Make It Sound Like It Belongs: Seamless Sound Effects in DaVinci Resolve

We spend thousands of dollars and countless hours creating an amazing cinema look—but the best visuals in the world can’t survive being paired with a sound effect that doesn’t sound like it belongs.

A bad sound effect is worse than no sound effect. Silence? The ear can forgive. A footstep that lands like it’s happening inside your skull—while the actor is forty feet away? Your audience will be talking, but not about what you want.

So here’s how I make sound effects belong. Not a rulebook — just what works for me, and why.

Two things make a sound belong

For me it comes down to two: space and dynamics.

• Space — does it sound like it’s in the room the picture shows? Close or far, tiled bathroom or open field, behind a door or right in your face.
• Dynamics — is it as loud, or as soft, as it should feel — without clipping into a distorted mess?

Get those two right and almost anything sits. Get them wrong and the most expensive sound library on earth still sounds pasted on top.

The good news is that a free plugin called Place It does the space half for you, and does it well — we’ll get there in a minute. The dynamics half stays with you, but it’s mostly one habit, not a degree in audio engineering.

Pick the room — don’t become a sound engineer

Here’s the trap I lived in for too long. Resolve has a built-in Reverb effect, and it’s genuinely powerful — and, if you don’t already know what you’re doing, genuinely brutal. Dozens of dials, half of which I couldn’t have explained to you, and I’d lose twenty minutes coaxing one footstep into sounding vaguely outdoors.

The thing is, reverb done by hand asks you to be a sound engineer. Convolution — which is what Place It uses — just asks you to pick a room. For me that’s the whole difference: mixing my own paint versus pointing at the color I want.

Reverb still earns its place as a final nudge. But guessing at decay times to make something sound closer or farther? That’s the part most of us — video editors, not mixers — were never really equipped to do by feel.

The whole workflow, in one rule

Two stages, and the first one is optional:

1. Neutralize — only if the sound is messy. A recording off your phone, something captured in a room that wasn’t quiet, or busted audio baked into footage you were handed — that gets cleaned up first, until it’s neutral. (I use iZotope. There’s a free route too — more on both in the examples.)
2. Localize — always. Drop it into the timeline and balance it against everything around it, using Place It for space and plain old level for fit.

And here’s what saves you hours: a clean sound effect from a real library — Soundly, Epidemic, Artlist — skips stage one entirely. It’s already neutral. As a concept it feels responsible to “clean everything,” but in reality, running a pristine library file through cleanup just costs you time. That pass is only for sources that actually have a problem.

”Loud enough” — compared to what?

One shift in how I think about loudness, before we get our hands dirty. I don’t really care how loud a sound file is on its own. What a meter says about the whole clip in isolation, its LUFS number — for the kind of work most of us are doing, that’s not what I’m listening to.

What I care about is this sound, in this moment, against the sound right before it and the music right after. Loudness is relative. A scream doesn’t land because it hits some number — it lands because the quiet right before it set the trap. So I balance by ear, in context, and let go of chasing absolute values I was never going to hear anyway.

Now, the tool.

Meet Place It (and it’s free)

Place It is a free plugin from a company called Soundly — it’s at getsoundly.com/tools. Under the hood it’s convolution reverb built on true-stereo recordings of around forty real spaces. Which is the plain version of saying: it drops your sound into an actual room that someone went out and recorded.

Getting it into Resolve

Once it’s installed, you might have to point Resolve at it:

1. Preferences → System → Audio Plugins. If Place It isn’t listed, add the folder it installed into, then make sure it’s enabled.
2. Save, and restart Resolve.
3. Back on the Edit page, open the Effects panel → Audio Effects → VST. It’s in there now.
4. Drag it onto the clip you want to treat, and its control panel pops up.

Reading the panel

It looks like almost nothing, and it gives you a lot. Two big dials carry most of it:

• Left dial — Source. This emulates the thing the sound is coming out of, grouped by size — a tiny phone speaker on up to big systems. It has an on/off switch and a slider for how strong the effect is.
• Right dial — Space. The room you’re putting the sound into. This is the one I live in.
• Presets (the top dropdown) are ready-made Source + Space combos — “radio living room,” that sort of thing.
• The Wall slider in the middle muffles the sound, like it slipped behind a wall.
• The Space side has a decay control for how long the tail rings out.

The habit I’d build: for real-world sound effects, turn Source off. You almost never want footsteps sounding like they’re coming out of a phone speaker — you just want the room. Source off, Space on. For me that’s the move most of the time.

You can record this stuff yourself — yes, on your phone

Before we fix anything, a small bit of freedom: you can capture a surprising amount of usable sound on the phone in your pocket. Pouring a glass. Footsteps on gravel. A jacket zipping, keys hitting a table, a door clicking shut. You don’t need a thousand-dollar field recorder to start making your own foley.

The catch is that raw phone audio is never finished. It carries the wrong room, a layer of hiss, and a level that wanders. It’s a source, not a sound effect — which is exactly why it goes through stage one: neutralize it, then place it. (That’s the wine glass, a little further down.)

Worked example 1: Footsteps for a crowd walking away

Multiple sounds? Multiple sound files. It’s very easy to become lazy with these kinds of projects. Let’s take two or more people walking. As a concept, it seems like we could just take the same set of footsteps and use it twice. In reality, different people walk with different gaits, and unless you want it to sound like one person with four shoes, you need to find or create different footsteps. No one marches in sync — so what I’ll do is find multiple sound files, then vary the level between them. No one is the same weight or distance. Pan them a little apart so they sit in slightly different spots across the stereo field, and suddenly your ears think it’s a group of people.

Something to keep in mind: you actually don’t want people thinking about these kinds of sound effects at all. The goal is to not trip the mental alarm bells that draw attention to something that should’ve stayed innocuous.

And then there’s the leaving. If those footsteps hold the same volume while the actors shrink into the distance, your ear catches it instantly. So I drop Place It on each footstep track — Source off, a Space that matches the shot — and automate two things across the clip: the level coming down, and the Wall slider creeping up, so the sound goes a little more distant and muffled the farther they walk. The level fades, the room swallows them, and they’re gone.

Worked example 2: Filling a wine glass (recorded on a phone)

Now the opposite situation. There’s no perfect “wine pouring into my glass, this exact close-up” file in any library — so I just record it. Phone, a real glass, a bottle of water standing in for the wine. A few takes.

Because it came off a phone, it’s a messy source. So this one does get stage one.

Neutralize it (iZotope). I bring the take into iZotope and get it to be neutral: for me, I knock out the background hiss, EQ off the boxy, boomy coloration that the phone might have baked in, and set a sane level. What I mean by sane is just a reasonably loud sound that I will later on level inside DaVinci. I’m not trying to make it sound good or perfectly placed into my scene here — I’m trying to make it sound generic, so I can place it cleanly in the next step.

No iZotope? It’s paid, and you don’t need it to start. Audacity (free) will handle the neutralize pass — its noise reduction plus a little EQ get you most of the way. If you want to grow into the paid world later, iZotope RX Elements is the cheap way in; and if you already run Resolve Studio, its Voice Isolation can stand in for quick cleanup.

Localize it (Fairlight). Now Place It — Source off, Space set to whatever room the scene lives in, and gentle. A glass being filled is an intimate, close sound; I want a whisper of the room, not a cathedral. Then I nudge the clip in time so the rising glug lines up with the liquid climbing on screen.

For me the lesson is this: a phone recording you cleaned and placed beats a “close enough” library clip almost every time — because it’s your glass, in your room, doing exactly what the picture is doing.

Worked example 3: A scream that lands — without blowing out

This is the one that lands on my desk constantly: footage with a scream that’s already blown out. Recorded too hot on set, the waveform clipped flat across the top, distorted into a digital crackle.

So let me save you the hour I spent learning this the hard way. On consumer gear — 16-bit — a clipped scream is almost never coming back.

As a concept, it feels like there should be a repair for it. There’s a tool in iZotope RX called De-clip, and the promise is that it reconstructs the flattened peaks. For lightly clipped audio, it genuinely does. But a scream isn’t lightly anything — it’s the loudest moment in the take, clipped flat for its whole duration. De-clip works by reading the intact samples on either side of a clipped spot and drawing the missing curve between them; when the whole top of the wave is gone, there’s nothing on either side to draw from. I sat with several of these for over an hour. Not one was salvageable. (32-bit float gear is a different story — but if you’re handed a clipped 16-bit scream, the kindest thing is to accept it’s gone.)

So the real question isn’t how to repair it. It’s what to put there instead.

Best: re-perform it (ADR)

The right answer is to bring the actor in and record it again. In a perfect setup you put the scene on a monitor, hit play, and have them scream along with their on-screen self — over and over until it feels and looks right. Two things make this win. They’re matching their own mouth and body, so it syncs. And it’s a real performance, which means all the involuntary stuff comes with it for free — the breath, the ragged catch at the start, the panting on the way down. That’s the texture that sells a scream, and it’s exactly what you can’t fake.

(Record it clean, though — low, with headroom. The whole point is to not recreate the problem you’re fixing.)

Second best: a library scream

When you can’t get the actor back, you download one. It works, but now you’re rebuilding by hand what ADR gave you for nothing:

• You have to fake the roll-off — the shape of the attack at the front and the decay at the tail — so the scream rises and falls with the mouth opening and closing on screen. A library clip starts and stops on its own terms, not your actor’s.
• You lose the natural human sounds — the breath, the pant, the ragged edge. The best you can do is layer those back in from separate breath and effort elements, then run the whole thing through Place It to seat it in the room. It’s assembly work, and honestly it never quite equals a real take.

So it never happens again

Most of this pain is a capture problem, and capture problems are preventable:

• Record in 32-bit float if you can. The gear that does it — Zoom and Tascam handhelds, even phone recorders now — is cheap, and 32-bit float essentially can’t destroy a peak; you just pull it back down in post. The 16-bit ceiling is becoming a choice.
• Run a padded safety track. A second channel knocked down 12–18 dB. The main clips, the safety survives. This one even works on 16-bit, and it’s free if your recorder or camera does dual channels.
• Set your gain for the loudest moment, not the dialogue. If a scream is coming, record everything embarrassingly low. You can always bring a quiet scream up. You almost never bring a clipped one back.

Quick reference

What not to worry about yet

You can happily ignore a whole pile of “real” audio-engineering worries while you’re still learning to make things belong:

• Surround and Atmos. Stereo is plenty. Get placement right first.
• Delivery loudness specs and LUFS targets. If you’re not handing off to broadcast, balancing by ear in context is what matters.
• Expensive sound libraries. The free tiers plus your phone will take you further than you’d think.
• A field recorder — beyond a cheap 32-bit one, anyway. Squeeze your phone dry first. Buy the gear when the phone is genuinely what’s holding you back.

Get the space right, get the level right for the moment it lives in, and clean up only what’s actually dirty. Do that and the sound stops being something you added — it just becomes part of the scene. Which is really the whole job: not making people hear the sound, but making them believe the picture.

This is a living guide — I’ll keep adding scenarios as I run into them. If there’s a sound you’re fighting with, tell me, and it might end up here.