How to Fix Multi-Voice Recordings With Multiband Compression
An Introduction to Equalizing Audio With Multiband Compression
08.01.19 | Podcasting | By: Adam Townsell
One of the most prevalent scenarios that we at Resonate (and I’m sure you DIY’ers as well) run into with podcast recordings, is more than one voice recorded onto the same audio track. This isn’t always a problem, but it does limit what we can do when mixing and has the potential to adversely affect the finished product. One of the tools that I’ve found useful in tackling multi-voice podcast recordings is Multiband Compression. Before we dive deeper into this topic, let’s take just a second to define a few things:
Equalization: the process of adjusting the balance between frequency components within an electronic signal.
When the above definition mentions frequency components, it is referring to the range of audible frequencies present in a sound; in our case, a person’s voice. In very broad terms, we think of low, middle, and high frequency ranges when equalizing sound. So for example, if someone’s voice has too much low frequency content it could sound boomy and indistinct; too much high frequency content can make someone’s voice sound harsh and difficult to listen to. In the mixing process, we try to find a balance between the two using equalizers (EQ) to adjust the sound. Below are two graphics showing an EQ set to remedy the two scenarios I described:
When voices are each recorded separately, it’s relatively simple to adjust EQ settings on both voices to correct any issues and optimize how they sound. When recorded onto the same track, however, any EQ adjustments that you make will be made to both voices. So, if I have one voice with too much low frequency content, I could cut it to balance their voice. If the other voice (or voices) on the same track already have the right amount of low end (or not enough!) they would also have low end removed and could wind up sounding thin and brittle. Thankfully, EQ can be combined with another tool to mitigate this issue.
Dynamic range compression (DRC) or Compression: an audio signal processing operation that reduces the volume of loud sounds or amplifies quiet sounds thus reducing or compressing an audio signal’s dynamic range.
Compression is another commonly used audio processing tool, used to reduce the dynamic range (the difference between the softest and loudest sounds) of a recording. When using a compressor, the engineer can set a volume threshold; any sound that crosses above this threshold gets turned down by a certain ratio by the compressor. This brings us to the heart of the issue: how to use multiband compression when equalizing multi-voice tracks.
We’ve covered frequency bands, EQ, and compression in (very) broad terms; a Multiband Compressor, then, is a compressor that is split into several different bands across the frequency range. You may also hear them referred to as an Active or Dynamic EQ. This allows you to set separate compression parameters (such as threshold) for each different frequency band. So to continue our same example, I could create a low-frequency compression band and adjust my threshold so that the person with the booming voice triggers it (causing low frequencies to be turned down), while the other voices are left relatively unaffected.
To start treating this track, I set up a multiband compressor with two frequency bands in the low range. The screenshot below shows these settings in action while the boomy voice is talking. The yellow line in the middle shows the change in the volume level across the frequency range, and you can see that the two low bands I set up are reducing the volume by a decent amount. This is because the booming voice has much more low frequency content, causing those frequencies to cross over the threshold I have set. I also have a fairly long release time set up (release is how long it takes for the compressor to stop reducing the volume and return to normal), which I find smooths out the sound of the processing and makes it more effective for overall balancing.
The next screenshot shows the exact same settings, but while the thinner, less boomy voice is talking.
If you look at the yellow line again you can see there is very little compression occurring in the low frequency bands on this voice.
The same principles apply when trying to tame an especially harsh vocal recorded onto a track with others that don’t have the same issue. The screenshots below demonstrate this example; harsh voice speaking first, mellow voice speaking second.
The possibilities extend beyond the two simple examples provided here; you may have someone with lots of content in the mid range that sounds very nasally and have to set up some bands there; I have even found myself in situations where I was using a decent amount of compression on the low end as well as the high end, leaving the middle range untouched.
There are some limitations to using this tool on multi-voice tracks. Each voice will have content across the entire frequency range, even though they may be more or less than others. This means that it can be difficult, or in some cases impossible, to process one voice and leave the others completely untouched. It is important to use caution and listen critically when applying multiband compression on this type of track. While this technique will never be as effective as recording each voice separately and having total processing freedom, it can help you find a workable middle ground for your tracks.
Here are a few solid, quality multiband compressor plug-ins that are available for your digital audio workstation:
This is the plug-in that you’ve seen used throughout this article. This one is useful because you can create up to six frequency bands anywhere you need them with a single click on the interface. The only bands that appear are ones you create, so you don’t have anything extra to check on or distract you as you work. It is also extremely easy to move, adjust settings, bypass, and split bands as you need to. It also has a helpful real-time frequency analyzer for those moments when you can hear that something is wrong but can’t quite pin it down.
Izotope Ozone Dynamic EQ
Detail is the name of the game on the Dynamic EQ from Izotope, part of their Ozone suite of plug-ins. When opening this Plug-in, four bands are activated by default, with two more available. The interface presents you with detailed real-time information whenever you click on a band. You can easily change the shape of a band, see a representation of the gain reduction being applied and swap between digital or analog characteristics. Like the Pro-MB, it provides a real-time frequency analyzer across the frequency spectrum.
C6 is an update of the classic C4 multiband compressor from Waves. The main difference here being that C6 gives you 6 bands to work with rather than just 4. What’s interesting about this plugin is that four of the bands start in place (low, low-mid, high-mid and high) and the remaining two can move freely anywhere you need them to be. The four starting bands can also be adjusted. This feature makes C6 a good starting point for those who may be unfamiliar with multiband compression and overwhelmed by some of the options on the others.
We hope this blog has been helpful for you as you continue to improve and hone the audio quality of your podcast. If you want to learn more, schedule a call with our team. We would love to learn more about you and your podcast and find ways to make podcasting easier for you.
Resonate Recordings is a comprehensive podcast production company. Headquartered in Derby City–Louisville, Kentucky–we are committed to developing partnerships with our clients, not just performing transactions. Since 2014 it’s been our mission to make podcasting easy for businesses, brands, entrepreneurs, and individuals. We do this by providing support with podcast launch, podcast consulting, podcast editing, podcast production, and other creative podcasting services. If you have questions or are looking to start a podcast, our in-house team is available and ready to help! We would love to schedule a call with you and learn more about your podcast needs and answer any questions you may have. We look forward to hearing from you soon!
- Decrease Podcast Listener Drop-off Through the Art of Sound Design
- Best USB Microphones
- Announcing Black Mountain Media and Culpable Podcast
By: Adam Townsell
Adam earned a BA in Economics with a minor in Music from Berry College in 2010. Adam enrolled in graduate school at Middle Tennessee State University in 2013, graduating with a Master’s in Recording Arts & Technologies in 2016. As an Audio Engineer, Adam focuses on premium editing services & oversees our team of Audio Editors.