How to Mix Vocals So They Sit Right in Your Song

“…if you make the vocal too loud, the track will sound small.”

Anyone learning how to mix vocals knows this moment: the beat sounds huge, the vocal sounds great, and together they sound like two songs playing at once. Turn the vocal up and the beat shrinks; turn it down and the words get lost. The order below fixes that in any DAW (the program you make music in). It comes from Max, the founder of Mix Elite, and his video on the six mistakes that destroy vocal mixes, with real settings and a screenshot for each move. If you mix in FL Studio, the full stock-plugin chain is in our guide to mixing vocals in FL Studio with stock plugins.

How to mix vocals step by step (the vocal chain order)

Mix a vocal in this order: even out its volume, tune it if it needs it, cut the rumble with EQ, compress it, tame the harsh S sounds with a de-esser, then add reverb and delay on their own tracks. Each step makes the next one easier, so don’t skip ahead to the fun part with the reverb.

1. Record it well, then cut out loud breaths, mouth clicks and noise between the lines.
2. Even out the volume, so every line reaches your effects at a similar level.
3. Tune it if it needs it, before the other effects.
4. EQ: cut the rumble and shape the body of the voice.
5. Compress until every word is equally clear.
6. Tame harsh S sounds with a de-esser.
7. Add reverb and delay on their own tracks, and make them duck under the voice.
8. Make the vocal sit in the mix by letting it turn the music down a little while it plays.

Mix Elite’s free 4.5-hour FL Studio course squeezes the core of this chain into one word, V.O.C.A.L.: volume automation, optimize with EQ, compression, attenuate harshness, and layer with effects. Attenuate just means turn down. The sections below follow that same order.

Why do my vocals sound like they’re sitting on top of the beat?

Vocals sound like they sit on top of the beat when the two never react to each other. Usually the vocal is a little too loud or too uneven, it fights the beat for the same middle frequencies, and it’s bone dry while the beat has space around it. Even it out, give it a reverb that ducks, and let it dip the music about 2 dB whenever it plays.

It’s probably the most common vocal mixing mistake of all. Everything sounds good on its own, but as Max puts it, “you just literally slap the vocal on top of the instrumental and called it a day.” Each cause below has its own fix further down the page:

• It’s too loud or too uneven. Loud lines jump out and quiet ones sink. Even out the volume first, then raise the vocal 1 dB at a time until it sits just above the beat.
• It fights the beat for the same frequencies. Chords, synths and guitars crowd the mids, right where the voice lives, so cut a little room for it in the beat.
• It’s dry while the beat has space. A dry vocal over a roomy beat sounds like it was recorded somewhere else. Give it a reverb on its own track that ducks while the singer sings.
• The beat never makes room for it. Let the vocal push the music down about 2 dB while it plays, with a sidechained compressor on the beat.

Start with a good recording

Gold in equals gold out. No plugin can add detail that the mic never caught, so the recording matters more than any trick further down this page.

For the full setup, from mic placement to input levels, see how to record vocals in FL Studio.

Every microphone colors your voice a little, almost like a built-in EQ. Condenser mics sound brighter than dynamic mics right out of the box, but that’s not the real problem, because you can always turn the highs up or down later. The real problem is distance. The closer you sing, the more low end the mic picks up, which is called the proximity effect. Stand too far away and the low end never gets recorded, so boosting it later only brings up mud and noise.

Before you hit record, sing your loudest line and lower the gain on your audio interface until its light never turns red. Record a few takes even when one feels perfect, and record the voice dry, with no effects, so you can still change everything later. Then cut out loud breaths, mouth clicks and the noise between lines, or keep a breath where it sounds natural. For gear, see our pick of the best microphones for singing and our vocal recording tips.

Even out the volume before you add any effect

This is the first of the six mistakes in the video above. When a singer moves closer to the mic, the vocal gets louder. When they lean back, it drops. Your compressor and every other effect react to how loud the vocal is, so an uneven vocal gets uneven treatment. The loud lines get crushed and the quiet ones barely get touched.

There are two ways to fix it. A vocal rider plugin rides the level for you, and Waves Vocal Rider is the most popular one. By hand, you draw a volume envelope, a line that raises the quiet parts and lowers the loud ones until the vocal looks even.

1. In FL Studio, click the vocal’s waveform and choose Edit sample to open it in Edison.
2. Switch to the volume envelope.
3. Right-click to add points, and drag them up where the singer gets quiet (1). In the free course, the quiet parts go up by about 3 dB at most.
4. Drag the result back into the playlist when it sounds even.
5. Edison opens on the Master track, so remove it from there once you’re done.

Set the input level (gain staging)

Mistake number four is close to this one: the vocal goes into your effects too quiet. Compressors and saturation (a plugin that adds warm distortion) behave differently depending on how loud the sound coming in is. Some saturation plugins act one way from 0 to 50% and another way from 50 to 100%. A quiet vocal makes you crank them. That can bring up noise.

So normalize the vocal, or simply turn it up, before it hits your plugins. That matters most when you use a preset, because the compressor in a preset needs a healthy level to work with. In FL Studio, open the clip’s channel settings and turn on Normalize (1), or raise the VOL knob (2).

Should you tune the vocal?

Only if it needs it, and before the rest of the chain. Gentle pitch correction pulls notes onto the right pitch and suits pop and EDM. The obvious robotic auto-tune sound is an effect, and it suits rap and live shows. A free tool is enough to start: the free course uses Graillon 2, with its smoothing set to 10 ms. At 1 ms it sounds robotic, and around 20 ms it sounds soft.

Our guide on how to tune vocals like a pro walks through it, and here are the best free autotune plugins and free pitch shifting plugins if you need one.

How to EQ vocals

EQ, short for equalizer, turns certain frequencies up or down. Low frequencies are the deep rumble in a sound, and high ones are its hiss and air. The second mistake in the video is putting the same EQ curve on every vocal. The usual curve cuts the lows and lifts the highs. Both moves make sense. The part that changes from voice to voice is the body in between.

Start with a low cut (1), also called a high-pass filter, which removes everything below a point. Keep it around 100 Hz. Leave the lows in and the voice sounds like a voiceover or a podcast. Push the cut too high and the main vocal goes thin, which is fine for backing vocals and ad-libs but not for your lead. Many producers also lift the highs a little (2), because pro vocals are rich in highs and most home recordings aren’t. If one spot sounds harsh, sweep a narrow band to find it and cut only 2 to 3 dB there. The de-esser handles the rest.

Listen for the body. A singer who was close to the mic has lots of it, and one who stood far away has very little. Hip-hop usually wants more of it, so the voice feels right in front of the listener. So don’t copy a curve. Cut the lows, then try a little more or a little less body with the beat playing. Sometimes the recording is already balanced and leaving it flat is right. And if an artist sends you a vocal from another studio, they may already have used an EQ while recording.

Match your vocal to pro vocals with a reference

Be careful with vocal EQ presets. A preset curve can suit one voice and ruin the next, so compare your vocal with a reference instead.

Mix Elite’s team built a noise file shaped like the average of many hit-song acapellas, which are vocals with no music. You compare your vocal against it in SPAN, a free analyzer plugin that draws a sound’s frequencies as a curve.

1. Put an EQ and then a compressor on the vocal first, because the reference was made from compressed vocals.
2. Load SPAN and show the reference as an Underlay (1).
3. Set the display to an average, and match the two levels.
4. EQ until your vocal’s curve (2) looks more like the reference. In the video’s example, that meant a low cut, a lot more highs, and a small dip around 600 to 800 Hz.

It’s a guide, not a rule. The reference is only an average, and you always EQ a vocal to fit your own beat. A vocal that sounds odd on its own can sound just right with the music.

Where do vocals sit in EQ?

Rumble sits below about 100 Hz, the body in the low mids, a boxy tone often around 600 to 800 Hz, harsh S sounds around 4 to 6 kHz, and the air at the very top. Every singer is a little different, so use this map as a starting point and let your ears decide:

• Below about 100 Hz: rumble and noise. Cut it.
• The low mids, roughly 100 to 300 Hz: the body. More sounds close and full, too much sounds muddy.
• Around 600 to 800 Hz: can sound boxy. In the video’s example it was a little too loud.
• About 4 to 6 kHz, sometimes up to 7 to 10 kHz: harsh S sounds. That’s the de-esser’s job.
• The top end: air and shine. Pro vocals have plenty of it.

How to compress vocals without crushing them

Compression turns the loudest parts of a sound down, so the loud and quiet parts sit closer together. The threshold is the level where the compressor starts working, and the ratio is how hard it pushes down. Mistake number three is overdoing it, and the ratio isn’t the real problem. The threshold is.

Max’s rule of thumb is to “set your threshold as low as the quietest word is in your lyrics.” Then every word stays easy to understand, and the loud accents stop stabbing your ears.

In FL Studio, use Fruity Limiter switched to COMP mode (4). Lower THRES (1) to that quietest word, then set the RATIO (2). Start soft, around 2:1, or 3:1 to 4:1, and only go higher if it sounds better. Mix Elite’s FL Studio walkthrough on compressing vocals goes up to 6:1 on a lead vocal, and our guide on how to compress vocals has those exact settings. Heavy compression is fine, as long as the threshold is set right.

Then check the attack (3), which is how fast the compressor grabs. Leaving it at 0 to 2 ms is fine most of the time. On rap vocals, test a slower attack of 30 to 40 ms, which lets the start of each word cut through the mix a little better. If you can’t hear a difference, just go back to fast.

Add warmth and glue with OTT (optional)

Once the vocal is under control, a little OTT, a free multiband compressor from Xfer Records, glues it together and adds a nice color. Keep its depth low, about 7 to 20% on a vocal. The free course uses 15%. It’s optional, so skip it if you can’t hear it helping. Our guide to OTT compression shows how it works.

How to tame harsh S sounds

Compression brings up the sharp S and SH sounds in a voice, which are called sibilance. A de-esser turns down only those sounds. They usually live around 4 to 6 kHz, and sometimes higher, up to about 7 to 10 kHz.

FL Studio has no plugin called a de-esser, but Maximus comes with de-esser presets. Open its preset menu and pick De-esser narrow-band or De-esser split-band (1). Then lower the threshold a step at a time until the S sounds calm down. Don’t overdo it. Too much dulls the vocal. Our post on 5 easy ways to treat vocals in FL Studio has more fixes like this.

How to add reverb to vocals without washing them out

Reverb puts a voice in a space, like a room or a hall, and delay adds echoes. Together they give a vocal depth, and they’re also the easiest part to get wrong. “Reverb usually is something beginner producers overuse since it basically blurs the imperfections,” Max says. The words lose their edges. The mix turns to mush.

For reverb settings by song style, step by step, see how to add reverb to vocals.

Start with the settings, whatever reverb you use. Turn the SIZE up (1), because very small sizes ring with a strange echo on most vocals. Then set the decay, how long the reverb lasts (DEC, 2). Keep it short for rap, so the words don’t smear into each other, and longer for slow vocals.

To time it to your song, use the free reverb and delay calculator on our site. In Mix Elite’s vocal reverb tutorial, a whole bar came out to about 1.8 seconds, so each reverb tail faded right before the next busy line started.

Next is pre-delay (DELAY, 3), a short gap before the reverb starts. A lot of people overlook it, but no pre-delay sounds unnatural, because in a real room the sound has to hit a wall before it comes back to you. Anything under 50 ms sounds natural. The tutorial tries 50 ms, then a tempo-synced 58 ms, and halves it to about 28 ms when the vocal starts to feel mushy.

Last comes the most important setting: the low cut (L.CUT, 4). It stops the reverb from muddying your low end. The free course sets it around 700 Hz, and in a dense mix you can push it as high as 1 kHz.

Put the reverb on its own track

The vocal (1) keeps its own chain, and the reverb lives on its own track (2). Turn the reverb fully wet, with no dry signal, because the dry vocal already plays on its own track. Now you can treat what goes into the reverb without touching the vocal: cut the lows, cut the highs where the S sounds live, and dip the muddy low mids, all before the reverb.

Make the reverb duck while the singer sings

Now for the move that makes it sound pro. It’s called sidechained reverb, and few moves matter more for a vocal that sounds finished. Sidechaining means one sound controls another. Here, the dry vocal turns the reverb down while the singer sings, and the reverb opens up in the gaps between lines.

1. On the reverb track, add Fruity Limiter after the reverb and switch it to COMP (1).
2. Right-click SIDECHAIN (2) and pick the vocal, so the vocal can control it.
3. Lower THRES (3).
4. Set the RATIO (4) somewhere between 4:1 and 8:1.

Go easy here. It should feel better, not sound obviously different. If you can hear the reverb pumping, raise the threshold or lower the ratio, and if it cuts in too hard between words, turn up the release (REL, 5). Some people use a dynamic EQ for this job, but Max prefers plain compression here, because to his ears it sounds a bit better.

Add a delay the same way

Delay works best on its own track too. Send the vocal to a new mixer track, put the delay there, and take the dry signal out, so you only hear the echoes. An EQ before the delay helps in the same way it does on the reverb.

In Fruity Delay 3, set the TIME (1) with TEMPO SYNC on (2), so the echoes land on the beat, for example on quarter notes, eighth notes or triplets. Ping pong mode (3), with the delay panned to one side, makes the echoes bounce left and right. Then route both the reverb and the delay through one more track with the sidechained limiter, so they both duck under the vocal. For the difference between the two effects, read reverb vs delay.

How do you get vocals to sit in the mix?

Make the vocal and the beat react to each other. Even out and compress the vocal, give it a reverb that ducks, then let the vocal push the music down about 2 dB whenever it plays, with a sidechained compressor or a dynamic EQ on the instruments, not the drums. The beat makes room, and the voice sits inside the song.

This is the last of the six mistakes in the video, and the fix is sidechaining again, this time on the instruments. If you have the beat as separate tracks, leave the drums out of it. With a finished two-track beat, you duck the whole thing, as in this short video:

First, send a copy of the processed vocal to its own mixer track and name it Signal. It only drives the sidechain, so it shouldn’t play through your speakers a second time.

The simple version: a compressor on the beat

1. Put Fruity Limiter on your instrument track or bus and switch it to COMP (1).
2. Pick the vocal copy as its SIDECHAIN input (2).
3. Lower THRES (3) until the music dips a little every time the vocal comes in. In Mix Elite’s stock-plugin vocal video, about 2 dB is plenty.

The full routing, click by click, is step 9 of our FL Studio vocal mixing guide.

What EQ cuts make room for vocals in a beat?

Cut the instruments, not the vocal. Find where your vocal is busiest, usually in the mids where its body sits, and turn that same range down a little on the chords, synths or guitars that fight it, never on the drums. Keep the cuts gentle, and use a dynamic EQ if you only want the dip while the vocal plays.

Two sounds fighting for the same frequencies is called frequency masking, and the louder one hides the other. Mix Elite’s frequency masking tutorial finds the spot to cut in four steps:

1. Put an EQ on the vocal, boost one narrow band and sweep it through the mids to hear where the voice is strongest.
2. Save that EQ as a preset and drag it onto the instrument that fights the vocal.
3. Turn those bands down on the instrument instead of up, and keep the cuts gentle so the beat doesn’t sound hollow.
4. Listen with the vocal playing, then compare it with the cut turned off.

A plain EQ cut stays on even when nobody sings, which can leave the beat dull in the gaps. A dynamic EQ with a sidechain input fixes that. In the free TDR Nova, set a band where the vocal is busiest, pick the vocal as the sidechain, and lower the threshold until the band dips only while the singer sings. The next section shows a ready-made version of the same idea.

The better version: duck only the vocal’s frequencies

A plain compressor turns the whole beat down, bass and all. Ideally, only the part of the music that fights the voice should move. The video above does this with a multiband sidechain preset in Patcher, FL Studio’s tool for building your own effect chains.

Low Cut (1) and High Cut (2) set the range of frequencies the preset works on. Sidechain (3) sets how deep the music ducks, and Release (4) sets how fast it comes back. As the note on screen in the video says, a trick like this doesn’t make an obvious difference at first, but it makes a world of difference to the mix.

If you’d rather use a ready-made plugin, the most common one for this job is Trackspacer. Sidechain it from the vocal and set it somewhere between 8 and 20%.

How loud should vocals be in a mix?

Loud enough that every word is easy to understand, without the voice jumping out at you. There’s no single number, because it depends on the song. Raise the vocal a little at a time, about 1 dB per step, until it sits slightly above the beat, never in big jumps.

Always judge it with the whole song playing. Listeners never hear your vocal on its own, so a vocal that sounds perfect in solo can still be buried, or too loud, in the mix. A reference track helps too. Pick a song you love in the same style, flip between it and your mix, and compare how loud the voice sits.

Background vocals, doubles and ad-libs

Everything above is for the lead vocal, and the extra layers need less. Backing vocals and ad-libs don’t need the body a lead needs, so you can low cut them higher, which keeps them out of the lead’s way. Keep the lead in the middle of the stereo field, so it stays the focus.

For the details, read how to mix background vocals, why double tracking makes recordings fuller and our guide to mixing hip-hop, rap and R&B vocals.

More vocal mixing guides

Every guide in our vocals series, grouped by topic:

• Mixing: the full stock-plugin vocal chain in FL Studio, how to compress vocals, ways to treat vocals in FL Studio, hip-hop, rap and R&B vocals, background vocals and double tracking.
• Recording and gear: recording vocals like a pro and microphones for singers.
• Tuning and plugins: tuning vocals, free autotune plugins, free vocal plugins, Waves plugins for vocals, pitch shifting plugins and vocoder plugins.

Start with the order on this page, finish one song with it, and then go deeper on whichever step gave you the most trouble.

Frequently asked questions