In this article, we hope to help you understand and implement the best compression settings for podcast and speech.
Your compressor must be set on lower ratios with a reasonably low threshold. Using slower attack and release times also might be quite beneficial, both of which should be set accordingly to the flow of your recording. Since we are looking for very subtle compression, we must use softer knee settings, which makes compressed and original sounds mix gradually. Sometimes you do not need any compression at all, but instead, you are better off with some subtle de-essing and gentle limiting.
Now let’s take it a little deeper.
Compression for speech and podcast
Whenever you have a desire to talk with people via podcast or record yourself explaining certain matters, you would want your recording to sound professional, right? Well, what sets apart all of the professional podcast recordings is that they always sound natural, and most of the time, listeners could not tell that something was changed. And there are a lot of ways to do that.
First of all, you could always buy some expensive microphones with, for example, tube preamps, which in itself will give you a very subtle compression and a silky high-end, or you can buy no less expensive digital consoles that are made explicitly for podcasters. Those consoles have already pre-made settings that could be recalled on the go. If you can afford all of those things, good for you, but for the rest of us, mere mortals, it is crucial to stay within a budget, often quite limited, and to be able to achieve the most with less.
Let’s gear up first
If you want your recordings to sound even remotely professional, you obviously can forget about using the built-in mic of your laptop or any other kind of consumer built-in recording gear. Those things exist only for convenience, and even if they sound decent for your personal home projects, you could not possibly achieve anything even close to professional sound.
Unfortunately, you are going to spend some of your hard-earned money on a couple of microphones (we recommend condenser) and an audio interface. The mic does not have to be expensive, just with a high noise to signal ratio, that is important because no matter what compression settings for speech, they will also make any unwanted noise louder as well.
As for the audio interface – the only thing you should be considering is how many people will be talking at the same time. And that will tell you how many preamps should your audio interface contain. Also, acquiring a couple of pop filters could be a good idea. It is very hard to get rid of popping sounds with software, so you might as well fix it beforehand.
Every time you are using a compressor, it is always useful to consider applying some EQ before. Unless your hardware has some onboard low and high pass filters engaged, you may do it with the EQ plugin. If your microphone has a frequency response that doesn’t particularly compliment your voice, you might fix it with EQ as well by reducing frequency bands that you do not like. Unless you are recording in a perfectly soundproof studio, you might apply some noise reduction. There are some great plugins on the market that can extract all undesirable noise with very little destructive influence on the original sound.
Now we are finally ready to compress our speech recordings. If by chance you are unfamiliar with what compressor is, you should know that it is a digital or, in some cases, an analog tool that reduces the dynamic range of any audio signal. You may perceive it as a very clever volume knob that makes loud parts of your audio quieter and quieter parts louder, which in turn makes the audio sound contained and refined. There are a lot of various types of compressors with even more different GUI designs, but they all pretty much have the same set of controls that could be applied. The most important are the attack, release, ratio, threshold, and knee.
The threshold is a controllable parameter with which you can tell the compressor at what volume of the input signal the compressor should start to work. When it comes to the speech, it is very crucial not to over compress your sound. So, higher thresholds are not something we are looking for in this particular case. Usually, lower thresholds of 5-7 dBs over the peak of your recording would be your best bet or at least a good starting point.
The attack tells the compressor when it starts to work after the signal hits the threshold. Faster attack times will compress the initial transients and might make your voice-over sound quite unnatural, but if the attack would be too slow, you might miss some quiet parts that needed to be compressed. So, an attack is usually set on a case-by-case basis. A good starting point would be 20-30 ms, then you will have to play it by ear.
Release shows us when the compressor stops compressing the signal, the goal here is to preserve the initial transience as much as possible and to make all of the quieter parts louder. So, if the release would be too slow, your sound may acquire some pumping effect, and if it would be too fast – you might get some percussive effect. None of these things sound natural when it comes to podcasts. The release settings will very much depend on the actual timing of your speech. A good starting point will be somewhere around 100-150 ms, from there, you should adjust it how it sounds better.
When it comes to compression setting for speech, the most important must be the knee. This parameter controls the relation between original and compressed signals. Hard knee makes the transition fast and very obvious and on the contrary soft knee makes it subtle and gentle. So, because our goal here is to have the most natural-sounding compression, obviously, we are going to use a softer knee. Set it somewhere around 8-10 dB and if you still hear how the compressor works, make it even softer.
The ratio shows how much the compressed signal would be quieter than the original one. That means that, for example, when you set your ratio to 10:1, your signal would be 10 dB quieter over each one dB over the threshold. So, naturally, we are looking for lower ratios, somewhere from 2:1 to 5:1. See what fits your particular recording the best.
Things to avoid
Unless you have some period-specific podcasts with some vintage vibes, chances are you wouldn’t get many benefits from compressors that emulate vintage gear and thus have a lot of coloration. You are better off with some digital compressor that sounds as clean as possible.
It is also very important not to overdo things. Some might argue that over-compressed sound is actually worth more than the sound without any compression. In music production, engineers tend to stack up compressors since different parts of the recording require different dynamic enhancements. When it comes to podcasts, even if you have a bunch of people talking in different mics, it is usually better to set just one compressor on the master bus.
Sometimes you actually might not use a compressor but instead, apply some other types of dynamic processing. You might use a limiter, which is a compressor with hard-set attack and release controls. The limiter makes your recording to sound louder by drastically limiting your dynamic range. Be very mindful not to overdo it since too much limiting will suck all the life from your recording.
If you have issues with hard-sibilance, which naturally occur in a human voice, you might want to use a de-esser, which is a type of compressor that works on the specific bands of the frequency range. De-essers usually have very few controllable parameters except for a threshold, so essentially they are very easy to use.