Get Background Music to Sit Perfectly Under Your Voice in DaVinci Resolve
A walkthrough of normalizing, silence removal, the Ducker, and master-bus loudness targets so music and voice stop fighting.
Posted
2 days ago
Duration
Format
Tutorial
educational
Views
21.6K
926 likes
57 · 43
Big Idea
The argument in one line.
A clean voice-over-music mix comes from layering simple steps: normalize each track to its own target, duck the music under speech, then verify loudness against the delivery platform.
Who This Is For
Read if. Skip if.
READ IF YOU ARE…
A YouTube creator editing talking-head videos in DaVinci Resolve who struggles to keep background music from drowning out the voice.
A video editor who normalizes audio by ear and wants concrete dB and LUFS targets to aim for.
A Resolve user who has never opened the Ducker's advanced controls or the Fairlight loudness meters.
SKIP IF…
You edit in Premiere or Final Cut and only want general mixing theory, since every step here uses Resolve menus.
You need broadcast or cinema mixing standards beyond the YouTube and social targets covered here.
TL;DR
The full version, fast.
Music and voice sit together when each track is normalized to its own target and the music is ducked under speech. Normalize the voice with the YouTube preset (about -1 dB peak, -14 LUFS), remove silences with Ripple Delete Silences while locking other tracks, and normalize music to a -12 dB true peak. Then tune the Ducker's look ahead, rise, recovery, and hold until short gaps do not pump. Finally, check the Fairlight meters against the YouTube targets and add a compressor and limiter on the master.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
Normalization only scales the whole track so its loudest peak hits the target; it does not touch anything in between, so one cough can set the level for the entire clip.
The YouTube normalize preset targets roughly -1 dB peak and -14 LUFS, which can run hot if you already record near -9 dB.
Music should sit much lower than voice: normalizing it to a -12 dB true peak is a sane starting point.
Resolve's Ripple Delete Silences works on any version, and locking the other tracks keeps it from shifting your video or music.
The Ducker is sidechain compression: the voice track triggers the music to drop whenever someone speaks.
Ducker look ahead decides how many frames before speech the music starts dropping, so it never lands late.
Rise time turns the Ducker's brick-wall drop into a gentle slope, and recovery does the same when the music comes back up.
The hold setting keeps the music down for a set time after speech stops, which stops short gaps from causing audible whooshes.
You can mute the mic track and still hear the Ducker work, which makes the short-gap swells easy to audit.
Volume and loudness are different: two tracks with similar peak dB can feel very different in perceived loudness, which is why Fairlight shows both bus and LUFS meters.
YouTube's targets are close to what TikTok, Instagram, and Facebook want, so one delivery preset covers most social platforms.
Compressing the vocal itself raises its perceived level against the music, so you may need to add more ducking afterward.
A light compressor plus a limiter on the master catches stray peaks without changing the tone much.
There is no correct music level; some creators run it loud and some bury it, so choose the effect you want.
Takeaway
Set levels first, then duck, then verify
MIX ORDER
A reliable music-under-voice mix follows a fixed order: normalize each track, remove noise gaps, duck the music, then check loudness against the platform target.
Normalize voice and music separately, with voice near -1 dB peak and music near a -12 dB true peak, so the baseline already favors speech.
Remove silences on the voice track only, with other tracks locked, so background noise vanishes between lines without shifting your edit.
Tune the Ducker's look ahead, rise, recovery, and hold together, and listen to the shortest gaps since those are where pumping shows up.
Trust your ears for the final ducker depth, because the right music level depends on the style of the video.
Check bus peak and LUFS meters against the delivery platform, aiming at -1 dB and -14 LUFS for YouTube and most social sites.
Add light compression to the voice and a limiter on the master, then revisit the Ducker since compression changes the balance.
Glossary
Terms worth knowing.
Normalize audio levels
A Resolve command that raises or lowers a whole clip so its loudest point hits a chosen target, without changing the balance inside the clip.
True peak
The highest level the signal actually reaches, including inter-sample peaks, measured in dB. It sets how close a track gets to clipping.
LUFS
Loudness Units Full Scale, a measure of perceived loudness over time rather than raw peak level. YouTube targets about -14 LUFS.
Ducker
A Resolve audio feature that automatically lowers a music track while a chosen voice track is speaking, using sidechain compression.
Sidechain compression
Compression on one track that is triggered by the level of a different track, used to push music down when someone talks.
Ripple Delete Silences
A Resolve audio operation that detects quiet sections by dB threshold and minimum length, then removes them from the selected track.
Fairlight
The audio page in DaVinci Resolve, with mixing tools, dynamics, and loudness meters.
Limiter
A processor that stops audio from going above a set ceiling, used on the master to catch rare peaks.
The Script
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
17px
If you're a content creator or video editor, there might come a time where you're trying to edit a project where you have someone speaking and have a music track going on down below.
Now trying to get those two things to work well together so the music's not too loud can be really tricky. But there's a few things that you can do to make that a lot simpler. Let me show you what those are.
So here you can see I've got the actual clip that I used for the intro of this video. And what I'm going to do is grab a music track and bring it down below that footage. And you can see just by looking at the waveforms themselves, the audio track from my vocal is much lower than the music track, and you can hear it when I play them.
So one of the first things I like to do is to try to normalize the tracks to levels that make sense for the project I'm working on. I'm going to start with my audio track, me speaking into the camera. I'm going to right click on that.
I'm going to go up to normalize audio levels. And because I am making this stuff for YouTube, I'm actually going to use the YouTube setting from the dropdown menu. Sometimes that can be a little too hot.
I tend to record around a minus nine DB and the YouTube setting is going to target about a minus one DB. One of the things you have to understand about normalization is it takes the entire track, looks at the very loudest parts and then makes that hit the targets. So if I had coughed, clapped my hands those would be the loudest parts of this track and it would normalize everything so that that loudest part is right around minus one db minus 14 luffs and then everything else would just come up along with it it doesn't separate any sections it's just basically turning the whole track up or down to hit a certain peak at its loudest part so i'll use youtube a lot of the times and it gives me a pretty good result you can see right in there that i've actually got that peak is what it's using to normalize where the rest of the stuff is kind of where you'd like it to be.
When we listen through and mute the audio track, it sounds decent. If you're a content creator or video editor, but there's a lot of spaces in here some between words some things i left and i want to actually get rid of or quiet down those spaces in case there's any noise going on in the background and there's a bunch of different ways to do it but one of the things you can do in any version of davinci resolve is actually cut those silences right out let me show you how to do that so with that track selected i'm going to go to the upper menu and i'm going to go to clip down to audio operations and i'm going to choose ripple delete silences and i'm going to get this pop -up window that starts deciding what a silence is.
Let me zoom in here a little bit so you can see what's going on. Now, these red sections you're seeing in my camera track with my audio, those are ones that Resolve is considering silence. You can decide what level in decibels, how loud a silence would be.
If I pull this back, you've moved it back so it's too loud to be considered a silence. I'd have to move that up quite a bit, and I'm watching for that red to select just the spaces in between where I'm talking. That's pretty much filled up that area.
Now, the prehead option, if I move that to the right and to the left, you'll see the bright end of that red section starts to move. Prehead is basically where that silence is going to be removed, how far before I start talking again, the presection.
And I want that fairly close to where I start talking. The post tail is after I'm done talking, when does it start considering its silence? So if I slid that to the right or left, then everything at the end of where I finished talking would start being affected.
And I want that fairly tight as well. And then below that, it's got the minimum to strip, how many frames. You can see right over here, this has only a few frames.
And if I raise that too high, it's saying it meets the decibel levels, but I'm telling it not to strip anything that's under 16 frames. And that is, so I can pull that all the way back to... 2 I think is the lowest value and it'll start pulling all of that out in there.
Now once I have this sort of dialed in so it looks like it's selecting all of the spaces pretty well in the way that I want without affecting my actual vocal. What I want to do is go right here and I want to turn on the position lock. That's going to stop any of that from moving.
And then right on here, I'm going to lock this main video track because I don't want that to be cut at all. I want that to remain intact. And I'm also going to lock the music track underneath, make sure the lock is white because I want that track untouched as well.
So now the only thing that should be affected is my vocal track and it should only be removing the sections without ripple deleting and moving anything together. Now I click Remove. Now you can see it's removed all those silences around my vocal track, which is cool because now it's nice and quiet at those parts in between.
I don't have to worry about background noise. Video Editor. There might come a time where you're trying to edit a project where you have someone speaking and have a music track going on down below.
Cool trick to know. Now we can unlock everything once again. We don't have to keep anything locked.
But I do want to normalize this music track right here. So the way that I'm going to do that is I'm going to right click on that music track and go up to normalize audio levels. And for music, I don't want it nearly as loud as my vocal mic might be.
And I have that pushing at near minus one. For music, I'm usually going to start somewhere around a minus 12 dB. So I'm going to switch from YouTube and I'm going to scroll up to true peak.
And I'm going to set that for a minus 12. And then I'm going to hit normalize and watch what happens to that track. You can see it pulls it way down and makes it much quieter.
And if we start playing through, let me lift this up a little bit. You can start hearing the balance between the music track and my vocal track. If you're a content creator or video editor, there might come a time where you're trying to edit.
It's better, but it's still a little loud when I'm talking. When I finish talking and then the music plays right in this space, it's not too bad. Going on down below.
So now the thing we want to do is take that track and whenever I start talking, we want to squash it down. Now you could go in there and use audio keyframes to keyframe the volume and turn it down every time you began speaking. But an easy way to do that is use a feature called the Ducker.
Let me show you that one. With this music track selected, I'm going to click into this gray area on the left -hand side of that audio track. And you see right at the bottom here, it says Ducker.
Now this feature is available in every version free or studio. So you'll be able to use this if you have the most recent version of Resolve installed. Let me turn on the Ducker and let me kind of explain what this does.
Basically, it's what we used to call sidechain compression back in my old audio days. And it's using compression to push down the volume of the music track every time it's triggered by the vocal track speaking. And if I crank this ducker way up, you'll be able to hear that in action.
Someone's speaking and have a music track going on down below.
So you can hear how it let the music be louder when I wasn't speaking and then quieted it down when I was. But dialing that in so that it works with your track is going to be something you'll do individually based on the kind of recording that you're working with and the project results that you're looking for. Now, in this case, I want mine fairly tight.
I would like it to kind of shoot up fairly quickly when I'm done talking and quiet down right before. So I'm going to pull this Ducker level back just for the moment. I don't think I need it quite that hot, but I'm going to open up these advanced controls and there's four different features here.
Let me explain what they do. The look ahead feature is going to decide how long before you start talking again that it should turn the music down. It's just looking ahead to hear when you're going to start talking, and it'll turn the music down either right before it or a few frames back before it.
So if I cranked up the look ahead, then I put the playhead right back here in this space, listen to how it will turn it down before I start speaking. So it was a little bit before, and if I took that same option and turned it all the way down, you'll hear the music duck almost exactly when I start speaking.
So that's a good one to leave up just a little bit to get a little bit of look ahead going on. This will totally be up to your particular project and your preferences. now the rise time is a slightly different one the rise time controls from the point where it's playing full volume and you start talking it's kind of a brick wall if you have it set to zero but the rise time sort of does a nice slow scoop down so if i pull this back here and i open up the look ahead time and crank up the rise time you'll hear it kind of softly pull the music down not as much of a brick wall now try so it's a smoother transition turning it down into your next vocal line.
Now, the recovery option at the bottom is kind of the opposite. When it's been squashed down and you're talking, when you stop talking, it will go right up to full volume. And if you want it to slowly come upwards and give it a little bit of a smoother increase, then what you would do is increase the recovery time.
Let's crank that way up and you can hear it right here where I stop talking and the music starts playing again. Crack going on down below.
Do you hear that smooth rise? Now, the only other option you have is the hold. If it squashes it down, regardless of the rise or recovery time, how long should it hold it down after you stop talking?
If I were to leave everything else down but crank the hold, it knows it's squashing it right here while I'm speaking, but the amount of time it holds it down for before it lets it come back up is going to be much longer. Music track going on down below. You can see it happened quite a few frames after.
So now we have to decide how to mix all four of these advanced options to work with what we're trying to create as a final product. Because I want this music to come in fairly quickly in that open space, I'm going to work with that first, and then I'll see how the other sections are affected. I'm going to move the look ahead up just a little bit.
I want a little bit of rise time. I want a little bit of recovery in there, and I'm going to bring the hold up just a little bit more, just to see how that sounds. I want this fairly quick right on this section.
Let's take a listen to where I'm speaking. space where the music comes back up and how it squashes it back down when i start again on down below now trying to get those two things not terrible it's a little bit harsh but i can tweak that a bit if i don't like it you can actually mute your main audio track the microphone track and it can still trigger this so you can turn the audio off completely and just listen to how it's being affected right now i've got my microphone track muted You hear that go up and down?
What this is really good for is you can go to some of these smaller spaces where there's no audio coming from my microphone and see if that music is going up and down. Because in those spaces, I don't want the music to quickly turn up and turn down. That would sound kind of weird because those spaces in between these sections are very short.
So let me listen to that.
That one's so fast. Can you hear it? There's a quick spike where it's allowing it to raise up and then compressing it quickly.
Sounds like a little whoosh sound. So that means I'm going to have to increase either the hold or the rise or the recovery. I think the way we'll start is bring up the rise, bring up the recovery a little bit.
So that'll push in those two crossfades together. Push up the hold just a tiny bit and let's listen to that section again.
That's actually pretty good. I'm not hearing any swells in those small spaces. I'll check the other ones.
Pretty good. Now let's listen to it with the vocal in and see where we want the overall ducker level right here. Now that we've got the advanced options kind of dialed in, we can now decide overall how much we want that squash down.
And that's really up to you. Some people like the music a little louder. Some people like it pushed way down.
Don't let anyone tell you what the right level is. It's the effect you're going for. Creators like Peter McKinnon use a lot of music very loud in their videos.
Some other creators like to have it way, way down low. You choose what's right for you. But let's back and let's listen if you're a content creator or video editor there might come a time where you're trying to edit a project where you have i wouldn't mind a little more duck than that i'm going to bring it up a tiny bit someone speaking and have a music track going on down below now trying to get those two things to work well together so the music's not too loud can be really tricky but there's a few things that you can do to make most of that seems pretty close i can tweak it if i want overall though Once I get these kind of working together, what I'm going to be paying attention to is the overall output for the delivery platform.
That's just a fancy way of saying, well, where does the thing go once you're done exporting it? In my case, I tend to upload to YouTube. So what I'm going to do at this point is switch over to the Fairlight page.
We move my playhead all the way back. And I'm going to go to the upper right here. And you see this section, the meter section.
If you don't see it, make sure you have meters turned on right here. That enables these meters. And make sure in the three -dot ellipsis, you have...
this set for whatever it is that you're trying to use as a delivery if you were making content that was going to be set to netflix you could use this option you can use all these different options to figure out what peak meters you want to see youtube is pretty much the same levels as almost any social media platform so whether it's tiktok instagram facebook any of those you can select youtube and it'll give you the meters you want and that's going to give us targets of a minus one for the main output or the bus and that's going to give us a target of minus 14 lufts for the overall loudness and i've talked about this before some people don't understand the difference between volume and loudness you can have two different pieces of audio that might be eq'd very different one might have a lot of bass response in it one might not have much at all and they could actually be very similar decibel levels but one sounds louder when you hear it it appears to be louder and that's what we refer to as loss it's perceived volume what your ear is hearing as the actual volume of it not just what the meters are reading so that's
That's why we have two sets of meters. So we're going to go for minus one on the main output and minus 14 for the loudness. And as I play this through, I'm going to look at these meters.
If you're a content creator or video editor, there might come a time where you're trying to edit a project where you... Right now, everything is fairly quiet. Watch the bus meter here.
It's way down below minus five. You have someone speaking and have a music... And over here, you can see the Lufs meter, this is the minus 14 mark, is far below that as well.
Track going on down below. I'm going to try to compress my audio a bit and get those peaks somewhere up near those two targets so that they don't peak higher than minus 1 and minus 14, but get up pretty close to them. That'll give me a volume that's not too quiet or too loud when I do upload it to YouTube.
Let's start by going to the master output and let's open up the dynamics by double left clicking right here. And I want to pay attention to these levels here. Content creator or video editor there.
And I'm going to turn on the compressor here and I'm going to bring that ratio up to maybe closer to not quite three to one, but 2 .4 to one. That's just how much compression I'm putting on it. And I'm going to increase the knee, which is how soft that angle is going to be.
in the way that it attacks the compression. I turn up the knee, you can see that point kind of softens over. And let me listen through to how that affects the audio.
If you're a content creator or video editor, there might come a time where you're trying... i'm going to bring the compression up just slightly by moving that back meaning the compression will kick in at a lower decibel volume trying to edit a project where you have someone speaking and that's actually pretty good now i can use this makeup game to raise that up and start trying to pay attention to those targets and try to get closer to each of them have a music track going on down below now trying to get those two things to work well together so the music's not too loud can be really tricky That's a little bit better.
You can see that both of those came up. I can reset this and I can start that again. But there's a few things that you can do to make that a lot simpler.
Let me show you what those... You can see down here, I'm literally hitting minus 14. So it might be a little too hot.
So what I'm going to do is leave the compressor on because the audio doesn't sound terrible, but I'm going to add a limiter and I'm going to bring that way up. I just want to very carefully limit the top so that if there are any very slight peaks, that limiter will stop it from going past a certain level. Little bit of compression, little bit of that limiter together could do really good things.
Now trying to get those two things to work well together so the music's not too loud can be really tricky. But there's a few things that you can do to make that a lot simpler. I wouldn't mind hearing a little more compression on my vocal itself.
The overall mix isn't bad, but let me add a compressor right here to my main vocal. And I'm just going to turn that on, soften up the knee, and just try to compress my own vocal itself because it's peaking more than the music is. And creator or video editor, there might come a time where you're trying to edit a project.
Let me solo it so I can hear just this effect on and off. Things to work well together so the music's not too loud can be really tricky. it doesn't change the overall tone too much but it definitely controls some of those peaks and just holds it together a little bit better i'm not adding any eq or anything right now i'm just trying to dial in the relative levels if i want to i can bring some eq into this and improve the sound of that vocal a bit but let's turn off the solo and listen to the entire mix and watch the targets if you're a content creator or video editor there might come a time where you're trying to edit a project once i've added compression i can hear that the music actually sounds a little bit louder than i like when i'm hearing my vocal because that compression is going to move things together and tighten them up so i can go back over to the edit page i can select that music track right here and on the ducker section i can raise that up a little bit more and add a little more ducking if you're a content creator or video editor there might come a time where you're trying to edit a project where you have someone speaking and have a music track going on down below
Now trying to get those two. I wouldn't even mind increasing some of the other hold rise time and recovery because it does sort of have a harsh introduction. I'd like it to be a little bit smoother when it comes in low.
Now trying to get those two things. That sounds a little bit smoother to me, but you might like something a little bit different. Now, all of the music I'm using in this video came from today's sponsor, Storyblocks.
Storyblocks is a 100 % human -made stock media library that's built to support creators. Their straightforward pricing makes using Storyblocks really simple. One price, unlimited downloads.
That gives you access to unlimited high -quality video, music, sound effects, images, DaVinci Resolve templates, and more. And unlimited downloads lets you test and experiment to find just the right asset that you need for every project, every time.
And every download supports a real working artist. Now, I've been using Storyblocks long before they ever were a sponsor because they save me time. I can go to Storyblocks and find things that might take me way too much time, money, or effort to film, record, or capture, and I can download as many things as I want until I find the perfect puzzle piece for my current project.
And Storyblocks is currently offering 15 % off of yearly subscriptions for a very limited time. If you want If you want to go check out Storyblocks for yourself, I'll leave this link down below.
Play around with these features and let me know what you think. And if you have any questions about what we did here today, feel free to hit me up in the comments section down below. And if you want to learn more about how to edit with DaVinci Resolve, click on the video that I have on screen now or the ones that I'll link down below.
Peace.
The Hook
The bait, then the rug-pull.
Music under a voice is easy to get wrong: too loud and nobody hears you, too quiet and it adds nothing. This walkthrough builds the mix step by step inside DaVinci Resolve, from normalizing to ducking to a final loudness check.
Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.
Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
A single yellow Text+ title becomes a bouncing, shaking, screen-blended character once you stack keyframes, camera shake, and a handful of whoosh sound effects on top of it.
From snapping every clip to the playhead with one keypress to auto-ducking music under a voiceover, a tour of the DaVinci Resolve features most editors skip past.
A screen-recorded walkthrough of DaVinci Resolve Studio's new MCP connection, where Claude imports footage, cuts best takes, strips dead air, and syncs a second camera, all from typed and dictated commands.
00:00
00:21
00:41
00:47
01:07
01:18
01:32
01:46
01:59
02:15
02:29
02:43
02:57
03:11
03:26
03:40
03:54
04:08
04:22
04:37
04:51
05:05
05:25
05:33
05:48
06:02
06:23
06:30
06:45
06:59
07:13
07:27
07:41
07:56
08:10
08:24
08:38
08:52
09:07
09:21
09:35
09:49
10:04
10:18
10:32
10:46
11:00
11:14
11:27
11:43
11:57
12:06
12:26
12:40
12:54
13:08
13:23
13:37
13:51
14:05
14:19
14:34
14:48
15:02
15:16
15:30
15:45
15:59
16:13
16:27
16:41
16:56
17:10
17:24
17:38
17:46
18:09
18:24
18:32
18:48
17:07
14:14
26:26
10:55
29:09
12:27