Compression For Podcasts
Compressors are perhaps the most misunderstood and misused of all the tools podcasters have at their disposal to edit podcasts.
They’ve been on every single episode of a big name podcast that you’ve listened to, yet you are probably a little shaky on what they’re doing, and when to use them. Today we’re going to make clear what a compressor does, what all of its components are supposed to do, and the best places in your FX chains to use them. We’ll also go through the simplest free compressor that you can start using right now to get your podcast levels more in line, and the best way to set it up. It might seem tricky right now, but getting a good handle on how compressors work can save you a lot of time in the editing bay trying to get levels right.
They’ve been on every single episode of a big name podcast that you’ve listened to, yet you are probably a little shaky on what they’re doing, and when to use them. Today we’re going to make clear what a compressor does, what all of its components are supposed to do, and the best places in your FX chains to use them. We’ll also go through the simplest free compressor that you can start using right now to get your podcast levels more in line, and the best way to set it up. It might seem tricky right now, but getting a good handle on how compressors work can save you a lot of time in the editing bay trying to get levels right.
These are the settings I use to start to rein in the peaks of a voice track. I put it in before the EQ, to ensure that if the level is too hot, the EQ won’t boost the signal into clipping range, so any compressors later on in the FX chain won’t inadvertently boost more of the signal into clipping range with make up gain
A compressor is a way to control the dynamics of our voice. It does this by bringing down the loudest parts of our signal (the beginnings of words and sentences, puh and tuh sounds etc.), so that there is less dynamic range between the loudest and quietest parts of the podcast. When the dynamic range is smaller, you can amplify the whole signal more before it starts to clip.
For example, let’s say you record an episode and one speaker is very quiet when he talks, but laughs extraordinarily loud. When he talks, his voice is much quieter than everyone else’s, peaking at -21 dB, so you would like to bring the volume up. However, his laughter peaks at -1 dB, so you could only bring up the level of his track by 1 dB before his laughter starts clipping, and yet his speaking voice would still be much too quiet, because the dynamic range is 20 dB.
With a compressor, we can bring down the volume of the loudest parts of his laughter to -11 dB, meaning our dynamic range would now only be 10 dB, and you could raise the whole track’s audio now by 11 dB before it started clipping. All of a sudden, you have a lot more leeway to match the audio of our quiet talker with the other speakers in the podcast. This is what makes compressors so powerful.
Unlike an equalizer, which is a static effect that applies the same processing to the incoming signal no matter how the signal changes, a compressor is reactive, and only affects incoming audio when a certain parameter is met.
Because compressors deal in volume, the parameter it looks at is the loudness of the incoming signal and it only starts to apply processing once the audio crosses a Threshold that we determine. In our example above, because we wanted to only bring down the volume of the laughter, we could set the threshold at -20dB and all of his speech (which peaked at -21 dB) would pass through the compressor unaffected.
A compressor is a way to control the dynamics of our voice. It does this by bringing down the loudest parts of our signal (the beginnings of words and sentences, puh and tuh sounds etc.), so that there is less dynamic range between the loudest and quietest parts of the podcast. When the dynamic range is smaller, you can amplify the whole signal more before it starts to clip.
For example, let’s say you record an episode and one speaker is very quiet when he talks, but laughs extraordinarily loud. When he talks, his voice is much quieter than everyone else’s, peaking at -21 dB, so you would like to bring the volume up. However, his laughter peaks at -1 dB, so you could only bring up the level of his track by 1 dB before his laughter starts clipping, and yet his speaking voice would still be much too quiet, because the dynamic range is 20 dB.
With a compressor, we can bring down the volume of the loudest parts of his laughter to -11 dB, meaning our dynamic range would now only be 10 dB, and you could raise the whole track’s audio now by 11 dB before it started clipping. All of a sudden, you have a lot more leeway to match the audio of our quiet talker with the other speakers in the podcast. This is what makes compressors so powerful.
Unlike an equalizer, which is a static effect that applies the same processing to the incoming signal no matter how the signal changes, a compressor is reactive, and only affects incoming audio when a certain parameter is met.
Because compressors deal in volume, the parameter it looks at is the loudness of the incoming signal and it only starts to apply processing once the audio crosses a Threshold that we determine. In our example above, because we wanted to only bring down the volume of the laughter, we could set the threshold at -20dB and all of his speech (which peaked at -21 dB) would pass through the compressor unaffected.
Here is a compressor from Ableton Live, used here to show a GUI you might be more familiar with. The yellow circle in the centre is our Threshold, the point at which the compressor would normally take effect, except the Knee (the three dotted vertical lines around the threshold mark) is set to 6 dB, in this case applying gradual compression 3 dB before the threshold, The yellow dot is the level our signal is currently at, and the orange GR (gain reduction) shows a visual representation at how much effect the compressor is having on the signal, which the compressor is then bringing up the total signal by, because Makeup Gain is on
Once audio crosses that threshold, our Ratio determines by how much the compressor will reduce the dynamic range. Setting a good ratio is the thing that seems to trip up so many new podcasters, because it is tough to wrap your head around exactly what a ratio is.
When you’re setting a ratio, you’re setting the amount of incoming volume (in dB) it will take to make the compressor go up 1 dB above the threshold. So a ratio of 2:1 means for every 2 dB above the threshold the audio is, the compressor will only rise by 1 dB, while a ratio of 10:1 would mean the compressor only increases a decibel for every 10 dB of incoming audio above the threshold.
Once audio crosses that threshold, our Ratio determines by how much the compressor will reduce the dynamic range. Setting a good ratio is the thing that seems to trip up so many new podcasters, because it is tough to wrap your head around exactly what a ratio is.
When you’re setting a ratio, you’re setting the amount of incoming volume (in dB) it will take to make the compressor go up 1 dB above the threshold. So a ratio of 2:1 means for every 2 dB above the threshold the audio is, the compressor will only rise by 1 dB, while a ratio of 10:1 would mean the compressor only increases a decibel for every 10 dB of incoming audio above the threshold.
Here is two compressors with the exact same settings except for their ratios. Notice the significant difference in the angle of orange line in the compressor on the right, where the ratio is 10:1. Also look at the significant increase in gain reduction, compared to the compressor on the left, whose ratio is set at a much more modest 2:1
Think of the ratio as the price the incoming audio has to pay to increase the loudness of the track. If you’re trying to prevent clipping, you’d make the incoming audio pay a lot more to ensure it doesn’t clip your audio, perhaps 20 dB or more to increase the volume by 1 dB, but if you’re just trying to smooth out audio for a more consistent tone, you might have charge less (have a lower ratio) of 1.5-3 dB. I really hope this metaphor is holding up for you.
However, sometimes you might not want the compressor to simply activate when audio passes the threshold and then shut off once it goes back down, especially if you’re trying to smooth out audio rather than prevent clipping. This is where the Knee comes in handy, because it basically “smooths out” the compressor, applying compression after the threshold proportional to the knee amount. So if the ratio is set to 10:1, and the knee is 10 dB, the compressor would only apply the full ratio after the signal was 10 dB over the threshold.
Think of the ratio as the price the incoming audio has to pay to increase the loudness of the track. If you’re trying to prevent clipping, you’d make the incoming audio pay a lot more to ensure it doesn’t clip your audio, perhaps 20 dB or more to increase the volume by 1 dB, but if you’re just trying to smooth out audio for a more consistent tone, you might have charge less (have a lower ratio) of 1.5-3 dB. I really hope this metaphor is holding up for you.
However, sometimes you might not want the compressor to simply activate when audio passes the threshold and then shut off once it goes back down, especially if you’re trying to smooth out audio rather than prevent clipping. This is where the Knee comes in handy, because it basically “smooths out” the compressor, applying compression after the threshold proportional to the knee amount. So if the ratio is set to 10:1, and the knee is 10 dB, the compressor would only apply the full ratio after the signal was 10 dB over the threshold.
Here the two compressors have the same ratio, but the knee on the compressor on the right slowly starts applying compression over 16 dB of volume, as opposed to the compressor on the left, which starts compressing the full amount at the threshold.
After that, we have to decide the Attack and Release of our compressor. The attack is how long it takes the compressor to react after your audio crosses the threshold. To tame peaks, we want it to be relatively short, even as low as 0 ms for digital compressors.
The release, conversely, is how long the compressor still applies processing after the audio is no longer above the threshold. This is to ensure that if audio peaks multiple times within a short period, the release will still be applying compression to the track, even if attack was too slow to catch the subsequent peaks. Some compressors will also have an Auto Release function, that will continually adjust the compressor’s release to avoid over compression and pumping.
These are the five basic parameters every compressor will have. However, if you’re following our guide to podcast editing, you’ll notice that our first compressor is Reaper’s ReaComp, which has a few more parameters. Like Pre-comp, which looks ahead of the signal and applies compression before audio is about to cross the threshold. Some people call it a “negative attack” because it has a less than 0 ms attack time.
It also has High and Low pass filters, which allow you to specify which frequencies the compressor will react to. If someone bangs the table excitedly as they talk, for example, you could adjust the high pass filter to 100 Hz, so it would only look at the signal from that frequency upwards.
The ReaComp also has options for Auto Make Up, which will bring up the whole signal by the amount of gain reduction it applies to the peaks, useful for making our quiet talker louder. It also features Classic Attack, which makes it operate more like old compressors, that would gradually stop compressing the signal if it went way over the threshold, however this is not particularly useful for us. What is useful to us is the RMS size, which basically smooths out the loudest and quietest parts of the audio according to a moving volume average. Try moving this upwards to 20-50 ms if you’re having trouble getting a consistent tone from a vocal.
After that, we have to decide the Attack and Release of our compressor. The attack is how long it takes the compressor to react after your audio crosses the threshold. To tame peaks, we want it to be relatively short, even as low as 0 ms for digital compressors.
The release, conversely, is how long the compressor still applies processing after the audio is no longer above the threshold. This is to ensure that if audio peaks multiple times within a short period, the release will still be applying compression to the track, even if attack was too slow to catch the subsequent peaks. Some compressors will also have an Auto Release function, that will continually adjust the compressor’s release to avoid over compression and pumping.
These are the five basic parameters every compressor will have. However, if you’re following our guide to podcast editing, you’ll notice that our first compressor is Reaper’s ReaComp, which has a few more parameters. Like Pre-comp, which looks ahead of the signal and applies compression before audio is about to cross the threshold. Some people call it a “negative attack” because it has a less than 0 ms attack time.
It also has High and Low pass filters, which allow you to specify which frequencies the compressor will react to. If someone bangs the table excitedly as they talk, for example, you could adjust the high pass filter to 100 Hz, so it would only look at the signal from that frequency upwards.
The ReaComp also has options for Auto Make Up, which will bring up the whole signal by the amount of gain reduction it applies to the peaks, useful for making our quiet talker louder. It also features Classic Attack, which makes it operate more like old compressors, that would gradually stop compressing the signal if it went way over the threshold, however this is not particularly useful for us. What is useful to us is the RMS size, which basically smooths out the loudest and quietest parts of the audio according to a moving volume average. Try moving this upwards to 20-50 ms if you’re having trouble getting a consistent tone from a vocal.
Most compressors will also have a Lookahead function, that will allow the compressor to perfectly anticipate transients (the beginning of words), that works similar to ReaComp’s Pre Comp
Now that we know what a compressor is, you should know that there are very few times in which one compressor will give you the best results in podcast editing. This is why I always caution against people using their mixer’s compression into their DAW, because you usually end up applying a 10:1 ratio at a low threshold, flattening the dynamics completely and giving your voice a very boomy quality that is more akin to radio djs. Instead, we at Podigy typically use 3 compressors operating at 2-3 dBs of gain reduction each, one before EQ to start to bring down the peaks, and the other two after EQ to smooth out the tone more evenly. This way you’re still getting the volume you need in your track, but it keeps the dynamics of regular speech, and you can still hear natural inflection of voices.
The easiest way for beginner podcasters to get multiple compressors working in tandem is Sonic Anomaly’s Trileveler 2, which works by having three compressors set to look at fast, medium and slow attacks that constantly adjust and work with each other to make sure your audio has a consistent volume. Just as we’ve outlined, the fast tames the peaks, the medium deals with the main sections of your voice, and the slow makes sure the whole signal is hitting our loudness standard (-19 LUFS for podcasts). In fact, looking at the three meters down at the bottom of Trilevler operate probably does a better job of explaining how to use multiple compressors in tandem than I ever could.
Now that we know what a compressor is, you should know that there are very few times in which one compressor will give you the best results in podcast editing. This is why I always caution against people using their mixer’s compression into their DAW, because you usually end up applying a 10:1 ratio at a low threshold, flattening the dynamics completely and giving your voice a very boomy quality that is more akin to radio djs. Instead, we at Podigy typically use 3 compressors operating at 2-3 dBs of gain reduction each, one before EQ to start to bring down the peaks, and the other two after EQ to smooth out the tone more evenly. This way you’re still getting the volume you need in your track, but it keeps the dynamics of regular speech, and you can still hear natural inflection of voices.
The easiest way for beginner podcasters to get multiple compressors working in tandem is Sonic Anomaly’s Trileveler 2, which works by having three compressors set to look at fast, medium and slow attacks that constantly adjust and work with each other to make sure your audio has a consistent volume. Just as we’ve outlined, the fast tames the peaks, the medium deals with the main sections of your voice, and the slow makes sure the whole signal is hitting our loudness standard (-19 LUFS for podcasts). In fact, looking at the three meters down at the bottom of Trilevler operate probably does a better job of explaining how to use multiple compressors in tandem than I ever could.
For TriLeveler settings (we explain how to get it in the Complete Guide to Podcast Editing) I would keep these settigns largely the same as listed above, except for setting all the "Max Gain" settings to 0, and all the Max Attenuation to 10 dB, and manually adjust the input trim. This way Trilevleler never unexpectedly raises the signal, only brings it down.
This brings us to where to put a compressor in your FX chains, as it can significantly change the sound of your podcast. Typically, on a single voice track, you want to put it after any noise reduction and gating, have one bring down the peaks before your EQ, and then your medium and slow response compressors after the EQ but before your De esser, which is our next episode. We do this because we need to incorporate the boost or cuts the EQ will add into our signal into the final amount. We put it before the de esser, otherwise the compressor would just bring up the de essed sound to its original level, which isn’t very helpful.
But what is helpful is our Complete Guide to Podcast Editing, which incorporates what we’ve discussed here into the FX chains we build to help save you time while editing podcasts.
This brings us to where to put a compressor in your FX chains, as it can significantly change the sound of your podcast. Typically, on a single voice track, you want to put it after any noise reduction and gating, have one bring down the peaks before your EQ, and then your medium and slow response compressors after the EQ but before your De esser, which is our next episode. We do this because we need to incorporate the boost or cuts the EQ will add into our signal into the final amount. We put it before the de esser, otherwise the compressor would just bring up the de essed sound to its original level, which isn’t very helpful.
But what is helpful is our Complete Guide to Podcast Editing, which incorporates what we’ve discussed here into the FX chains we build to help save you time while editing podcasts.