How Pros Mix Vocals: My Full Vocal Chain

Pro vocal chains all come down to the same basics: tuning, corrective EQ, compression, de-essing, tonal EQ and saturation, with reverb and delay on sends. Every expensive vocal plugin and one-knob fixer does exactly these jobs under the hood. Below is my full vocal chain, the one I use on my releases for Sony & Universal, with every setting.

So, you've probably been told that the next vocal plugin will finally make your vocals sound like the radio. Stop paying for promising vocal plugins and one-knob fixers. Once you understand what these plugins actually do under the hood, you'll see it always comes down to the same basics. Here's how the pros do it, and how I mix mine.

Quick intro if we haven't met: I'm Laurentius, a music producer, songwriter and DJ from Zwolle. I've been releasing music since 2014 for major labels including Sony & Universal, and independents like Paraíso, Soave and LoudKult. This is the vocal chain I use on those releases. Here's more about my background if you want to get to know my work!

Contents

• How do pros mix vocals?
• Vocal chain order: the eight jobs under the hood
• How to mix vocals: dry first, effects last
• Laurentius's perfect lead vocal chain
• Laurentius's backing vocal chain
• Laurentius's acapella bus chain
• Vocal FX sends: my reverb and delay settings
• Ready to make your vocals sound like a record?
• Frequently asked questions

How do pros mix vocals?

Pros build their vocal chain from the same moves: they tune early, clean up the frequencies, control the level with compression, de-ess, add air and colour, and put reverb and delay on sends that they automate. The exact order differs per engineer, but you'll find these moves again and again in published pro vocal chains. A few examples from Sound On Sound's Inside Track interviews and Universal Audio:

• Tuning first: Tom Barnes of TMS put Auto-Tune as the first plugin on every vocal track of Lewis Capaldi's Someone You Loved.
• Two compressors, each doing a little: Universal Audio calls the 1176 and LA-2A combination "a classic for a reason". Many Grammy-winning producers like Robert Orton (Lady Gaga, Just Dance) use two different compressors in a row the same way.
• De-essing after compression: Robert Orton puts his de-esser towards the end of the chain, because compression and EQ bring the S's up. Jonathan Low (Taylor Swift, willow) also runs FabFilter Pro-DS after his 1176.
• Air around 12 kHz: Tom Elmhirst (Amy Winehouse, Rehab) added air with a Pultec around 12 kHz, and Ian Kirkpatrick (Dua Lipa, Don't Start Now) boosted 12 kHz for air on the vocal too.
• Filtered effect returns: James F Reynolds (BTS, Fake Love) cuts the low end and high end on his reverb send.
• Automated sends: Robert Orton automates his delay sends, and Chris Lord-Alge (My Chemical Romance) automated his long delay throws.

My chains below use every one of these moves, and where I go against the textbook, I'll tell you why.

Vocal chain order: the eight jobs under the hood

Strip the marketing off any vocal plugin and you're left with these eight jobs, in the order I use them:

1. Cleaning: the noise between your phrases.
2. Tuning: your notes where they belong.
3. Corrective EQ: the mud, the boom and the harsh spot.
4. Levelling compression: every phrase at the same volume.
5. Character compression: peak control that thickens the vocal.
6. De-essing: the S and T sounds back down.
7. Tonal EQ: air and presence, on purpose.
8. Saturation: harmonics and glue, usually tape or tube.

Does the order matter? Yes, it does. Cut the mud before you compress, otherwise your compressor reacts to the mud. Add your air after de-essing, otherwise you're boosting the S's you just removed. A one-knob plugin does five of these jobs at once and hides the controls from you, and Logic Pro has a stock plugin for all eight.

How to mix vocals: dry first, effects last

My approach is actually very simple. I always mix my leads and backings completely dry, until they cut through the mix on their own. Only then I add the reverbs and delays, extra vocal effects, and only then I glue everything together on the acapella bus. If your vocal only works when the reverb is on, the reverb is hiding a problem you haven't fixed yet.

My vocal template is one parent group called Acapella, with three groups inside it:

• Leads Dry: all your lead vocal tracks. The lead chain sits on this bus.
• Backings Dry: all your dubs, harmonies and adlibs, with their own chain.
• Vocal FX Sends: your reverb and delay returns, 100% wet.

Your whole vocal, effects included, ends up on one bus, so when a label asks for the acapella, it's all in one stems bounce. Sending your vocals to a mixing engineer instead? Here's how to export stems in Logic Pro.

Before any of these chains, three things happen on your raw vocal tracks:

• Gain staging: The settings in this blog are my starting points for a lead vocal. I make sure a raw vocal averages around -18 dBFS, with peaks no higher than -12 to -10 dBFS. If you record hotter than that, you'll sacrifice a clean recording.
• Noise gate: Use Logic's stock Noise Gate and pull the Threshold up until your background noise is gone. If it gates too much and cuts off breaths or word endings, back it off and volume automate the noise away instead.
• Pitch repair: If the performance is really off, fix it note by note in Flex Pitch, Vovious or Melodyne before you process anything. The tuner on your bus is exclusively for polish, not for rescuing a bad vocal performance.

Laurentius's perfect lead vocal chain

This is the vocal chain on my leads bus, all from top to bottom, just like in your Logic Pro mixer:

These are starting points. Always trust your ears.

Tuning: A speed of 15ms to 25ms holds a pop lead in place, but still keeps the natural slides between notes. Always match your output level to your input, so tuning never changes your volume. I use UADx Topline Vocal Tune because it sets its Speed in milliseconds, and its natural modes leave the small pitch movements alone. Stock Logic alternative: Pitch Correction. Set Root Note and Scale/Chord to your song, and use Response to set how fast it corrects. Any Auto-Tune works here too.

Corrective EQ: On a deeper voice, bring your high pass down until the chest starts to sound thin, then back off a little. To find the harsh spot, sweep a narrow boost between 3.5 and 4.5 kHz on your loudest chorus phrase, and cut where it hurts. Both cuts are dynamic, so they only work when that area jumps out and your quiet lines keep their body. I use Pro-Q 4 for its dynamic bands, which is the one thing Logic doesn't have. Stock Logic alternative: Channel EQ with static cuts, and a bit gentler: around 1 dB on the mud and 1.5 dB on the harshness.

Compression: I use two compressors doing two different jobs, the same way I explain in my post on Logic's stock compressor. First, the Vintage Opto (Logic's LA-2A style mode) is my 'vocal leveller'. Aim for 2 to 4 dB of gain reduction on the loudest phrases, and no reduction on the quiet parts. Then, the Studio FET (Logic's 1176 style mode) thickens the vocal and catches the peaks the Opto left behind. My parallel trick: set its Mix to 75 percent, so the natural dynamics of the singer stay underneath the compressed version. Turn Auto Gain off on both and set your make-up gain by hand, so you compare tone and not loudness.

Important: most engineers actually run these the other way around, with the 1176 first and the LA-2A second. Universal Audio and Sound On Sound both call that the classic vocal chain. I level first, so my FET only has to thicken. And I'm not the only one: Steve Greenwell (Joss Stone) runs an LA-2A before an 1176 as well.

De-essing: Watch your reduction meter on the loudest S in your chorus. You want 3 to 6 dB there, and nothing on normal words. If it ducks every word, your frequency is set too low. I use Pro-DS because it shows you exactly which band it's grabbing. Stock Logic alternative: DeEsser 2, with Range on Split, the same frequency and Max Reduction at 3 to 6 dB.

Air and presence: The air boost uses a wide bandwidth, because you want a lift and not a spike. The presence boost at 5 kHz is narrower, and that's what makes your words readable in a busy chorus. Judge both on calibrated speakers or headphones. I check mine on my DT 990s with SoundID Reference. If the air turns brittle, it's your de-esser that needs fixing, not your EQ. I use the Pultec because it's the classic air EQ on pro vocals. Stock Logic alternative: one Vintage Tube EQ covers both. Use the top unit at 10 or 12 kHz for air, and the bottom unit's High Freq at 5 kHz with a little High Peak for presence.

Tape: Drive the input until your VU meter sits at 0, then turn the output down by the same amount. Louder always sounds better, so if you don't match the level, you'll keep the tape for the wrong reason. Cal +6 is also the calibration level UA recommends for 456 tape. I use the Studer because you can choose tape speed, tape formula and EQ standard, and every combination sounds different. Stock Logic alternative: ChromaGlow on Magnetic, Apple's tape model, on Apple silicon Macs.

Important: UA's own manual suggests the Studer as the first insert. I use it last, as the final colour on a vocal that's already under control, just like Jonathan Low ends his vocal chain on Taylor Swift's willow with tape-style saturation.

Want me to mix your vocals? That's one of the things I do for artists and labels from my music studio in Zwolle. Check out my mixing services.

Laurentius's backing vocal chain

So, how do you mix backing vocals? Your backings have one job: making the lead bigger, without competing with it. That's why this chain is harder and darker than the lead chain. It sits on my Backings Dry bus, and covers all dubs, harmonies and adlibs.

These are starting points. Always trust your ears.

Timing first: Before any plugin, your doubles and harmonies need to hit exactly with the lead. Here's one paid plugin I do highly recommend: VocAlign. It lines your backings up with the lead automatically, and makes them SO much cleaner. I use VocAlign 6 Pro. In Logic, you can do the same by hand with Flex Time, it just takes a lot longer! Heads up: according to Synchro Arts, VocAlign's ARA mode needs Logic running in Rosetta on Apple silicon Macs.

Tuning: Set your speed faster than on the lead, 0 to 10ms. Tight tuning is what makes a stack sound like one instrument, instead of five different people. Stock Logic alternative: Pitch Correction with a faster Response.

EQ: The high pass sits higher here, because your lead already has the body covered. The 2 to 3 dB cut between 2 and 3 kHz is the carve: that's exactly where your lead lives, so taking it out of the backings gives your lead space without turning anything up. The shelf above 10 kHz keeps the air on the lead. Pan your dubs 100% left and right. Stock Logic alternative: Channel EQ, with the same three moves.

Compression: The Classic VCA (Logic's dbx 160 style mode) works harder here, with a constant 4 to 6 dB and no parallel compression, so Mix stays at 100 percent. You're not after dynamics here, you want one steady bed behind your lead.

De-essing: Use a lot more than on the lead. Five stacked takes means five S's landing at the same time, and they add up quickly. Stock Logic alternative: DeEsser 2, with Range on Wide or Split.

Tape: 7.5 IPS is a slower tape speed than on the lead, and slower tape gives you less top end. That pushes your stack behind the lead, exactly where it belongs. Stock Logic alternative: ChromaGlow on Magnetic.

Laurentius's acapella bus chain

This is my vocal bus, which I call the acapella bus. Everything lands here: leads, backings and the wet effect returns. This bus is glue, so every move you make here is small. If you need a big move on this bus, the fix belongs further up your chain.

These are starting points. Always trust your ears.

Clean-up: Once all your vocals are summed, the rumble and mud of all those tracks add up, so this is where you catch it. Stock Logic alternative: Channel EQ.

Glue: The Vintage VCA (Logic's SSL style mode) at 2:1, with a 30ms attack and Auto Release on. Watch your meter: 1 to 2 dB of gain reduction, never more. If it goes past 2 dB, you're fixing a balance problem with a compressor, and that should be done with the faders on your groups.

Belts: Between 2.5 and 4 kHz is where a belted chorus note gets sharp. A dynamic band only pulls down those notes and leaves the rest of the performance untouched. I use Pro-MB because it lets you put a narrow band exactly on the problem. Stock Logic alternative: Multipressor. Set the crossovers so one band covers 2.5 to 4 kHz, and only compress that band. It's less precise, but it does the job.

Final de-essing, air and tape: The de-esser catches the sibilance that only shows up once your reverbs and delays are in the mix. One broad air boost over everything makes your whole vocal sound like one instrument. Then the lightest tape setting of all three chains. Stock Logic alternatives: DeEsser 2, Vintage Tube EQ and ChromaGlow, all used gently.

Want to hear where all of this ends up? Listen to my releases.

Vocal FX sends: my reverb and delay settings

This is the final stage, after your dry vocals already cut through the mix. All my sends live in the Vocal FX Sends group, 100% wet, and none of them are compressed. I dial in how much reverb and delay every vocal gets with the send knob, so there's no need.

Slap delay

The slap delay adds thickness, it's not an echo you should notice. Slapback started on tape machines, and Sam Phillips used it at Sun Records. Soundtoys puts slapback between 100 and 200ms.

I start from the 15 ips Studio Slap Wide preset in EchoBoy Jr. by Soundtoys: Mode on Wide, time mode on Time at 126.2 ms, style on Studio Tape and Glide off. The Mix is almost fully wet, and Feedback and Saturation both sit just above minimum. Low Cut goes just before 12 o'clock and High Cut to around 2 o'clock, with Input and Output at 12 o'clock. If you can hear the slap as a separate repeat, it's too loud or too long.

Stock Logic alternative: Tape Delay with Tempo Sync off, Delay Time around 125 ms, Feedback at its lowest, Low Cut around 200 Hz and High Cut around 8 kHz.

Long hall reverb

This reverb gives your record its size, and the pre-delay keeps your words readable while the tail stays big. My go-to is Crystalline by Baby Audio. I sync Start (the pre-delay) to 1/32 and End (the decay) to 1, with Shimmer off, the Ducker all the way down and Dry/Wet fully wet. At 120 BPM, a 1/32 pre-delay is about 62ms.

Free alternative: RAUM by Native Instruments, a free download through Native Access. Choose the Airy algorithm, sync the Predelay and set the Mix fully wet.

Stock Logic alternative: ChromaVerb on the Vocal Hall room type. Sync the Predelay to 1/32 or set it to 20 to 40ms, set the Decay to 2.5 to 4 seconds, Dry to 0 and Wet to 100%, because it's a send. In the Output EQ, add a high pass around 250 Hz and a low pass around 10 kHz.

Long delay

This is the delay people actually hear. I use EchoBoy Jr. again: Mode on Ping-Pong, time on Note at 1/8, style on Studio Tape and Glide on. The Mix is almost fully wet, Feedback sits around 10 o'clock and Saturation just past 12 o'clock. Low Cut goes to 12 o'clock and High Cut just past 12 o'clock, with Input and Output at 12 o'clock.

My pocket trick: look at the Low Cut and High Cut on both of my delays. These only filter the echoes, so the repeats lose their low end and top end, and sit in a narrow band behind your vocal instead of taking over. Works like magic!

Stock Logic alternative: Stereo Delay, with Routing on Ping Pong L/R, Tempo Sync on, 1/8 notes, Low Cut around 300 Hz and High Cut around 6 kHz.

Optional: the reverb build-up trick

You probably know this sound from songs like Stay by Zedd and Alessia Cara, or Lean On by Major Lazer: the last word before the drop swells into a wash of reverb, and then the drop hits. Set up a second reverb send with a long decay, fully wet, and high passed around 300 Hz. Then automate the send up over the last word or two before the drop, and cut the dry vocal right after, so only the tail carries into your drop. Pull the send back down on the downbeat.

These are starting points. Always trust your ears.

Ready to make your vocals sound like a record?

Now you have all three of my vocal chains, my sends, and the reason behind every step. My advice: build it with stock plugins first, and only buy a plugin once you know exactly which job you want to do better. Good luck with your mix!

Want me to mix your vocals, or the whole track? Check out my mixing services, or tell me about your song on my contact page. Working on a full record? Here's how working with me as your producer goes. And if your track is a banger, send it over anyway, because I'm always looking for new music to play in my live DJ sets.