Full Transcript

·YouTLDR

4 Techniques That Will Change Your Outlook On Mixing

37:245,921 words · ~30 min readEnglishTranscribed Apr 23, 2026
AI Summary

The secret to professional mixing is moving beyond technical tool knowledge to understanding how human brains interpret sound. By treating compressors as distortion boxes and reverb as an EQ tool, you can shape character and space more effectively than by simply turning knobs.

Most producers struggle because they lack a mental framework for translating what they hear into specific technical adjustments, leading to 'knob-twiddling' fatigue.

Section summaries

0:00-1:00

Introduction

optional

General overview of why mixing is hard; skip if you want to get straight to the techniques.

1:00-5:00

EQ & Vowel Sounds

watch

Essential framework for identifying frequencies without the 'sweep and destroy' method.

5:00-21:00

Compression as Distortion

watch

Deep dive into how attack/release times create harmonic character; includes a kick drum demo.

21:00-24:00

The Compression 'Number One Trick'

watch

Explains why volume automation should almost always precede compression.

24:00-31:00

Reverb as Tonal EQ

watch

A unique perspective on using reverb to thicken instruments like acoustic guitars and snares.

31:00-35:00

The 3 Delay Windows

watch

Technical breakdown of how delay times translate into human spatial perception.

35:00-37:00

Course Outro

skip

Promotion for the Audio University membership and community.

Key points

  • EQ Vowel Reference Points — Every frequency range corresponds to a specific human vowel sound (e.g., 250Hz is 'oo', 1kHz is 'ah', 4kHz is 'ee'). This uses the mouth's natural resonance as a built-in ear training tool.
  • Compressors as Distortion Devices — Compressors don't just control volume; they reshape waveforms into square-like waves, which generates harmonic distortion and saturation.
  • Reverb as a Tonal/EQ Tool — Instead of using reverb just for space, you can use it to reinforce specific harmonics of an instrument (like adding 'thickness' to a snare) through parallel processing and EQ'ing the reverb return.
  • The Three Windows of Delay — Delay is categorized into three functional zones: Binaural (1-10ms for localization), Early Reflections (15-35ms for room size), and Late Reflections (35-60ms for space character).
Compressors make square waves. Gabe Herman
Compressors are not intelligent. You are intelligent. Gabe Herman

AI-generated from the transcript. May contain errors.

Language
0:00

Throughout the process of learning how

0:01

to mix music, many people hit a wall

0:03

where they know something's wrong, but

0:06

they can't quite [music] name it.

0:07

Whether the kick is fighting the bass or

0:09

the vocal is just sitting on top of

0:10

everything instead of being inside the

0:12

mix as a cohesive piece. Whatever it is,

0:15

you end up turning the same knobs back

0:17

and forth for about an hour, and somehow

0:19

[music] you end up worse than where you

0:21

started. If you've done that, leave a

0:23

comment below. I definitely have done

0:26

that, and I'm sure we're not alone. It's

0:28

not that you're doing something wrong.

0:30

What's missing [music] is the connection

0:32

between what you're hearing and what you

0:34

can actually do about it using the tools

0:36

available.

0:38

That connection

0:39

>> [music]

0:39

>> is what a good education gives you, but

0:42

it's harder to find good education

0:43

>> [music]

0:44

>> than it should be. A lot of mixing

0:45

videos on YouTube, my own included, tend

0:48

to teach you how tools work,

0:50

>> [music]

0:50

>> which is helpful information, but it's

0:53

less common to find videos that tell

0:55

[music] you how to translate what you're

0:56

hearing into adjustments on those tools.

1:00

Today, I want to walk you through four

1:02

concepts, one for each of the primary

1:05

tools you're probably already using: EQ,

1:07

compression, reverb, and delay. Each one

1:10

is a specific idea that gave me a

1:13

framework for when to choose the tool

1:15

and what to listen for once the tool

1:17

[music] is in the signal chain. These

1:19

come from the mixing and ear training

1:21

program at Audio University, which is

1:23

taught by myself and [music] Gabe

1:25

Herman, who's been teaching audio

1:27

engineering and mixing professionally

1:29

for over 20 years. I'm going to pull

1:31

some clips from the course so you can

1:33

hear Gabe explain these ideas directly,

1:36

>> [music]

1:36

>> and if any of this lands for you, the

1:38

program is linked in the show notes

1:40

below.

1:44

Let's start with EQ because I think it's

1:46

where most people spend most of their

1:48

time, and it's also where most people

1:50

are working much harder than they really

1:52

need to. The typical approach goes

1:54

something like this. You hear something

1:56

off in the mix, you open an EQ plugin,

1:59

you sweep a boost around until you find

2:01

a frequency that sounds bad, and then

2:04

you cut it. And the logic here makes

2:05

sense, and it is really a good tool and

2:09

a good technique that I still use at

2:11

times.

2:12

The problem is that it can lead to more

2:14

problems than it solves. If you apply a

2:17

big boost while you sweep through the

2:18

frequency spectrum, you can make almost

2:21

anything sound terrible at almost any

2:24

frequency, which means you end up

2:26

cutting things that weren't actually

2:28

problems in the first place. There's a

2:30

faster way to get your bearings with EQ

2:32

that one of my teachers taught me.

2:34

Certain frequencies correspond to vowel

2:36

sounds. Your mouth is a resonant cavity,

2:39

and when you shape it to say a vowel

2:41

sound, you're resonating at a specific

2:43

frequency. Once you make that connection

2:45

between what frequency corresponds to

2:48

what vowel sound, you have a reference

2:50

point every time you open an EQ.

2:53

Around 250 hertz, you get the "oo" sound

2:56

like the word boot.

3:02

Move up to 500 hertz, and it shifts to

3:05

an "o" sound like boat.

3:10

At 1 kilohertz, you're in "ah" territory

3:14

like the word father.

3:20

2 kilohertz is more like "a" as in face.

3:27

And 4 kilohertz is an "e" sound like the

3:30

word meet.

3:36

Above [snorts] 4 kilohertz, you start to

3:38

get into the sibilance range. 8

3:40

kilohertz gives you a clean "s" sound.

3:43

Higher than that, it gets sharper,

3:45

almost like an "s" mixed with a "t" at

3:48

16 kilohertz.

3:49

Below 250 hertz, you can focus on the

3:52

physical sensation of the sound in the

3:54

body. 125 hertz, you tend to feel in

3:57

your chest, and 63 hertz, you feel

4:00

lower, especially when you're listening

4:02

through full-range speakers or a system

4:04

with subwoofers.

4:06

What this gives you is a way to connect

4:08

what you're hearing to where you need to

4:10

go on an EQ. A vocal that sounds boxy

4:13

might have too much energy around 500

4:16

hertz or 1 kilohertz, somewhere between

4:19

"o" and "ah".

4:21

A mix that sounds harsh usually is a

4:24

buildup of energy sitting between 2

4:26

kilohertz and 4 kilohertz, between "a"

4:30

and "e".

4:31

I'd encourage you to download the free

4:33

ear training guide linked below this

4:35

video for a quick reference for each of

4:37

these reference points I've just shared

4:39

with you. You can also download it at

4:40

audio

4:41

universityonline.com/eartrainingguide.

4:44

This was a huge revelation for me when I

4:46

learned it, and even if many engineers

4:49

don't use this exact method, they

4:51

probably have some tricks of their own,

4:53

reference points for what they're

4:55

hearing.

4:56

Here's Gabe describing how he approaches

4:58

an electric guitar in one of the mix

5:01

sessions within our course. This guitar

5:03

really speaks from kind of the back of

5:06

the mouth. [music]

5:08

It's got like an

5:11

kind of [music]

5:12

vibe to it, and I use my mouth and I use

5:15

the sounds in my [music] mouth to EQ all

5:17

the time because it helps to kind of

5:19

place things in my body, and then I can

5:22

internalize them better.

5:23

>> That instinct, translating what you hear

5:26

into something your body already knows

5:27

and understands, is a foundational

5:30

element of ear training that actually

5:31

sticks with you. The key is repetition.

5:34

You hear it, identify it, apply a

5:36

change, and check the result. And over

5:39

time, you stop guessing and start using

5:41

EQ with a goal in mind.

5:44

Okay, compression. Most people learn to

5:46

think of a compressor as a dynamics

5:48

tool. It catches loud moments and pulls

5:51

them back, and that's accurate, but it

5:54

misses something else that's important

5:56

about what compressors actually do to

5:58

the sound. And after watching Gabe teach

6:01

this, it really started to click for me.

6:04

Here's how Gabe puts it in the course.

6:06

Compressors make square waves.

6:08

They are uh

6:10

really not generating the the effect

6:12

that most people associate with dynamic

6:14

range control. While they do control

6:16

dynamics, it's the interaction of their

6:19

dynamic range control

6:21

um devices and and components that

6:25

generate um harmonic saturation and

6:27

harmonic distortion

6:29

in a way that is different than just

6:31

dynamic range control. And so we need to

6:33

be able to think about dynamic range

6:35

control separate from harmonic

6:36

distortion, and you might get further

6:39

learning how to use a compressor not

6:40

just by hearing how the dynamics of the

6:42

signal change under compression, but at

6:44

the same time learning to hear what the

6:47

harmonic qualities are of too much

6:49

compression

6:50

um or poorly timed compression, etc.

6:54

You should focus on the intent. Why do

6:56

we use a compressor? What changes do we

6:59

want to see in our sound? What should

7:01

stay the same?

7:03

Think of distressers uh uh think of

7:05

compressors as distortion boxes and pair

7:07

them with complementary source material.

7:09

So, if you have um a uh

7:13

a very fast transient in a snare drum

7:15

signal, say you've recorded your snare

7:17

with an API uh 512 mic pre with an op

7:21

op-amp circuit, um those do a great job

7:24

of preserving transients. It may be ad-

7:27

uh advantageous for you to uh then use a

7:30

compressor paired with that signal

7:31

that's able to handle very fast

7:33

transients.

7:34

Um so, think about the um the the tool

7:37

as being um the way that tool is

7:40

designed to be applied to a specific

7:42

type of sound, and that might help you

7:44

develop uh a better strategy, and we

7:46

will discuss strategies as another part

7:48

in the series.

7:50

Lastly, I want you to remember, and this

7:52

is the most important thing, and you'll

7:53

maybe get tired of me hammering on them

7:55

so hard throughout this, but compressors

7:57

are not intelligent. You are

8:00

intelligent. Compressors as distortion

8:03

devices, not just as dynamic

8:05

controllers. Every time a compressor

8:07

acts on a signal, it's not just turning

8:09

down the loud parts, it's changing the

8:12

shape of the waveform, and that change

8:14

produces harmonic distortion. The

8:16

character of that distortion comes

8:18

directly from your attack and release

8:21

settings. A fast attack means the

8:23

compressor clamps down quickly and

8:25

starts reshaping the waveform almost

8:27

immediately. And a fast release means it

8:30

recovers very quickly. The more

8:31

aggressively the waveform gets reshaped,

8:34

the more harmonic distortion is

8:36

produced, and the more the character of

8:38

the sound changes. Slower settings mean

8:41

less reshaping, less distortion, and a

8:43

cleaner sound overall. That's a big part

8:46

of why two compressors with different

8:48

time constants can sound completely

8:50

different at the same ratio and

8:52

threshold. The timing settings aren't

8:55

just controlling dynamics, they're also

8:57

controlling character.

8:59

Here's Gabe working on a kick drum using

9:01

compression specifically to shape the

9:04

tone.

9:05

I want to talk about setting attack and

9:07

release times and how I might use a

9:09

compressor to shape the tone of a kick

9:10

drum.

9:11

In this session, I've got just one track

9:14

we're going to be listening to, which is

9:15

the kick in mic. Let's give it a listen.

9:28

In our previous session on EQ, I

9:30

discussed how I used this frequency

9:32

curve to enhance the sound of the kick

9:34

drum. I'm going to bypass that so we can

9:36

hear the difference of the signal

9:38

without the EQ versus with the EQ.

9:42

Without it.

9:47

With it.

9:52

What I'd like to point out is that we

9:54

used an EQ in in previous session to

9:56

enhance the low frequency content and to

9:58

make the uh articulation of the kick

10:00

drum a little stronger while reducing

10:02

some of the frequencies that were

10:03

shadowing the upper mid-range um closer

10:07

to around 200 hertz. We call this the

10:09

mud range.

10:11

Um today when we install a compressor

10:14

behind this signal we're going to be

10:16

changing the envelope and introducing

10:18

harmonic distortion to that signal.

10:21

When I think about what I want to change

10:23

about this signal, I have to be very

10:24

careful.

10:26

If I were looking at the frequency

10:27

responses drawn on the waveforms of our

10:30

edit window, you'll see very fast

10:32

transients followed by some pretty quiet

10:35

resonances.

10:37

Uh of course, this is what the waveform

10:38

looks like before the EQ, not after.

10:42

If I wanted to do that, if I wanted to

10:44

see what it looks like, I would have to

10:46

commit the audio up to this level

10:49

which I'm going to do right now.

11:01

As I do this, what I'm going to see is a

11:04

change in the waveform that will

11:05

actually represent what's different now

11:08

that I have

11:10

um made the uh EQ rendered.

11:14

What we see now is that before the body

11:17

of the kick drum, which was a little bit

11:19

weak, has been made a little bit

11:20

stronger.

11:21

But we still have these very strong

11:23

transients

11:24

followed by some relatively quiet

11:27

resonance of the kick drum.

11:29

What I'd like to do is use a compressor

11:31

to tame the transients and make these

11:34

sections here after the transient a

11:36

little bit louder.

11:38

If I use my cursor tool to select just

11:40

this amount of time that is my transient

11:42

for this kick drum hit I can look up and

11:45

see that it's about 1 millisecond of

11:47

time here from start to end.

11:50

A difference of .096

11:52

uh milliseconds.

11:55

Or a tenth of a millisecond I should

11:57

say.

11:58

So when I set up my compressor, what I

12:00

want to be very careful of

12:02

is that I'm not making the time too

12:04

long. If I make the attack time too long

12:07

I'll uh end up setting up a compressor

12:09

that's too slow to be able to react to

12:11

the signal.

12:12

And that means I'll end up compressing

12:14

what's happening on the back end of my

12:16

signal, this later part which is the

12:18

body

12:19

and the resonance in the decay of the

12:21

tone of the of the envelope of the

12:22

signal.

12:24

That wouldn't be good at all because

12:26

what I really want to do is squash the

12:27

transient, not the body. So I need to

12:30

keep my attack times pretty quick.

12:33

Lucky for me this stock compressor that

12:34

comes with Pro Tools is very quick. I'm

12:37

going to set the attack time to 10

12:38

picoseconds. That's way fast enough to

12:41

be able to uh uh

12:42

uh get out in front of this transient.

12:46

At the same time I want to make sure

12:47

that my release time is also very fast

12:49

because I want the compressor to be able

12:51

to stop compressing relatively quick.

12:54

I need to reduce the volume here. If I

12:57

go with something strong like 6 to 1

12:59

compression that's going to create many

13:02

more harmonic uh

13:04

uh uh uh

13:05

auto order harmonic uh elements in my

13:07

signal than I think I really want.

13:10

I like the way this sounds right now.

13:11

It's warm, it's punchy, it's thick. If I

13:14

start introducing too many auto order

13:15

harmonics to the signal, it's going to

13:17

change the timbre of the signal to be

13:20

more crunchy, more gritty, maybe a

13:22

little bit more aggressive, which is

13:23

fine for some music, not in this case.

13:26

So I'm going to set my attack time to be

13:28

very fast, my release time to be very

13:30

very fast. I'm going to set my ratio to

13:32

something a little bit more prudent, say

13:35

2 to 1.

13:39

I'm going to be very judicious with my

13:40

threshold. I want to make sure that the

13:42

signal is just barely coming up over the

13:44

threshold. Remember the further down I

13:47

bring the threshold, the more of the

13:48

signal will be squared off.

13:50

So I really don't want to be too

13:52

aggressive with this or else I might end

13:53

up having it sound a little bit crunchy.

13:56

Just to demonstrate what that sounds

13:57

like though, I am going to show you what

13:59

it sounds like when we go a little too

14:00

far.

14:01

But let's start with just trying to find

14:03

a good starting point. So I'm going to

14:06

bypass

14:07

uh this

14:08

compressor. I'm going to select a little

14:10

bit of uh signal here and we're going to

14:12

loop it.

14:16

>> [music]

14:19

>> Now we have our

14:21

our kick drum soloed.

14:23

I'm looking at the meter and I'm seeing

14:25

that the threshold the way it's set

14:26

right now

14:27

the needle just basically pops right

14:30

over the edge of this threshold line

14:33

just when the transients hit.

14:36

I can also see that it rebounds back

14:37

pretty fast.

14:39

I'm going to up my threshold just a

14:40

little bit more.

14:45

And I'm going to slowly bring it down up

14:46

until I start to see

14:49

some orange flashes that come all the

14:51

way down to close to about -3.

14:55

More than that and I'm going to start to

14:56

really be introducing a lot of harmonic

14:58

distortion.

15:00

The other thing I need to be careful of

15:01

is that there are some hits that are a

15:02

little bit stronger than others. There's

15:04

no consistency here.

15:06

So maybe my goal is to make all the

15:08

transients consistent

15:11

and then work from there.

15:12

So I'll bring down my threshold

15:33

to where all of my transients are

15:34

basically getting hit a little bit.

15:36

This will make sure that all of these

15:37

tops are getting brought down.

15:41

Now

15:42

I'm going to put in uh take the bypass

15:44

out and we'll hear what that sounds

15:46

like.

15:49

Ah instantly

15:51

we can hear that squaring off of that

15:52

waveform, that harmonic distortion.

15:57

>> [music]

16:02

>> Too much.

16:03

I'm going to give back a little bit.

16:06

>> [music]

16:17

>> That's a little better but I can still

16:19

hear some crackle. Now I'll start to

16:21

slope slow down the release time a

16:22

little bit.

16:32

Now I should be able to back my

16:33

threshold in a little bit more without

16:34

hearing as much

16:36

of that distortion because my compressor

16:38

isn't acting as strong. It's not

16:39

creating a sharp a square wave.

16:52

That sounds better. Now if I bypass it

16:58

and with the compressor in

17:04

it's quieter. The transients aren't as

17:06

loud and if we listen very carefully, we

17:08

might actually even be able to start to

17:10

hear some of those harmonics creeping

17:11

into the tone of the signal.

17:13

I'm going to make this a little bit

17:14

louder though. I'm going to use my

17:15

makeup gain now

17:17

to make the overall signal louder.

17:20

That should help to bring out these

17:21

resonance that I'm trying to bring out

17:23

in my tone. But as I do, you'll also

17:25

start to hear the harmonic distortion

17:27

come out from the compressor.

17:31

>> [music]

17:37

[music]

17:40

>> Without the compression

17:45

with it.

17:52

In my description earlier

17:54

uh in my presentation, I discussed that

17:56

we should think about compressors as

17:57

time-based processors.

17:59

What I think we can hear

18:07

to hearing a sound that's actually quite

18:08

long in its resonance. It takes a lot

18:11

longer for the sound to go away.

18:13

It's because I'm compressing the signal

18:15

and I'm bringing up I'm compressing the

18:17

transients and I'm bringing up all the

18:18

noise floor stuff that's happening

18:20

later.

18:21

So what's happening is you could think

18:23

of it as as as controlling dynamic

18:24

range, but what I'm really doing is

18:25

making the sound last longer

18:28

which is a time domain process, right?

18:30

I'm making the sound last longer,

18:31

therefore we're going to hear it a

18:32

little bit clearer. It's going to sound

18:34

louder to our ear.

18:37

The other thing is I've introduced a lot

18:38

of harmonic distortion and we can hear

18:40

that the the kick drum has gone from

18:42

being kind of low and punchy and uh

18:44

ticky to being kind of from from from uh

18:49

loose and flappy.

18:51

These are bad descriptive words but

18:53

they're the best we can do when we're

18:54

talking about subjective uh qualitative

18:57

uh assessment here. Um I can measure the

18:59

harmonics but I can describe it as being

19:01

different in some sort of way that's not

19:03

just the harmonics but the quality of

19:06

the sound.

19:08

So we have to figure out a way to

19:11

balance the uh uh uh

19:14

uh balance the the the issues here. We

19:16

we have On the one hand I'm I'm

19:17

succeeding at compressing and and

19:20

suppressing the transient of the signal.

19:22

On the other hand

19:24

as I'm doing that I'm also bringing up

19:26

the harmonics in the back end of the

19:27

signal and that's not very good.

19:29

So we have to find a way to kind of come

19:32

to terms here. So I'm going to uh

19:35

uh introduce a compromise.

19:37

Perhaps what I want here is not as much

19:39

compression.

19:41

Maybe I can take 1 and 1/2

19:45

to 1

19:46

instead of 2 to 1.

19:49

That should help. I can also soften my

19:51

knee.

19:52

By softening the knee, I'm making it so

19:54

that I'm not squaring off the

19:55

compressor. The com- The compressor is

19:57

not squaring off the waveform quite as

19:58

aggressively. It's going to curve around

20:02

that compression point, which should

20:04

help to preserve some of those

20:07

transients that had that sine wave type

20:11

element to them, so that they're not

20:13

suddenly becoming flat mesas.

20:16

Um that should also help. So, let's try

20:18

introducing some knee and we backed off

20:20

the ratio a little bit.

20:25

>> [music]

20:31

>> Uh that's much better.

20:34

>> [music]

20:35

>> Without the compression.

20:40

With it.

20:43

>> [music]

20:45

>> Here we're hearing a little bit less of

20:47

an aggressive harmonic tonality to the

20:49

compression. If compression adds

20:51

harmonic distortion every time it acts

20:54

on a signal, then using it to fix uneven

20:57

dynamics means adding coloration to

21:00

every moment you're trying to correct.

21:02

The compressor doesn't know it's dealing

21:04

with a word that got swallowed or a

21:06

guitar phrase that was played softer. It

21:08

reacts to level and it reacts the same

21:11

way every single time. You're smarter

21:14

than a compressor, though. For the

21:15

dynamic problems that are really about

21:18

performance, a phrase that's too quiet

21:20

or a syllable that got lost, you might

21:23

be better off going in manually. That

21:24

could mean adjusting the clip or region

21:26

gain, which happens before the signal

21:29

hits the compressor. That way, you

21:31

smooth out the performance before the

21:33

compressor even sees the signal. Then

21:36

the compressor can focus on what it's

21:38

actually there for, adding texture and

21:40

character, if you want texture and

21:43

character.

21:44

Here's how Gabe explains it. You should

21:46

try to get 90% of the way there with

21:49

your tracking technique, your clip gain

21:52

automation or with volume automation,

21:54

and then use the last 10% of that

21:56

control. Um maybe that's done with a

21:59

compressor.

22:00

Okay? And And this is the number one

22:02

trick that once I show my students how

22:04

this works, like they're suddenly

22:06

compression just becomes a lot more

22:07

intuitive.

22:09

It's funny. There's things about dynamic

22:11

range that have to do with performance

22:14

and not um harmonic distortion. So, if

22:17

you want to control the dynamics of a

22:19

signal,

22:20

you shouldn't reach for compressor to do

22:22

that.

22:23

Reach for um if you're in the tracking

22:25

stage,

22:27

repositioning the microphone. Maybe it's

22:29

pulling the microphone further away, so

22:31

that you don't have quite as much of a

22:32

dynamic range between loud bits and

22:34

smaller bits, because everything's less

22:35

focused. That can still sound really

22:37

great. In fact, pulling the mic back um

22:41

might help you

22:42

um be more successful at using

22:44

compression while tracking, because the

22:46

compressor isn't trying to do the heavy

22:48

lifting of pulling the levels down when

22:49

it goes significantly above the

22:51

threshold, but most of that's just been

22:53

taken care of naturally by the uh the

22:56

inverse square law and the fact that

22:57

you're pulling the microphone away,

22:59

you're pulling the mic away, so you get

23:01

less proximity effect in the vocal. Um

23:03

the dynamic range of the vocal will get

23:06

naturally more even. And then when the

23:08

compressor is activated, it's working on

23:10

a much smaller percentage of the signal

23:12

than if it's right up in the mouth of

23:14

the performer, where it's having to do

23:16

an awful lot. You're just going to get a

23:17

lot of harmonic distortion from that,

23:19

not necessarily dynamic range control

23:21

that you might want in the presentation

23:23

of a performance.

23:25

So, um

23:26

my suggestion is to try to get

23:30

the dynamic range part of your signal

23:33

under control using automation and be

23:36

intelligent about which syllables you

23:38

want to emphasize. Put a little flourish

23:40

on the last line of every vocal. Um

23:43

uh think about words like uh and the and

23:46

from and the small syllables that

23:47

sometimes get lost. Try to boost those

23:50

up with gain control, um gain

23:53

automation, so that they're more

23:55

present, because a compressor isn't

23:57

going to be listening to the lyrics and

23:58

deciding, I didn't understand what they

24:00

said right there. I should have

24:01

compressed that better.

24:03

Compressors are not smart. They do not

24:05

take intelligent uh

24:06

intelligence into into consideration

24:08

when they make their changes. They act

24:10

aggressively and consistently, and they

24:13

will not bring humanity into your

24:15

tracks.

24:16

The goal isn't to avoid compressors.

24:19

It's to use them for what they do well

24:21

and stop asking them to solve problems

24:24

they can't fully understand.

24:28

You're probably already familiar with

24:30

the standard use of reverb in a mix. You

24:32

put something in a space to make it feel

24:34

more real or you add a shared room to a

24:37

mix, so everything feels like it belongs

24:39

together. Both of those are useful, but

24:42

when I watched Gabe explain another

24:43

situation where he uses reverb, it took

24:46

me by surprise.

24:47

He thinks of reverb in some cases as an

24:50

EQ tool.

24:51

Not just a metaphor for EQ, but as an

24:54

actual way to control the tonal balance

24:56

of the full sound.

24:58

When a room resonates, it reinforces

25:00

certain frequencies and lets others

25:02

decay faster. So, by choosing the type

25:05

of reverb, shaping its frequency content

25:07

and controlling its decay time, you can

25:10

augment the tonal character of the

25:12

instrument in a way that feels more

25:14

tonal than spatial. Here's Gabe working

25:17

on an acoustic guitar that's been EQ'd

25:20

and compressed and is sitting reasonably

25:22

well in the mix, but still feeling a bit

25:24

pointed and edgy.

25:26

He goes to reverb in this case.

25:29

People would think of reverb as being

25:30

spatial, but I'm actually going to use

25:31

this to create timbre. Make the the body

25:34

of the guitar feel a little bit more

25:36

mellow. As I'm listening to this, it

25:38

just feels like the guitar is very kind

25:39

of pointy and edgy. So, what I've done

25:43

here is I've created a an aux return and

25:46

a send

25:47

that's called a GT verb and I'm going to

25:49

send this acoustic guitar to this

25:52

return.

25:53

And on this return, I'm going to insert

25:55

a reverb and I don't want anything too

25:58

special. It doesn't have to be something

25:59

fantastical. In this case, I'm using

26:01

Dverb.

26:03

I'm going to turn the gain reduction all

26:05

the way off, so that it's not pulling

26:06

anything out of there and I'm going to

26:08

set the room parameter to small.

26:10

And I really don't want much here, maybe

26:12

about 178 milliseconds.

26:14

That's all I really need to do. And then

26:17

I want to pull the high frequencies down

26:19

to maybe oh, I don't know, 1.6k.

26:24

So, I'm going to put the fader up here.

26:27

Let's hear what this sounds like. I'm

26:28

just going to solo the acoustic guitar

26:30

and we'll see what this reverb sounds

26:32

like.

26:35

>> [music]

26:38

>> If I mute the reverb,

26:45

that's pretty cool.

26:46

It gives it a room. It gives it a space.

26:48

It definitely makes it feel more stereo.

26:50

And I'm starting to get into the

26:51

direction of it sounding warmer, but I

26:53

haven't really gotten there. So, I'm

26:55

going to go into like my EQ here after

26:57

this reverb and I'm just going to make a

26:59

bump at around 200 hertz. What I'm also

27:02

going to do, so that it doesn't get in

27:03

the way of the bass, is I'm going to

27:04

pre-delay this signal maybe about 10

27:07

milliseconds or so. And I'll stop at 12.

27:10

That seems That seems feasible. And I'm

27:13

probably going to want to de-emphasize

27:15

the high frequencies as much as I can.

27:17

So, I might just throw in a a nice

27:18

gentle shelf here

27:20

to further that along. So, I'm really

27:22

tilting the the tonality of this and

27:25

this this hopefully it makes the room

27:27

reflect the low harmonics that are

27:29

bouncing around inside the reverb

27:31

program. That's going to make the guitar

27:32

feel thicker without adding low

27:34

frequencies via EQ into the sound

27:37

directly.

27:39

>> [music]

27:43

>> And I feel a little space between where

27:44

the reverb starts and where the guitar

27:46

ends.

27:52

So, you can hear how the guitar is

27:53

starting [music] to take on that darker

27:54

tone.

27:55

Now, I'm going to do one more thing. I'm

27:57

going to compress the reverb. What we're

27:59

going to try to

28:00

achieve is take all of the fast-acting

28:03

stuff that's in there and kind of pull

28:05

it out and let it release really slow.

28:08

We're going to set the ratio of

28:10

compression to maybe 5 6 to 1 and a

28:13

really soft knee

28:15

like this. And I'm just going to pull

28:16

back the threshold, so we're burying the

28:18

reverb kind of in constant compression.

28:27

Then I'm going to use the output gain to

28:28

bring it back up in my mix.

28:43

And you'll feel [music] that instantly

28:44

my guitar sounds and feels warmer, like

28:47

a warmer bodied instrument.

28:49

If I took this out, [music]

28:54

not only do we lose the space, but you

28:56

can sense that [music] the low harmonics

28:57

just kind of disappear out of it.

29:01

The guitar gets warmer and thicker

29:03

without any change to the dry signal

29:05

itself. He's not EQ'ing the guitar. He's

29:08

EQ'ing the room and what it gives back.

29:11

The same principle applies across

29:13

different instruments. Here's Gabe again

29:15

on a snare drum, deciding he wants the

29:17

snare to sound thicker and fuller.

29:21

>> [music]

29:30

>> Okay, now just a short reverb.

29:46

You can almost hear how it's giving it

29:47

of a flamy feel to it.

29:49

And again, this is just the dry snare by

29:51

itself.

30:01

>> [music]

30:02

>> Now with our processing.

30:07

>> [music]

30:12

[music]

30:17

>> So hopefully you can hear that there's a

30:18

change not just in the presentation of

30:20

the stereo image or the space around the

30:22

snare, but also a temporal shift in the

30:24

snare.

30:25

I could have made the snare feel thicker

30:26

with EQ,

30:28

but the room can do a lot of that work

30:29

for me as well. And rather than try to

30:31

make that individual track uh thicker or

30:35

uh try to push more low frequencies into

30:37

say a compressor which might react

30:38

negatively to it, instead I'll just do

30:40

some parallel processing in the reverb

30:42

and try to get the reverb to add

30:44

harmonics to my dry signal and hopefully

30:46

make the snare uh sound more full and

30:48

lush.

30:49

Different instrument, same idea.

30:52

Once you start thinking about the reverb

30:54

return as something you can shape

30:55

tonally,

30:57

the way you use reverb will change.

31:00

The central idea Gabe comes back to

31:02

throughout the delay section of the

31:03

course is that time and space are the

31:06

same thing. Delay is how we perceive

31:09

distance. When a sound arrives at your

31:11

left ear slightly before it arrives at

31:13

your right ear, your brain doesn't hear

31:15

two different sounds, it hears one sound

31:18

coming from the left side. And that tiny

31:20

timing difference is how you locate

31:22

things in the world around you. And

31:24

you've been doing it your whole life

31:26

whether you thought about it that way or

31:27

not. Think about how you judge distance

31:29

with your eyes.

31:31

Close one eye and you lose depth

31:32

perception.

31:34

The brain needs two slightly different

31:36

views of the same thing to build a 3D

31:39

picture. And hearing works in a similar

31:42

way.

31:43

Delay is one of the most important tools

31:45

we have for creating dimension and

31:47

depth.

31:48

What makes this practical is

31:50

understanding that different amounts of

31:52

delay do completely different things.

31:55

Gabe organizes them into three time

31:57

windows. The first is the binaural

31:59

window from roughly 1 to 10

32:02

milliseconds. In this range, your brain

32:04

doesn't separate two instances of the

32:06

same sound into two distinct events. It

32:10

sums them together and reads the timing

32:12

as a localization cue. A few

32:14

milliseconds of delay on the right side,

32:16

for example, makes the whole thing feel

32:19

like it's coming from the left with no

32:21

panning required. Some engineers use

32:23

this concept called the Haas effect to

32:26

push things wider than the physical

32:28

speaker boundaries. The catch is mono

32:30

compatibility though. When you sum a

32:32

binaural delay to mono, the two versions

32:35

of the signal interfere with each other.

32:38

And what might have sounded huge in

32:40

stereo sounds phasey in mono. So it's

32:43

always worth checking your mix in mono

32:46

before committing. The second window is

32:48

early reflections from about 15 to 35

32:51

milliseconds. This is where delay stops

32:53

being a positioning tool and starts

32:56

being an acoustic tool. At 15

32:58

milliseconds, the brain has just enough

33:00

processing time to register the delayed

33:03

signal as something separate from the

33:05

original. And done intentionally, it can

33:08

register to the listener as a reflection

33:10

of the nearby wall. A helpful guideline

33:13

is that 1 foot of distance equals

33:15

roughly 1 millisecond of delay. So 15

33:19

milliseconds implies a 15-foot path.

33:22

That means delay can be used to define

33:24

the size of the space you're trying to

33:26

create within your mix. You push it

33:29

higher in this range and the implied

33:31

space gets larger. This is how you make

33:33

a dry recording feel like it was

33:35

captured somewhere. And the concept can

33:38

also be implemented using the pre-delay

33:40

setting that you might find in a reverb

33:43

plugin. The third window is late

33:45

reflections from 35 to 60 milliseconds.

33:49

This range adds a deeper sense of room,

33:51

less about where you are and more about

33:54

the character of the space itself. The

33:57

back wall, the texture of the surfaces,

34:00

mono compatibility is less of a problem

34:02

here because the signals have enough

34:05

time between them that they're not

34:06

fighting each other as much. And the

34:09

level between the direct and indirect

34:12

delays tends to be wider and that also

34:15

results in less interference. Beyond 60

34:18

seconds, the brain stops reading the

34:20

second sound as connected to the first.

34:23

It becomes its own event. That's the

34:25

territory of rhythmic echoes and

34:27

slapback creative effects rather than

34:30

acoustic simulation. When you understand

34:32

how the brain interprets different delay

34:34

times,

34:36

you get a sense for what's possible with

34:37

delay. And the next time you want to add

34:40

dimensionality or special effects to a

34:43

mix, you can get close to the settings

34:45

you need without digging through

34:47

presets. So those are the four. EQ

34:50

through the lens of vowel sounds so you

34:51

have a reference every time you open an

34:53

EQ. Compression as something that shapes

34:55

the tone of a signal just as much as the

34:58

dynamics with clip gain handling the

35:01

outliers in the performance before the

35:03

signal reaches the compressor. Reverb as

35:06

a tonal tool, not just a spatial tool.

35:08

And delay as a set of time zones, each

35:12

one producing a different effect on how

35:14

the listener experiences the space

35:16

around the sound. There's a point Gabe

35:18

makes at the end of the course that I

35:20

don't think enough of us focus on while

35:22

we're mixing. It's one skill to use

35:25

compression and EQ to make individual

35:27

instruments sound good.

35:29

It's a different skill entirely to use

35:31

those same tools to make instruments

35:34

work together. So everything has its own

35:37

space and so that the whole thing feels

35:40

like one cohesive sound.

35:42

That second skill is what the course is

35:45

actually building toward. Information

35:47

about mixing or anything for that matter

35:50

isn't hard to come by anymore.

35:52

And that's exciting.

35:54

But the challenge is still to build a

35:56

skill that will actually stick with you.

35:59

And that takes structured practice,

36:00

feedback on your work, and time spent in

36:03

the material with other people working

36:06

through the same things. The Audio

36:07

University membership is built around

36:09

that exactly. The courses are taught by

36:12

myself, Gabe Herman, and other guest

36:14

instructors with decades of experience

36:16

mixing and teaching. And beyond the

36:19

courses, both Gabe and I are active in

36:22

the membership. So you can leave

36:23

comments directly on the lessons or ask

36:26

questions when something doesn't click.

36:28

And you can get a response from one of

36:29

us or from both of us and we'll each

36:32

share our perspective. We also run live

36:34

coaching events periodically where you

36:36

can ask questions face-to-face and

36:38

there's also a mix review feature where

36:41

you can submit a mix and get expert

36:43

feedback from Gabe himself. A lot of

36:45

people in the membership are working

36:47

through the same problems and hearing

36:49

responses to other people's questions

36:52

tends to move you forward in ways that

36:54

watching videos won't do on its own. You

36:57

still have to put the practice in

36:58

yourself though to actually hear

37:00

results. But having that support

37:02

structure around you changes how fast

37:05

that happens. The membership includes

37:07

the ear training lab and downloadable

37:09

multi-tracks so you can put the theory

37:12

into practice. If you're looking for

37:14

more than just videos to watch and

37:16

you're willing to put in the work, the

37:18

link is in the show notes below.

37:20

I'll see you in the next video.

More transcripts

Explore other videos transcribed with YouTLDR.

Get the TLDR of any YouTube video

Transcribe, summarize, and repurpose videos in 125+ languages — free, no signup required.

Try YouTLDR Free