short that the compressor produces a pumping effect (by pushing the audio signal down and then letting it back up again) because this isn’t a Daft Punk mix.
Not every compressor has this knob and some have simply a button (like dbx’s “OverEasy”). In case your unit has this or something similar, I’ll leave this section here for you to read as this parameter can have a dramatic effect on your sound (certainly for the better). The Knee refers to how the compressor engages uncompressed audio and transitions that signal into a compressed state.
There are typically two types of settings for this: a “soft” knee and a “hard” knee. A soft knee gradually applies compression to the signal as it comes closer to the threshold, providing a smoother transition between the uncompressed and compressed signal. A hard knee indicates that compression is only applied after the volume level exceeds the threshold, which can sometimes be jarring – depending on your Ratio setting.
This knob determines how much compression is actually applied to the signal after it crosses the threshold. The formula is expressed in a number relative to 1, i.e. X:1, where X is how many decibels the input is reduced by once crossing the threshold. For every multiple of X the output, in decibels, is 1. If the ratio is set as 1:1, that indicates that for every one decibel in, one decibel comes out – or no attenuation.
This is perhaps most easily expressed with an example:
Let’s say my Threshold is set to -10db and I have a semi-constant input (in our case, the streamer’s voice) that’s peaking around +2db (so we’re clipping here). If we set the ratio as 4:1 that means that for every multiple of 4db over the Threshold, only 1db comes out. So, in our example, since there is a 12db span between our Threshold (-10db) and the audio peak (+2db) only 3db of signal will output after the compressor attenuates the volume. Meaning, after compression, our streamer’s voice will only be peaking at -7db, no more clipping!
Ratios from about 25:1 through ∞:1 are generally considered limiting (see earlier) and are primarily used to ensure the signal does not exceed the threshold at all. Audio going above that is squashed back down due to the incredibly strong compression.
Recommended starting point: So, in general, you want to apply enough compression that it tames those high peaks but not enough that it squashes your voice. A good place to start is 3:1 or 4:1 and then dial up the threshold until you see about 2-3db of gain reduction on your regular speaking voice. Remember, too much compression can squish the life out of an audio signal, in this case your voice, so be especially attentive to the quality and timbre of your sound.
As mentioned earlier, compression can be used to raise the overall perceived volume of an audio signal. Once the compressor is set, the peaks of our streamer’s voice are being attenuated back down to a similar volume as the rest of the dialogue. Output gain allows us to “make up” for the attenuation by raising the overall volume of the new, compressed signal.
For example, if the gain reduction meter indicates that, on average, there is 6db of gain reduction we can turn up the output gain by 6db so that even the quieter speech will sound louder. However, if you were constantly clipping before, even while speaking (as in our example in the Ratio section earlier) you won’t want to turn this knob to compensate for all the attenuation, otherwise you’ll be back to clipping again. So pay attention to your starting point!
This aspect has mostly been covered already. As mentioned above, this is the point that compression is engaged. Only signal that passes this level will have compression applied.
Recommended starting point: The threshold is actually one of the only settings it’s nearly impossible to give a good estimate for as it entirely depends on your ratio and signal strength. This knob is usually set last and “dialed in” until the desired amount of compression is applied. That said, a good guesstimate is -28db or -30db for typical vocals. Depending on your Ratio, you’ll want to raise the Threshold until you see about 2-3db gain reduction on your regular speaking voice.
So we’ve got a microphone and now a compressor – you’re ready for Part 3. In the next guide we’ll cover how to get all this analog sound into your PC. We’ll cover different models as well as what they do and if they would appeal to a single-PC streamer or dual-PC streamer. I’ll post it as soon as I can format it properly and I’ll keep this guide, and the future ones, updated with links to all the guides. I really hope this was helpful, informative, or answered some of your questions!