I noticed my basic livestream audio production post started getting a LOT of rechosts lately, so I thought it might be a good idea to chost some more production how-to's for people looking for slightly more dynamic, advanced setups. For this guide, we're gonna learn how to install third-party VST audio plugins on OBS Studio, as well as Equalization and Ducking/Sidechaining.
"Why Should I Care?"
These are some advanced audio production techniques in this guide, so you don't really need them to have perfectly adequate livestream audio. Most streamers do what I detailed in the previous post at the most and just stop there, so it's fine if you don't think this is worth your time. However, these techniques can bring your production quality to the next level and set you apart from your peers. Very few streamers implement Equalization and Ducking in their livestreams, but you can tell when someone's done the work to make their audio shine.
While it may seem counter-intuitive, video quality isn't actually the primary turn-off for other viewers - it's audio quality that makes viewers either leave your stream quickly, or stick around for more. Especially on Twitch, you're always gonna have some level of video artifacting and distortion, but the human brain is capable of putting pieces of the messy visual puzzle together to figure out what's going on. Audio, however, is a much more delicate process for the brain, and it takes more effort to understand an audibly noisy situation than a visually noisy one. In addition to that, many viewers actually keep livestreams running in the background while doing something else - almost as if the stream is like a podcast with a visual element, where they can turn their eyes to the livestream if they hear somebody getting excited to see what's happening. In conclusion: the better your audio sounds, the better your stream will be.
A Refresher
In the last post we learned how to adjust noise gates, compressors and limiters. We learned that noise gates are designed to cut out noises below a volume threshold, which in practice would cut out the sound of your breathing or lip smacking (unless you're blowing straight into your microphone like it's a balloon). We learned that compressors are designed to flatten the volume of your voice between a range of intensities, from whispering to shouting. We also learned that limiters can be used as an extra safeguard against clipping, which is what happens when a sound becomes too loud for an audio source and causes extremely unpleasant distortion.
In this particular chost, we're gonna learn some more advanced techniques. We will learn how to install third-party plugins for use in OBS Studio. We'll use a third-party plugin to perform a technique called Equalization, which will help enhance the fidelity and clarity of our voice for our viewers. Finally, we'll learn how to use a technique called Sidechaining or Ducking to further separate your voice from other audio sources.
Installing VST Plugins
OBS Studio allows us to install third-party plugins called VSTs to add effects to our audio. There is a wide variety of effects you can use, like Reverb, Delay, Distortion, and Equalization among many others. We're interested in finding third-party plugins for Equalization, so first we're gonna want to find an EQ VST that works for us.
Ideally, we want a VST that's lightweight so we aren't bogging our computer down with effects that hog computing resources from our other tasks like video gaming. Using an effect that hogs too much computing power can lead to frame dropping, which often appears to viewers as inconsistent choppiness in the picture, clicking in the audio or even buffering. It just so happens I have just the ticket: REAPER ReaPlugs.
REAPER provides a lot of freeware effects that are incredibly lightweight on computing, but highly effective. Some of their plugins even perform complex tasks like Multi-band Compression, MIDI to VST conversion and even a plugin that lets you use fucking Javascript to design your own effects... i think. But what we're interested in right now is just ReaEQ, their highly customizable multi-band equalization plugin.
On the home page for ReaPlugs' VST suite, we'll download the 64-bit package. Once downloaded, we can unzip the compressed package. Next, we gotta put these equalizers where OBS can see them. Take a look on the OBS VST Guide and look for the folders listed on the page based on your operating system.
Next, we're gonna open (or close and re-open) OBS Studio and add a new filter to our existing microphone/audio input chain with a Noise Gate, Compressor and Limiter. This new filter is the VST 2.x Plug-in. You'll see that this new filter gives you only a dropbox to select your VST, and a checkbox to open the interface when the filter is active. We can leave the checkbox unchecked, since we don't need to constantly observe our VST after the settings have been adjusted to our liking.
In that dropbox you should see a list of VSTs including the ReaEQ plugin we just installed. If it doesn't appear, you may need to move the VST into one of the other folders listed on the OBS VST Guide. Once you select the ReaEQ VST from the list, you can now click the Open Plug-in Interface button.
And there it is, ReaEQ! The developers of OBS Studio actually use REAPER plugins to test VST filter functionality in OBS, so this is pretty much the hassle-free EQ to use. If this plugin is crashing OBS, then uhhhh look on the forums, cuz i've never had an issue with them lmao.
Equalization
The plugin's interface may look complicated if it's using language you're not really familiar with, but I can teach you another neat trick in OBS to help you hear what the equalizer is actually doing under the hood. Close the filter window for now. Under our Sources tab we're going to add a Media Source and select an audio file from your computer. Use a recording of your voice for this audio file. If you don't have one, turn off all the filters in your microphone's filter chain by clicking the eyeball icons next to each filter and then record a clip of yourself talking in OBS. I like to find a Wikipedia article of my choosing and read the first sentence in a normal voice, the second sentence in a loud and excited voice, and the third sentence in a quiet voice. Once we've recorded our voice clip, select the recorded file in the Media Source, enable the "loop" checkbox and click OK. Then in the audio settings menu for our Media Source (found in the audio mixer tab where all the audio sources are displayed with volume sliders and audio meters), we're gonna select "Advanced Audio Properties." On the same line as our Media Source, we're gonna click the dropbox menu that says "Monitor Off" and select "Monitor Only" so we can hear the Media Source being played.
Now that we can hear our voice playing back to us through a recorded clip, we're gonna add some filters to our Media Source. Add a new VST filter to the Media Source, select the ReaEQ plugin and open its interface. We've got a lot of settings to go through, but I'll give you a guided tour through your most effective tools for tuning the sound of your voice.
Each 'band' is on its own tab. By default, ReaEQ gives us four bands to work with. Four bands are certainly enough to accomplish what you're looking for. The first 'band' will be set to the Low Shelf type by default. We're gonna change this to a High Pass instead. Select the 'Type:' dropbox, and select the High Pass type.
A 'High Pass' will gradually slope the volume of all frequencies below the Band's range downward into complete silence. It's called a High Pass because it lets higher frequencies pass through the filter while it restricts lower frequencies. In the context of voices, this is incredibly effective for reducing the explosive, bassy sound that happens if you breathe or spit into the microphone, or when you say "plosives" like P sounds, T sounds and K sounds as in the words Power, Tail or Crush.
Now we're gonna move this band around. You can click on the Band 1 pip in the audio visualizer and slide it around, or use the Frequency slider. I usually set the Frequency to somewhere around 100 hertz. You can also adjust the 'Bandwidth' of the band, which does what it says on the tin: it adjusts the Band's width. I have mine set to a slightly more aggressive 1.78 rather than the 2.00 default setting, but setting it too aggressively can cause the High Pass filter to kinda boomerang on the end of the band facing the higher frequencies. That boomerang effect can be fine-tuned with another band, but generally it's just not worth the effort.
Next, we have Band 2. The type is set to Band by default. You can see how it works by dragging the 2 pip around on the visualizer. Bands will raise or lower the volume of frequencies around it. Adjusting the bandwidth can make this area of effect bigger or smaller. You can also hear the difference it's making in your voice recording.
While everybody's recording environment and equipment will be different, generally speaking it's a good idea to perform a technique colloquially labeled in the audio industry as "scooping the mids." This can create a warm, comfortable sound to your voice that some in the industry liken to sitting in front of a campfire. HOWEVER, not every audio rig will respond well to scooping the mids. Headset microphones won't really get their best sound out of this technique, and require a very finicky trial-and-error process to make them sound good. For now, we'll just assume you have a standalone microphone like a Blue Yeti or an XLR microphone plugged into an interface like a Shure SM58.
To 'scoop the mids,' we first need to give our voice a little low-end presence. Adjust the second band's frequency and bandwidth until you're happy with how the low, bassy frequencies in your voice sound satisfactory to you. It's a good idea to keep adjustments somewhat subtle; I'd say the absolute maximum adjustment I'd make is the line at +6 or -6 dB, which sit between the middle and vertical edges of the visualizer. I have mine set to 202 hertz with a 0.96 bandwidth, which slightly sharpens the dropoff for the High Pass to reduce plosive impact but maintain warmth in my voice.
Next we're gonna do the eponymous Scooping Of The Mids. Grab the third band and drag it slightly below its default position. You can shift it left and right to hear how it changes the tone of you voice, but keep in mind what it's doing to the other bands on the visualizer. It's fine to use one band to affect another in the equalizer, but it's just good practice to be mindful how bands can affect each other and thus affect the tone of the sound you're changing. I have my third band set to 806 hertz with a 2.00 bandwidth with a gain of 3.8. This cuts into the second band's area of effect, which reduces the volume of some of the low-to-middle frequencies that I find makes voices unclear and muddy, while enhancing the clarity of my voice thanks to the mid-scoop.
Now, we can just leave it there if we're satisfied, but if you want a little extra shine to your voice we can use Band 4 to adjust the higher frequencies in our voice. This can give us a bit of a sparkly, airy feel to our voice. Rather than using a High Shelf for this band, we're gonna use the regular Band type to make these adjustments. I usually leave my 4th band neutral or even drop higher frequencies, but I'm a weird edge case because I don't have teeth - as in, the actual bones in your mouth that you use to chew food. This means I tend to make wet mouth noises quite a lot while talking, and the higher frequencies are where those really annoying, clicky, lip-smacky noises LOVE to sit. When I want to add some of that sparkle though, I'll put the band at 8500 hertz with a bandwidth around 1.1 to 1.4 with a gain of 2.6. This gives me just a touch of airy feel without being tiring to the ear.
Now, take note of the frequency, gain and bandwidth settings you have for each band, then put them into the EQ on your microphone filter chain. This part is a little tedious, but having a live preview of your adjustments makes the whole process of EQing soooo much easier than trying to make your adjustments with guesswork and the even more tedious trial-and-error method.
Finally, we need to put this equalizer in the chain somewhere. Remember, the equalizer will make your voice sound different depending on where you put it in the filter chain. If we place the EQ after the compressor in the filter chain, our EQ will generally have a warmer, bassier sound to it and the adjustments we've made will be exaggerated. If we place our EQ before the compressor in the filter chain, your voice will have a more crisp, sparkly presence. If you want to hear how the EQ sounds before and after the compressor, you can place the EQ before the compressor and record a clip, then place the EQ after the compressor and record another clip. Then listen back to those clips and figure out what you like best. (Remember to mute the Media Source you used to adjust the EQ!)
Ducking/Sidechaining
Equalization is exhausting, I know, but this final part of the tutorial is a fun, easy little reward for doing all that work. Ducking and Sidechaining are two different words that mean the same thing; the only reason you use one or the other is just based on what industry you're working in. In broadcast engineering it's called Ducking, while musicians and audio engineers call it Sidechaining. Generally speaking, this technique takes the volume level of one audio source and uses it to adjust the volume of another audio source. Broadcast engineers use Ducking to automatically make the crowd noise from a sports arena quieter while commentators are talking, then automatically make the crowd noise go back to normal volume when the commentators stop talking.
Subtlety is key here. We want the Ducking to feel more like a gentle wave on the beach, rolling in and out slowly and softly. Setting the Ducking too aggressively can make the audio feel very 'crashy' and jagged, like slamming through choppy waters in a boat.
We're gonna add a Ducking/Sidechain filter to the sources we want to make quieter when we talk. This means we're gonna be adding a Sidechain filter to our various audio sources like music, video games and other applications that make noise. If you're just capturing all of the audio from a particular output source like your headphone output or speakers, we'll just add a filter on that.
My settings may not necessarily work for you because Ducking is all about subtlety, and your audio setup and use case will be different from mine. My settings are for simulation racing esports broadcasts, which means my goals are to make the commentators' voices clear and understandable while maintaining the volume and intensity of the exciting car noises. Cars don't talk, so there isn't a concern that speech from the game will become jumbled with the voices of the commentators - allowing me to use rather aggressive settings that most livestreamers would probably want to avoid. I'm using heavy compressors on the game audio too, so aggressive ducking is practically necessary to ensure the commentators are heard without them getting lost in the noise. For your use case as a let's play livestreamer or maybe a let's chatter, you probably won't be using settings like this - these settings are here just to give you a general idea where to start and what to do to get them closer to your goals.
You might have also noticed I've got two Ducking filters; one for my own microphone, and one for Discord so the people in my voice chat will also duck audio on this source when they talk. If you frequently livestream with friends in a voice call, using multiple Ducking filters can be useful. However, keep in mind that multiple ducking sources activating at the same time might cause a sort of "stacking" effect where the audio being ducked is actually getting ducked twice as much. This isn't necessarily a reason to avoid ducking - it should be generally understood to avoid talking over each other in a livestream - but just something to keep in mind.
To give you somewhere to start for your settings, I usually use a 3.50 ratio with a threshold somewhere around -20 dB. My attack is quite aggressive, but my use case for Ducking is relatively niche so you might be better off with a more gentle 10 to 20ms response. My release is also set quite aggressively for my particular situation here, but 300 to 400ms is a nice, gentle option.
When you're setting your threshold, you can take a look at your microphone's audio meter - the green, yellow and red bars that light up when you speak - and visually take note of the numbers next to the meters when you talk. This will help you translate the visual information of the meter into decibel numbers, which you can then use to fine-tune your threshold on the Ducking filter.
Before we finish off here, we're gonna select your Ducking source. Click the dropbox and find your microphone in the list. Repeat this process with other audio sources as necessary, and you're done!
If you wanna see how it sounds, you can open a game - maybe with some music in the background - and start recording. Take some time letting the game audio run through unfiltered, then talk a little bit. If you're not sure what to say, just talk about what you're literally doing in the game - interacting with the menu, where you're walking, what you see around you, stuff like that. You can also make some short, sudden noises to see how the Ducking filter reacts to them. A simple "bark" or the standard Minnesotan "oop" would be good. When you listen back to the clip you recorded, pay very close attention to how your voice and the other audio sources interact when you speak. If it feels like your talking is making other audio sources "sink" too much when you're talking at a regular volume, you can adjust the Threshold higher (for example, -20dB to -15dB). If you feel it's sinking too much when you're loud, you can drop the Ratio by .5 or so. If you want the other audio sources to sink slower when you start talking, move the slider on the Attack setting very slightly to the right. If you want the other audio sources to come back to normal faster when you stop talking, you can move the Release slider to the left.
And there you have it! Equalization and Ducking is now in your arsenal of production tools!
If you need any extra help or explanation, you can send me an ask or shoot me a DM on Discord with the username hootwheelz. If you DM me on Discord, make sure you say something like "hey i need some help with audio on my stream," because I get a lot of scammy phishing DM's. If you wanna pay for my help that's all cool, but not at all necessary - these tips are intended to help the entire community, and helping others is reward enough! (but also we live under the crushing reality of capitalism and i have no income so if you can afford to pay me, i'll be very grateful lmao)
Additional Comments and Information
This is just some additional information that could be useful or interesting to people, but isn't necessary for the purpose of just getting your EQ and Ducking settings applied.
If you've been adjusting EQ or Ducking for quite some time and can't quite seem to get things to sound right, give your ears a break. Our ears can quickly become accustomed to the adjustments we're making, and exaggerate or shift how we perceive the adjustments we're making. Take off your headphones and get out of your chair, and let your ears become accustomed to the natural audio of your studio. Ear care is incredibly important; as an audio engineer who loves going to concerts, I greatly regret not being more protective of my hearing. I deal with mild to moderate episodes of tinnitus now, and hearing damage can permanently affect how you perceive certain ranges of frequencies. Giving your ears a break and keeping audio volumes at a comfortable level - either by turning down the volume setting on your PC or wearing ear plugs when going to a concert or event - and reducing exposure to loud audio environments will help you prevent hearing damage.
Some minor discrepancies in hearing also come with age. I remember being in high school during the time that smartphones had just become commonplace in the teenage demographic, and one trick students used to hear phone notifications without the teacher knowing was using what's called a "mosquito tone." It's a very high frequency sound that most younger people are able to hear, but due to hearing deterioration as people age, adults may not be able to hear it. It sounded almost like a really, really small mosquito flying just barely within earshot. A majority of people who experience age-related hearing degradation won't have a noticeable impact when it comes to audio engineering since the industry doesn't often give much delicate thought to the extremes of the frequency ranges, but musicians and music producers might want to take note of this since low sub-bass frequencies and the higher frequencies caused by cymbals, white noise and harsh-sounding lead synths might sound different as a result of this degradation. Generally though, age-related degradation is almost unnoticeable until you get to around retirement age.
Andrew Huang, previously known as Songs To Wear Pants To, is a musician who has hearing issues that developed (if i remember correctly) in his teenaged years. He actually can't mix and master his music the way that sounds good to him, because the average hearing capabilities of humans are different from his - and as a result, adjustments that sound good to him might actually sound terrible to me. Hearing is an incredibly important yet fragile sensory processing element of the human experience, so you should always try your best to take care of your ears. This guide is made under the assumption that your hearing is somewhere around average, so if your hearing is quite a lot different from mine, you may want to ask an audiologist about your hearing, get tested and try to learn how it affects your ability to perceive frequencies in comparison to the average person. Unfortunately, not everyone has access to an audiologist for a variety of reasons. You might be able to find some rudimentary audio tests online that play a simple sound at a frequency you can change, so you might be able to grab a couple friends and compare how you hear the sound. It won't be perfect, but it might give you an idea how your hearing is impacted and help you compensate for that when doing audio engineering.
