Is your music drowning out your speech? Use these simple tricks to keep your music quieter than your dialogue.
You have your project going. You’ve got the right music (perhaps from our selection), you’ve got the right guests, subject, mics, and everything. You’re even starting to see your audience grow, with daily gains of tens and hundreds of viewers and listeners.
But there’s still one thing causing your show to fall flat.
As you watch it and rewatch it, and then watch a Hollywood trailer, you’ll notice something massively different, besides the BROOMS and the BRAAAHMS. You’ll notice that when the voice over cues in, he’s loud and clear, and somehow the music has moved out of the way.
It has ducked.
Audio ducking is a massively important principle that a lot of beginner creators miss or don’t seem to care much about. It truly changes the output of your creation, whether it’s a YouTube video or a podcast. Getting to know the principle will greatly change your game and push something that’s at the edge of “professional” finally over it.
What is Audio Ducking?
Notice how maybe some instruments won’t even play while the singer is belting out her morose lyrics, and when she breathes or takes a break, the guitar or trumpet might play an extra diddly? Or when the guitar cues up for a solo, all the other instruments kind of back down and let that dude shine? The same happens in talk shows, movies, voice overs, and more. The central sound that you should be listening to (typically the dialogue), is given more sonic space, and everything else is moved out of the way, or lowered (ducked).
Take note in the following example two things. At the intro, we get some nice music, that then comes down in volume when the speaking starts. Now move forward to 3:13. there is easily audible speaking over music. When the speaking stops, the music comes back in. Two great examples of what is meant by audio ducking.
There are two ways to achieve ducking. One is to use a compressor, which will duck your music automatically. This is a great technique to use, especially if you’re doing a live podcast and you don’t want to worry about fiddling with any knobs or faders while you’re chatting. It’s not perfect though, and depending on your show, you might have to change the settings a lot. The other way is to adjust the audio manually, to get the exact levels that you want and make sure nothing sounds a bit, well, automatic.
Luckily with Create Music, there’s a really easy way to handle manual ducking, which I’ll get to in a moment, but first I’ll start with using a compressor.
Compressors normally compress the sound, and generally have the effect of making everything into an even level. Sometimes it makes what’s loud soft and making what’s soft loud. In that way, it’s almost the opposite of ducking. But when applied and used against another audio signal and the compressor is only activated when the other signal comes in, it can also be used to get signal A out of the way.
You can find many software compressors online, some free and some quite expensive. For basic audio ducking though, there’s no reason to go expensive. Some video editors also come with audio compressors, and in a previous blog I went over using a compressor in DaVinci Resolve.
Pretty much, all compressors work the same. So I’ll walk you through ReaComp, Reaper’s native compressor (is Reaper for you? Here’s our blog on free or mostly free Digital Audio Workstations). Though ReaComp comes native on Reaper, you can use it on any DAW and any video production software that allows .vst plugins. Here I’ll go over audio ducking using Reaper (to understand preparing Reaper for a basic recording setup, check out our blog on that).
Put the compressor on the music track.
Since the music track is the track that will do the ducking, that’s where the compressor goes. Once it’s set up properly, the speech on track 1 will tell the music on track 2 to turn down. That means something has to tell the compressor to listen for the speech on track 1. Press ctrl+t twice to put in two new audio tracks. Then on your music track, click FX and select “VST: ReaComp”.
Link the tracks
Now we have to make the two tracks communicate with each other by linking them together via Reaper’s routing feature. Go down to the mixer at the bottom of the screen. Click on “Route” on track 1. Under – Sends – select “Add new send”. Then click on where it lists your music track. Then on the routing of your music track, under – Receives – click on the second “1/4” you see and select “(New channels on receiving track)” and “3/4”. And voila, they’re talking to each other.
Connecting the compressor
We now how to set the compressor so that it’s listening to the speech rather than its own track. To do this, go down to “Detector Input” and select “Auxiliary input L+R”. Now the compressor will only turn on when it hears the speech.
First thing is to make sure that the speech is already easily audible over the music track.
The following is a list of what each setting does. I’ve put in some starting recommendations, but by no means should you just set it and leave it. Set it and experiment with the knobs, finding out what works best for your setup.
Threshold: How loud does the signal have to be for the compressor to kick in. The compressor effects everything going above the threshold limit. If you want more of a flat sound, you’d lower the threshold, if you want more of a dynamic sound, you’d raise it up just shave off the uncontrolled peaks. For audio ducking, it’s best to start with about -20 db.
Ratio: This refers to how much compression is applied to the signal. To start with, set it at around 6:1 or lower. This means that if the signal crosses the threshold by 6 db, then it will increase the output level by 1 db.
Attack: With a short attack, the compressor will activate quickly (ie more sensitive) when your voice comes in. With a longer attack, it will take longer for the compressor to be applied. If the attack is set too low, it will sound choppy, but if it’s too long, then it might not even kick in to duck your voice. Reaper has it set automatically to 3.0 ms, and that’s fine to start with.
Release: This is how long it takes for the sound to return to normal after the signal your voice stops. Reaper has this set at 100 ms, which is a good starting point. If it’s set too low, then the music will bounce back in you’re breathing. It’s probably best to err a little long on the release for audio ducking.
Knee: Directly relates to the attack and the release. This sets the level of how fast the compressor will reach the ratio and how fast comes back down after the release, and it works by gradually dialing it in from that many db below the threshold. So a hard knee with a fast attack will be super choppy (good for mixing kick drums and pads on a dance track), while a soft knee with a slow attack will be very gradual (this is the end you want to be on for audio ducking). Reaper starts the knee at its hardest at 0.0 db. Once you find your preferred attack, start to soften that knee up.
Reaper has some filter settings as well which aren’t really important for audio ducking, so leave them alone.
Create Music (Automation)
Dealing with compressors sounds complicated, right? That’s why we built a handy audio automation tool into our music app.
First, upload your video so you can check it against the music. If you have a podcast, you’ll have to convert it to a mp4 or AAC.
Then click on the “Automation mode”. That will turn on a yellow volume line over all your objects. Where you want to add a node, just double click. You’ll have to add two nodes in order to duck, and then use the first or second node to adjust how gradually you want to duck. Then on the other end of where your vocals are, add two more nodes to gradually raise back up the volume.
You want a nice, gradual ducking, so that the listener doesn’t notice/isn’t distracted by the change in volume. You can use the same feature for fading in and out.
Once you’ve got your audio ducking down, whether through compression or with Create Music’s automation feature, then your project will sound that much more professional. It’s the little things that can make big impacts.