Notes on 'Feast your Eyes on Me'

Jan 24 2022 Album Diary

I just released my third singer-songwriter EP called “Feast your Eyes on Me”, which means it’s time for an album diary!

In this article, I’ll explain the process behind the development of this EP, focusing mostly on what I did, the mistakes I made, and the lessons I learned from that.

What’s the idea?

I’ve written literally hundreds of songs over the years. I was fed up with having them just sit on my hard drive and really want to record and publis them this year, which is why I set myself the goal of “one small EP per month”

This is the first one. Taking the lessons I learned last year, I tried to streamline my process (to make fewer mistakes, work more quickly), while working towards that “radio-ready” (or just “professional”) sound.

The main differences were:

Sketch Recordings? I guess you could also just call it a demo. Why do this? Because, with previous albums, I usually only had bits and pieces of the song recorded. The first time I actually put together the whole song, was when I was already recording it. Which led to many awkward surprises (weird transitions, things that didn’t sound well together, a sudden increase in tempo).

By doing these “sketches”, I have a full recording of what the album is going to sound like, and can fix any issue during that stage.

More Instruments? I’m still a singer-songwriter. And a beginning mix engineer. So I want to start with my simplest songs first, which means almost none of them would benefit from a full band (bass, electric guitar, drum, etc.).

So what instruments are we looking at? Piano, Strings (cello/violin), a soft sort-of-Drum (cajon/shaker/tambourine), Electric Guitars.

(Two songs on the album might be possible as a band version, namely “Maybe you were right” and “Short jumping”. But I decided to not pursue that for now, keeping things simple, and the general sound of the album consistent.)

Lessons Learned

Being my third album, I (luckily) didn’t make any major mistakes anymore or met other major obstacles. Which means this diary will be much shorter and I can just summarize the lessons I learned below!

Lesson 1: First beats should be strong

In almost all songs, whenever we cut to a new section (verse->chorus, chorus->verse, intro->verse, whatever), I did the following:

This meant the first beat of a new section has some extra impact. This just sounded way better and more natural to me. And it was part of the solution to one of my biggest flaws: making songs that sound too “same-y” across the board, with no clear sections or climaxes or powerful moments.

Lesson 2: Don’t make the jump to vocals jarring

Being a singer-songwriter, most of my songs start like this:

But … vocals must be quite a bit louder than guitar. (I made the mistake of soft vocals on my first album, which really taught me how loud they actually are on most tracks!)

So this transition from “intro” to “verse” was quite jarring and felt like someone had just stumbled into the room and starting yelling into the microphone.

What’s the solution? It’s terribly obvious in hindsight: make the intro almost as loud as the vocals. In most songs, the intro is just a few guitar tracks (usually doubles of the same chord progression). I simply selected them all and bumped up their volume, then listened to the transition (intro->verse), and repeated that cycle until it wasn’t jarring anymore.

Lesson 3: Soft-doubling of vocals

In the previous album, “doubling vocals” meant …

But I’ve learned that there are better ways! In many songs, if you listen closely, you can hear doubles playing really, really softly in the background. Even if it just sounds like one voice, even if it’s just a soft intro to a song with an inimate vocal, usually there are still doubles.

They are just very soft and out-of-the-way. Creating nothing more than some extra “thickness” to the voice, without our ears perceiving them as separate voices.

So that’s what I did. In many parts where I’d normally do only one voice, there can now be one or two doubles playing very softly in the background. And other parts where I might do doubles at the same volume, I’ve learned to

Lesson 4: Reverb & Delay

These are the lessons learned:

And, of course, increase the intensity and volume of reverb/delay during the climactic parts. (A chorus, a bridge, a solo, etc.)

Lesson 5: I need tricks when singing with headphones

I’ve tried all sorts of things, but when I have my headphones on, it’s just so unnatural to me that I keep singing off key.

I’m used to singing while holding and playing a guitar, or perhaps a piano. So I thought: why not just do that?

Most vocals on this album were sung while I was holding a guitar and softly plucking the note/chord I’m supposed to sing

Yes, it means some of that guitar sound sometimes bleeds into the vocal recording. Yes, it feels a bit like a handicap, not being able to consistently sing without this help.

But it’s the best I can do for now. And hopefully, by doing this over and over, my ears can adjust to having headphones on and learn how to sing well with them.

How did you do it before? By autotuning/pitching up the vocals afterwards. Believe it or not, I can actually sing :p So all my takes were exactly 1/4 note below what they should be. Pitching up would make it in-tune for 99%, which basically solves the problem. But this kind of editing adds artifacts that sound unpleasant. And it makes the vocals less natural and emotional, more digital and, well, emotionless.

If I can sing something in tune, naturally, it automatically sounds way better and meshes better with the other instruments, simply because it’s how the voice is supposed to sound. Even if it was a bad take, with many blue notes and a cracking voice, it still sounds better than a take that’s edited afterwards.

The point is this: I think the vocals on this album are a step above my previous attempts, and they only took a few takes each (and almost zero editing), making them considerably faster to produce and mix.

Lesson 6: Use bad mics to your advantage

The default microphone inside my laptop is terrible, as it should be.

However, that does mean it creates some different sounds, some extra resonance and frequencies, that my “good” microphone doesn’t add.

And I’ve learned that this is awesome for background vocals!

So, for many songs, I just softly sang the background harmonies into my laptop microphone. (Whenever I had time and felt my voice was good.) Adding those to the main voice just made it sound really good. And even better: I sing more freely without headphones, behind my laptop, so they are usually in tune the first try :)

Lesson 7: Decide whether to use drums BEFORE recording starts

The second song really needed some drums for the steady groove/beat … but when I started recording, I wasn’t sure if it was needed (or how they would sound), so I hadn’t recorded them yet. Recording went fine, a few days went by, then I tried to add the “support” for the guitar + vocal, including drums.

Guess what? My recordings weren’t exactly keeping a steady beat! They sounded great, but I sped up in some places, and slowed down in others, and matched all recordings based on what I heard. When I tried to put a standard, grid-aligned drum beat in the track, it just made it made it sound worse. Even though the drum was keeping time perfectly well, the other recordings weren’t, so it often happened that it sounded as if the drum was playing a whole different song completely!

So, I ended up having to painstakingly move the individual drum hits to match the subtle changes in tempo of my other recordings. Was a lot of work. And in some places, I just couldn’t figure it out, which is why there’s relatively little actual drumming in that song.

In the future, I’ll need to make a decision beforehand. And if there’s any chance of a drum being added to the song, just add it anyway to be certain, so I can ensure all future recordings match the groove.

Lesson 8: Multiple layers = lower volume

During the chorus, and the bridge/solo, there are usually many layers/instruments active. Even if they only play a simple melody, even if they’re quite soft, all that volume adds up to a super loud bit of music.

This causes an ugly discrepancy in loudness between e.g. a verse and the chorus after it. Yes, choruses should be slightly louder than verses (to add extra energy and power), but not that much louder. They were so loud that I just couldn’t hear the vocals I was singing on top of it. That’s annoying, as it makes the vocal sound more distant and “bad” by default, and makes me uncertain as a vocalist.

(It sounds as if someone is shouting at you from the other side of the street, while a fanfare is marching between you. You can’t hear what they are saying, and even if they were perfectly in tune with the music, it would sound odd and annoying to you.)

And, in the worst case scenario, this causes clipping and audio information is lost.

The problem is: just lowering the general volume isn’t a solution. It will still clip … and then just reduce its volume afterwards. (Additionally, such broad edits aren’t great, because there are always unforeseen consequences in some corners of the mix.)

Instead, I had to go through all layers and lower them all a few dB, until their total volume was the same as the verse before. This gave me some extra space to play with effects and vocal volume, to find the perfect extra “push” for the chorus.

In the future, I’ll need to keep this in mind during recording. Immediately lower the gain after the recording is done, or play less loudly in chorus parts. (Although the reverse is perhaps more useful: play more loudly during the verse, as that’s often the time where the only accompanimant is a lone guitar, and we need all the volume we can get to prevent noise slipping into those recordings.)

About the songs

If I have something special to say about a song, this is where I’ll do it!

1. At the Right Time

I didn’t know if this song needed the Western guitar or the Spanish one. So I decided to try both.

I recorded the Western first (completely, so even extra guitars and doubles). It sounded quite good, although some parts of the pattern were really hard to get right, as steel strings are just tougher and less suited for the fingerpicking here. And that meant some parts were a bit “off”, leading to some sections that were a bit messy.

I had to spend a few boring hours to line up many sections :p But I just could not get them right in one take!

Because it was getting late, I couldn’t record new stuff, so decided to first add the other instruments. (A bass, a piano, some violins.) It really added a lot to the previously guitar-only backing!

I had to watch out the other instruments didn’t overpower the guitar. Because they are played with my keyboard, using virtual instruments, their volume is just much louder by default. So I had to lower their volumes about 3 or 4 times before I was able to comfortably hear both the piano/violins and the guitar plucking.

As usual, the next morning I recorded the low harmonies with my morning voice, so I could sing much lower than I normally would. Feeling my voice was a bit burned-out, I decided not to push and stop singing for the day.

This meant that the main vocals were recorded after all other songs were done. With those done, the song was basically finished.

I’ve learned that good composition and good quality initial recordings lead to a strong dry mix, which means no editing afterwards is really needed. (Well, if I was a better audio engineer, I’d probably find tons of stuff to improve. But compared to earlier albums, these songs are 90% done by the time all recordings are done, which is a great improvement.)

In the end, I kept the steel strings for this song! The Spanish guitar was just too … soft to accompany this whole song. I did, however, introduce it later in the song (especially the bridge). For extra volume, some frequencies/warmth I was missing, and because the bridge chords are too hard for me to play on steel strings at the moment :p

2. Maybe you were Right

In classic Tiamo-fashion, when I grabbed my notes for this song, there were three blanks: intro -> do something, solo -> do that thing, outro -> do something

Fortunately, I still knew what “that thing” was, and it was indeed a fine solo. But the intro/outro had to be improvised on the spot.

(I’m on a schedule here and it takes a while to setup my old equipment. So I can’t just shut down everything and go back to the writing board for a few days, then go back into my studio.)

The guitar parts were pretty basic (one-strum chords, or strumming in a simple pattern). I didn’t see any way to make that more interesting, and after playing for a while I realised it shouldn’t be more interesting.

Instead, there are many silences in the song, which makes it ideal for riffs and licks! So, the leading guitar is really basic (just 2 layers of strumming), but it’s supported by fun melodies all around it (both low and high on the guitar).

The extra instrumentation was another tough nut to crack. Piano didn’t seem right. Nor strings. Drums would only fit at the end of the chorus (or bridge) as a bonus. Some wind/brass instruments felt better. But my favorite? Bass.

In the end, I added a loud and clear bass with lots of fun little melodies. (The “mid-high frequency content” of the song had to mostly be filled by the guitars and the brass.)

Lastly, the vocals. The chorus is 99% low (somewhat repetitive) notes, then ends on a really high note. So it was quite tough to sing those well, without sounding either dull (like I’m just stuck or forgot the lyrics :p) or straining.

Not sure what to make of this song, but at least it taught me a lot and sounds different from anything else I’ve made so far.

3. Play this Mess

The simplest song of the album. (I try to have at least one barebones, guitar + voice only song per album.)

At least, that’s what I thought. Yes, the chord progression and melody are relatively simple. But then the challenge merely shifts: now we need to make it sound great with just a simple guitar and vocal!

I chose the Spanish guitar for this one. It’s softer, suits the general tone of the song better, and I really liked my sketch recording of this song (which used that guitar).

In the end, I did add some extra guitar layers, and pretty simple piano/violins.

But 90% of the song is still just a single plucked guitar. And now that I have two microphones which I can mount on a single standard, I could record one plucked guitar with two mics simultaneously! Which means it just sounds much wider, and I can be more loose (and improvise a bit) with the fingerpicking. And I don’t have to double stuff afterwards, with lots of effort, all the time.

(Although I discovered one of my microphones doesn’t really have clean gain, so I can’t turn up its volume very far before I hear a lot of noise and high-pitched sounds :/ I ended up pointing that one at the louder part of the guitar. And just played more aggressively.)

Again, the low harmonies were recorded first, in the morning. I was simultaneously working on finishing my previous album, so I used the rest of my voice for re-takes of that. The actual full vocals were recorded later, after creating the basic skeleton of all songs.

4. Short Jumping

Right from the start, I knew I couldn’t just record one guitar part (with my nylon-stringed guitar) and do this song justice. It’s too upbeat, too uptempo, it needs more instruments and power.

So I decided to record the plucking with my Spanish guitar, but support it with chords (and maybe rhythm) using my Western guitar.

It was tough at first to find a good blend of the two sounds, so I spent a while adjusting volumes and trying different patterns, until I found something acceptable. (Not great, but surely better than if I’d only recorded a single guitar.)

I also felt that the bridge needed something like a trumpet, so I added it. But a digital trumpet (played through my plastic MIDI keyboard) doesn’t sound great, so I copied the same recording to several other wind instruments (such as horns) and played them simultaneously to get a fuller sound.

(Though some instruments are panned to the left and others to the right to actually make it sound wider and, well, more amazing with headphones on.)

Once added to the bridge, I knew I had to add them at other places as well. (Start small, but slowly build towards that big bridge.) So that’s what I did, and especially if you lower their volume quite drastically, trumpets and the like give a great atmosphere to a song. You barely notice it’s there, yet it fells the empty space well.

Lastly, when I recorded the final part, I created an accidental bass line at some point. When I heard this, I was like: this sounds amazing … I should make it, well, not-accidental! So I played an acoustic bass to add that line to the whole final bit.

Hearing the sound of that, I knew the bass could be added at other parts as well. Slow and soft at first, but with quick thumping notes around the bridge.

That’s how the basic composition of this song was build. When it was done, I felt some high frequencies or drum-like features were lacking, but I didn’t know what to do yet. So I wrote it down as a to-do and went on.

Later, I came back to add a shaker for the high frequencies, as well as some subtle piano parts. I also had to re-do some guitar takes, or edit them, to make the beat a bit more steady.

Lastly, this is a weird song for me, so recording the vocals was a bit of a challenge. (The melody is quite a bit lower than I usually write them, and I also don’t write many upbeat songs.) Eventually, I decided to loosen up (do some exercises, stretch, dance while singing) and just sing loads of takes until I had a proper one.

5. Feast your Eyes on Me

By far the most complex song on this album. Also, to me, one of the strongest I’ve ever written (in a general sense). Which is why it became the title track.

(And I also want to write a story someday with this song as its main theme. So in case you wonder what the lyrics are about, it’s an impactful scene of that story I hope to write someday.)

This did mean it was also very complex to record and mix.

It has many tempo changes, key changes, unique plucked guitar patterns (which are too beautiful to merely place in the background), and some different sections that had to transition nicely into one another.

To really make the unique guitar stuff stand out, I decided to double track all of it. Yes, the whole plucking for the song was recorded twice, then one layer send to your left ear, the other to your right. It took quite some editing to make the recordings line up, as despite my best efforts, there will always be misaligned notes and slight differences between the recordings that are too big to overlook.

(I used the Spanish guitar here, because the fingerpicking was even harder to pin down with the steel strings, especially if I needed to record it twice in the exact same way. Also, I like the sound of it better, fits the foreboding theme/lyrics more. And it provides some variation against song #1, which has a similar guitar arrangement.)

Once done, I … actually destroyed some of that editing, as it was now aligned too perfectly. This made the overall sound smaller and more phase-y, which I didn’t like. So I had to purposely misalign many notes because it just sounded better.

On top of this is a simple arrangement of drawn-out strings, subtle piano, and basic chords from my steel string guitar.

I can call that a “pro tip”: combine a Spanish and Western guitar for the ultimate effect! Playing basic (open) chords on the Western provides great high frequencies, and the strumming (and pick noise) add a great groove/beat/tempo to the song. Adding fingerpicking and (closed) chords on the Spanish really fills in the missing frequencies and adds embellishment.

In the end, this is an amazing song (don’t know how it just almost completely popped into my head one day), and I think the execution sort of does it justice. Some extra drums, a volume boost here and there, a tempo increase/decrease, there are loads of tricks going on behind the scenes to guide you through the song and provide the highest “climaxes” at the right times. Hopefully, this will transfer to the listener, and they will enjoy the song as much as I do.

Some notes on “mastering”

Thus far, to my inexperienced brain, “mastering” the songs just meant “increasing their volume until the overall loudness is roughly at the level Spotify wants”

But there’s obviously much more to it. And when I compared the waveform of one of my songs, with the waveforms of popular songs, it confirmed a thought I already had: my songs had way more dynamic range, which means there were large parts that were way softer than other parts, while most professional songs are quite uniform (in terms of volume).

The result is that my songs feel more unbalanced and a bit more “dull” or “in the background”. Because, well, they are literally lower in volume, and our ear automatically places that more in the background. (Some dynamic range is obviously desirable, you don’t want all of it to be full volume all the time. But my songs had a range that was just confusing.)

So, this time, I tried to make it more uniform.

First, I just tried several compressors with a really hard cutoff (to get most of it on the same volume). Why? Using a single compressor with huge gain always sounds bad, so it’s better to use multiple after each other.

But this was horrible. It worked, yes, but because this compressor acts on the whole track, it basically creates a pumping, harsh, weird mess.

So I simply used a single compressor, and set its input gain to a volume that would make the whole track almost as loud as it can physically be. Then I used a relatively soft threshold and cutoff, to ensure we don’t create some aggressive effect on the track.

After all, we’re just interested in making the volume the right loudness and somewhat uniform, without destroying the original mix.

This worked much, much better.

Additionally, most audio services recommend leaving 1 dB of headroom on your track. (So the “loudness peak” is at -1dB, instead of 0 or close to it.)

Thus, I used a limiter at the end to set this ceiling of -1dB, and checked to make sure it wasn’t activated like all the time.

(Which would mean the limiter is like a scissor just cutting through the whole track, lowering volumes everywhere, which obviously sounds terrible. You just want it to kick in once in a while at the absolute loudest parts, as some sort of safety mechanism.)

This is a significant improvement, getting my songs closer to the professional level. Actually, it’s a bit stupid that I’ve ignored it all this time, as mastering literally affects all songs, globally. Of course it’s really important to do it right!

Additionally, doing this revealed some unbalanced parts in the songs. Now that they were louder, now that the general volumes slightly changed, some mistakes stood out that were just too silent before.

Conclusion

This album was written in a week. (While recording other music, I always get inspired and basically write my next few songs right then and there. So this album was created during the finalizing of the previous album.)

It was fully recorded in a week. (It took a few days longer to fully finish. Also because I lost some days due to other responsibilities.)

Compare that to the previous album that took several weeks to get the first version, then again several weeks to improve and fix all the mistakes I made in the first try. And those songs were less complex and the end result far less profesional than I think this album is.

It’s hard work, learning how to do this audio stuff well. But seeing progress, becoming more productive, creating something like this within a few weeks … it feels good.

I know it’s not perfect. It could all sound cleaner, more “radio ready”, more exciting, better balanced, etcetera. But I feel like it’s unavoidable that I’ll learn how to do that if I just keep making these albums.

So, onward to more music!