Singing With Headphones (Part Two)

Jun 21 2021

Not so long ago, I wrote a post called “Singing with Headphones”. It explained the troubles I had with singing flat when I had headphones on, and possible ways to solve that.

Unfortunately, the techniques I ended up recommending did not (fully) do the job for me.

I was still inconsistent. Sometimes a whole recording session would be on-pitch (first try, nothing special necessary), but the next time a whole set of takes might have that same weird problem again: all my vocals were EXACTLY 1/4 note underneath the real note.

So I continued experimenting. I downloaded Reaper (DAW), which has a built-in plugin ReaTune which is very good (and fast) at recognizing pitches and even correcting them. With that tool available, I recorded and analyzed my vocals until I was sick of hearing my voice.

What did I learn? Well, let me explain!

The Problem

To reiterate, the problem was as follows.

Whenever I sing into a microphone, using headphones or some other accompaniment (such as a background track coming from the speakers), I tend to sing all the notes flat. (This doesn’t happen otherwise.)

I can “solve” this (most of the time) by simplying autotuning the whole take upwards by 50 cents. But that creates artefacts, often doesn’t sound natural, and I don’t want to rely on these kinds of fixes.

I’ve sent mixes to people around me, both with the autotuning trick above and without it (where I managed to just sing the whole song on-pitch), and they all respond the same way. The autotuned track sounds more “mellow” and fitting, but the lyrics are hard to understand and it doesn’t sound consistent. The real track is a bit more aggressive and energetic, but ultimately sounds more professional and intelligeble.

(On a related note, this also made me realize that vocals are more important than anything. I have NEVER received comments on the guitar playing, even when I knew there were mistakes and it was just a rough first version. Everyone listens to, and focuses on, the vocals. So from now on, I’ll probably record all songs “vocal first”. But that’s an article for another time.)

This deviation from the real note is slight enough that it often doesn’t sound outright wrong (it’s still very near the real melody), but it creates a vocal that sounds dull, lifeless, and cannot blend well with the rest of the tracks. (Because they are literally playing and singing different notes.)

So, if you have a mix where the vocal just will not blend in, or sounds dull and uninspiring, this might be your problem.

It, unfortunately, took me two months to find out that this was the cause — I was simply singing the wrong notes — which means I now have to re-record almost all vocals :/

Finding something that works

The quest now becomes: can I find a certain setup, a way of recording my vocals, in which I don’t have this problem?

Ideally, I would train myself to sing with headphones on. But that will take a long time, if I manage to do it at all, and I don’t want to wait any longer with recording and publishing my music.

So, for now, I wanted to find just something that worked.

Eventually, I made these realizations:

In all other situations, it’s a coin toss. Will I be on-pitch? Will I be completely off? I don’t know until I actually do the take and analyze it afterward.

Nevertheless, finding these situations that help my vocal recordings was a huge help. I was feeling really down, quite disappointed with myself, that weeks of vocal takes were just sounding bad all the time. Finding these situations that worked, and hearing that my takes were suddenly great, lifted my moods (and motivation) enormously.

So that’s the advice: try all sorts of stuff, analyze the recordings (using e.g. a plugin), and remember the things that worked best. Build on that.

How do we use this?

Solution 1: Guitar

Solution 1 has issues. Because I play the guitar at the same time, those notes will bleed into the microphone. If you solo the vocals on my album “Keeping the Sunset on my Side”, you’ll hear soft guitar notes in the distance everywhere. Fortunately, this isn’t an issue when you bring back the guitar accompaniment (it’s the same guitar, playing the same notes) …

… as long as I don’t play the guitar too aggressively. But if I play too softly, I can’t hear the note, and it doesn’t help me!

This technique, therefore, only works well if I can play the guitar somewhat loudly without problems.

(Additionally, on difficult songs, I would often make mistakes with the guitar notes, ruining the vocal take. It’s quite interesting actually: I can play the songs with my eyes closed, but if you ask me to ONLY play the lowest note of each chord, as SOFTLY as possible, I mess up regularly.)

Solution 2: No accompaniment

Solution 2 also has issues. I can’t do a long take, because I will “lose” my pitch anchor throughout. Additionally, I have no sense of timing (not even a metronome in my ear), so any take I lay down will have to be edited and cut severely to match the song.

But most importantly: you just can’t get the same feeling, and energy, and groove when you’re literally singing into the void. The room is silent, there’s not even a guitar to guide you, and you have to deliver a wonderful emotional vocal take?!

Solution 3: Studying the melody

This one surprised me the most.

I thought I “knew” the melody. I thought I was singing the right notes, but it just didn’t come through on the recording.

But once I looked more objectively (again, using a plugin), I realized some big gaps in my vocal technique. On certain melodies, at least.

For example, take song #6 from “Keeping the Sunset on my Side”. The bridge contains the phrase “I’d walk a while (with you)” several times, sung like a choir or big crowd (drawn-out, multiple voices layered on top of each other, etc.).

I just COULD NOT get it right! I would sing it time and time again, and each time it sounded bad. No two takes fit together. But I knew the melody, right? I heard myself singing the right notes … right?

Well, no. What should have been an A4, I consistently sung more like a G4 or G#4. Moreover, whenever I started a phrase, I did not hit that A4 immediately, I did some sort of quick “slide” upward, “searching” for the note.

So I took a step back. I wrote down the exact notes I was supposed to sing. Then I played them on my guitar (which was tuned!) and sung over them, five times, until I could get every note right.

Then I did another take … and it was almost immediately perfect :)

In hindsight, it’s only logical. How can you get a clean, professional vocal sound, if you don’t even know what you’re supposed to sing? If you’re just guessing at the right notes? Even if only a single note is off, it can break the whole sound.

From now on, I will do this, especially on the hardest parts. Determine the exact notes, practice singing those exact pitches, and then start recording vocals.

Solution 4: Higher notes

Well, this isn’t really a solution. More of a statement: my low notes need more practice. I tend to muffle them, not caring whatever pitch they are, as long as I get the mid-high notes right.

Of course, I could write extremely high melodies from now on … but I actually already do that and really don’t want to go any higher! No, I just have to train my low notes, keep paying attention to them.

(To give you an idea: that A4 I talked about earlier is a note that many male vocalists cannot even sing. For others it’s a high note that needs to be trained. For me it’s more of a middle-range note. The problem with that is, of course, that any note below F3 is unsingable for me, as it’s just too low. That’s why I automatically write higher songs and melodies.)

Solution 5: The “ooeh” sound

My vocal teacher actually once made a passing remark about the “oeh” sound being a good exercise for people with my voice type. I wish I’d listened earlier :p

This vowel is actually quite hard to pronounce for me. Which is probably a sign that I need to train it more and learn from the placement/feeling of singing that vowel.

Remember my example with “I’d walk a while” (song #6), which I just could not get right? Well, there is one long sustained note “youuuuuuu” in there. And guess what? I sung that perfectly on every take. Even when all the other notes were wrong, this one (also an A4) was always perfect. (The vocals you hear on the album are the raw first take, without any editing or autotune, that’s how on-pitch that vowel makes me sing.)

What’s the conclusion? Once I realized this, I started warming up with this vowel, and started singing it before any part I had difficulties with. It works wonders.

Why is this vowel hard for you? In case you don’t know who I am, I have serious health issues that produce a lot of strain in my neck/jaw/shoulders. On bad days, I literally cannot pronounce “eee” or “oeh” sounds properly. There is just too much tension, too little coordination in those muscles, for those sounds. (This is also the reason why I’m more interested in getting a good vocal sound in the studio, whatever it takes, rather than just saying “learn to sing better!")


So that was my adventure. I learnt a lot about the issues with my current voice, but also about its strengths, and which situations help me perform.

My greatest advice is, therefore, to just record yourself a lot and analyze the results. It’s the only way to know for sure if you’re on-pitch, if something helps you or not, if there are any patterns in your performances.

From now on, I do the following:

It’s still not perfect, because I realized there are still areas of my voice I need to train a lot, but it gets me to perfect recordings 90% of the time. Which is way better than the constantly off-pitch recordings I used to get 100% of the time :p

An alternative solution

I’ve noticed there are two different types of musicians in the world.

I am the first category. I’ve played and sung live for 10-15 years now. When I started recording myself, it was very weird to sing without holding an instrument, or to hear myself recorded and edit that.

So yeah, if that’s what you know and what you’re comfortable with, that creates issues when recording yourself.

If that’s the case, you have two options.

And even though you would do BOTH things in the ideal world, I’m starting to lean towards the second option.

If I buy an extra microphone, and a bigger audio interface, I can record myself AND the guitar at the same time. I can simply capture live performances, edit them to my wishes, and release those.

I’m pretty sure it will lead to better, stronger, more blended tracks, because I’ve always done this and know how to make it sound good. At the same time, it removes a lot of freedom (whatever the vocals, whatever the sound of the room, whatever tiny mistakes, you’re kinda stuck with it), and it would cost me a lot of money.

However, if that is no issue to you (you have money/gear, an okay room, and are a consistent performer), I would strongly suggest this alternative solution!

Recording stuff with headphones is actually quite a recent development. In the good old days, artists would often just record the whole group simultaneously (playing in the same room), or freely sing over speakers/other players accompanying them.

Update! I just read some very interesting ideas. Namely that our left ear is more attuned to rhythm, and our right ear more attuned to pitches. Many people report great results when giving people the beat/drums/rhythm guitar in the left ear, and their voice in the right. Will try this on my next album.