Notes on "Don't Forget Me"
With the release of my new EP called “Don’t Forget Me”, it’s time for another album diary!
I’ll explain the process behind the development of this EP. Usually it talks a bit about the ideas, themes, vision … and a lot about all the recording/mixing mistakes I made along the way and how I solved issues.
What’s the idea?
After doing “100% live takes” for all six songs in the last EP, I learned a few things:
- Yes, this captures a much more lively and natural vocal (and, to a lesser extent, guitar playing) performance
- But any editing afterwards is pretty much impossible. I can only stitch together the best parts of each take into a best fnial take, with some care taken to make the stitches seamless. (Even tiny edits, or tuning to the vocal, is impossible.)
- When positioning the microphones, it’s mostly about where they point. Really point the guitar mics away from the voice, and the voice mic away from the guitar, to get good separation. Do whatever is necessary to set it up this way. (The carton box that was used to send me my new headphone was big enough for me to use as a sort of low table to get everything in its proper place :p)
- It works wonders for more subtle/intimate songs. But for bigger songs, you’re just fighting against all the downsides of this method. (For example)
(What are “100% live takes”? It just means that I play/sing the song in its entirety and record that with three mics: two on the guitar, one on the voice. That live recording, of the full song, then becomes the basic element of the mix. It’s basically 90% done by the time I have that take. I did it because I’m a singer/songwriter, used to playing while singing, and I wanted to see if this led to better performances.)
For this EP, I wanted to take the next step and experiment further.
Two of the songs are driven by piano, meaning I’ll just use my previous method (of recording everything seperately). I do not currently own a live piano setup that I can record professionally.
One song is a rock song, with full band/instrumentation, so it doesn’t benefit from this method.
One song is really tricky (with tempo, timing, etcetera). I’ll probably do a live take, but with headphones on and a metronome to follow.
The others will be done with a live take again.
All in all, I’m hoping this EP …
- Introduces even more variation into my discography
- Gives me more practice with piano-focused songs and a sort of “hybrid” recording method
- Will sound even more professional and exciting than the previous ones.
About the songs
If I have something special to say about a song, this is where I’ll do it!
1. Don’t Forget Me
This is actually what inspired this album, its name, its theme.
I stumbled upon a YouTube video about the same sentiment where she, at some point, just says “don’t forget me, don’t forget me, don’t forget me, …”
From that, this song grew. All other songs were already (mostly) written in terms of melody and chords, but not so much lyrics, so I formed those around this theme.
I wanted to experiment with something new: making the main instrumentation a voice loop. So I recorded the phrase “don’t forget me” on as many different notes I could, combined those layers into a loop, and put that on the background of the whole song.
(I did a basic autotune and quantize/timing pass over all the vocals. As you’ll constantly hear all those layers saying the same thing, it’s not nice if there’s even one voice singing something slightly wrong. Once I was able to fit it perfectly into a few bars, I could just copy-paste to repeat. Then I added a generous amount of delay and reverb, and used an EQ to cut away the frequencies that the main voice should use to grab attention and stay forward in the mix.)
The verse melody doesn’t want a guitar, it wants a dreamy piano/synth background with long held notes. So I added that throughout the song. (Learned how to play with synthesizers for the first time!)
The instruments play nothing special: just the chord tones to form “dreamy” chords (such as the Maj7 variant). Add tons of reverb/delay to get near the sound I want.
I added a soft piano, a soft cello, and then went searching for a bit more ear candy than that. (It already sounded really full, atmospheric, big now … but not necessarily interesting, as it was just basic drawn-out notes.)
I was searching for something high, a bit twinkly, but didn’t want to use a harp or high piano again. I ended up on a flute with quite extra effects on it to make it less harsh. I also had a plugin for a “(Children’s) Music Box” (which I used for a game I made a year ago) that seemed to fit. Again, loads of delay and reverb.
And then I just had to record a strong vocal on top. Which isn’t easy, especially not in the bridge, as the notes are quite high and breath is at a premium. That’s also why I chose to record the parts separately for this song.
In the end, I did feel some percussion and low frequencies were missing, so I added a low thumping (foreboding?) drum in many parts. That was the last thing I did: besides some tweaks, fixes, basic plugins, the song was done by this point.
2. Could I Be Happy Here
In my opinion the “star song” of this EP.
Really great melodies, tempo/key changes, lots of variation and high points, and just a complete embodiment of its theme.
That also put some pressure on this song to became as good as I thought it should be. I tried to follow the guiding light of The Lumineers: I thought this really sounded like a song they could write.
As such, I used my Western guitar and kept the strumming to more of a “forceful thumping” than the smooth plucking I usually (subconsciously) go far. I added a tambourine early. I experimented with foot stomping/clapping/multiple yelling voices.
Mostly the multiple layers of voices stuck.
When recording such a song, you really miss some low and high frequencies. It starts to feel empty. Like something’s missing once you reach your big final chorus/bridge/climax.
I used the percussion to add the high frequencies. It’s what you usually do. And it felt wrong to add another instrument, like a piano, for that.
The low frequencies were harder. Other songs had already used instruments/techniques I’d usually use. I wanted to try something new. So let’s look at the Lumineers: they use low cello notes for that. Let’s do that as well!
Then the trick was: getting a good, strong vocal recording and a vibey, rhythmic guitar to match it well. And edit until it all flows properly.
Because man is this song difficult to manage. Why do I keep writing stuff like this?
Fun Remark: When I recorded the banjo, something just sounded off. Couldn’t figure it out. I was sure everything was tuned correctly. Later that day, I listened again on different devices, experimented with slightly changing the tuning of the full recording … and discovered the banjo sounds “correct” when detuned by -5 to -10 cents! Apparently, banjos are supposed to be slightly out of tune to sound well with other instruments. Did not know that.
Fun Remark: Recording the “cajon”, I wasn’t sure if it was better to mic the front (where you hit it) or the back (where the sound hole is). So I placed a mic on both ends to test. But I was met with a happy accident: when I recorded both, and played them together … it actually sounded great! I had accidentally found a setup that provided a nice stereo image of the cajon, without phasing issues, so I made sure not to move anything anymore and used that setup for all cajon recordings.
3. Out of Arms
This is, like, three songs pushed into one. At the same time, I put them together for a reason: they fit well and sound similar.
So I just approached it as one song, not three distinct parts that I should record/mix/view seperately. It feels similar in vibe to song #2, but different enough, more melodical and upbeat.
I bought shakers for this one. I also allowed myself more instruments just to add that sense of excitement: an electric guitar moving all over the place, a trumpet.
At the same time, the core of this song is just a recording of me singing while playing, going through the song. I thought that was the best way.
4. Let Your Mind in My Mind
This song gave me the toughest time. I wasn’t sure if I should keep it extremely intimate (that felt fitting for the verses, just me and the guitar, nothing else), or approach it more like a pop-song (which felt more fitting for the chorus/bridge).
In the end, I tried both. I did a few live takes through the whole song. I recorded parts of it seperately, like my older songs. Then I checked which I liked more.
You never guessed it: I used parts of both. Many extra instruments come in during the choruses and bridge, but are completely left out (or much simpler/softer) during all other parts, in which a soft recording of me and my Spanish guitar takes over.
5. Trip Around the World
I can’t have an album without its share of angry songs.
As soon as I wrote the melody for this song, it just had to be a rock song. I also noticed I hadn’t done solos/instrumental parts in a while, so I forbade myself from writing a bridge (or more verses/choruses), and kept lots of space for a solo.
Being a “full band rock song”, the process is a bit samey: record a basic drum pattern that gets more complex/full as time goes on, add a bass that just plays the important chord notes, slam on my electric guitar (double track that!), record multiple layers of vocals and make them sound a bit more aggressive.
I really wanted to go further than that, but my experience/motivation/equipment is a bit lacking.
Oh yeah, a rock organ is also nice. Especially when you use it to swell/rise into choruses or solos. Also, if you drop out the bass for a few bars, that explosion into a chorus feels even better. And I always make the bass guitar play more licks as time goes on. (Usually just walking between the notes, or a quick arpeggio/alternating pattern.)
In that sense, full band songs are perhaps easier, because there’s a rather easy predefined formula that just sounds great when you follow it. At the same time, it quickly sounds like all other rock songs and it’s hard to really excel here.
6. A Song for a Beautiful City
Sometimes, I lie in bed, late at night, darkness all around, just listening to sounds from the street. Sometimes drunk people are singing, cycling by, talking about something stupid. Sometimes you hear a party going on far away. Sometimes, especially when its Winter and its perhaps snowy, I just look out at the pretty lights.
That’s where this song comes from. It just popped into my head, laying in bed, in its entirety. (Of course, I had to write the full lyrics, refine parts of the melody, etcetera. But it was 70%-80% formed by the time I fell asleep.)
It’s a musical tribute to bustling cities, adventure in the night, pretty lights. I keep thinking of the song with a sort of “romantic night in Paris” vibe.
As such, I tried to capture that in the recording. An accordion, harmonica, subtle violins are present. The piano is more dreamy than that it really adds notes.
And the voice isn’t my regular singing voice, but more what I’d imagine a street singer to sound like, who is actually in Paris and seeing this every night.
Maybe that was a mistake. Can’t know until I’ve made the song and released it. But I think it will be perfect.
The song isn’t much more than that. More instruments would just ruin it. The melody doesn’t really change. It’s more soft, whimsical. As I said: a sort of tribute or lullaby one might sing in the background.
7. As The Banjo Plays (Bonus)
I often write completely new songs while working on an album. Most of them move towards the “a future album”, but this one was simple enough to include as a bonus track here.
As such, the main melody came to me as I was learning how to tune/play the banjo for those other songs.
The next day, I expanded upon it, wrote lyrics, learnt how to play it on a regular guitar. A few days later, after some practice, I sat down and played it, simply recording myself with a single microphone.
That’s all there is to say about this one, I think :)
Lessons Learned
1. About using VSTs
Most VSTs need ridiculous amounts of reverb/softening to sound realistic.
(A VST is a virtual instrument, which you can load into a track so you can play it using a keyboard plugged into your computer. At least, that’s how I use them.)
Most of these are recorded as dry and loud as possible. This means that you get their pure sound, as if your ear was only a few centimeters away from the instrument. Which basically always sounds harsh, lifeless, too loud.
With older albums, I tried to limit the use of VSTs for that reason, or lowered their volume drastically and only used them as background elements.
But I’ve learned that I was simply too insecure. I simply needed to become more extreme with my modifications.
A reverb should do a lot of work. Then you can EQ out any frequencies that other instruments want, or the mid-frequencies that demand attention. Sometimes a Phaser/Chorus/Doubler plugin is needed to “spread” the sound a bit more across the spectrum.
But the idea is that you can’t just combine a VST (of, say, a trumpet) with a real voice and guitar recording, and expect them to mesh well.
Bonus Tip: Most (professional) VSTs automatically shape the sound based on the velocity with which you press the keys. So if I play very aggressively on my keyboard, the trumpet will sound like someone blowing the shit out of it, but if I play softly, the trumpet will also sound more softly. Unless your VST is playing a solo and should be the absolute center of attention, you want the softer/quieter/more intimate sound. So just bring down the velocity, either while playing or afterwards, to be more near 50%, and it makes a huge difference.
2. Shakers are great
I finally decided to spend some money on shakers, a tambourine, some other small musical toys.
Yes, you can replicate it by shaking something else that makes a similar sound. Or using a VST again. But it’s a lot of work to get a sound that’s not great.
It’s just better to buy the real deal (shakers are like really cheap) and record them with a microphone, anytime you need it.
Once you start to listen for it, you hear these instruments everywhere, especially in the type of music I make. (Folk, singer-songwriter, more acoustic or guitar-focused, etcetera.) They can be the replacement for drums, but they are also often used more subtly. (Only at the start of each new segment, only far in the background to add some “twinkle”.)
There’s a reason why drums (or the beat) are the core element of every popular song. You need the frequencies they produce and the groove they imply to simply make everything better.
3. It really is all about automation
The more music I make, the more time I spent automating.
I find it faster than trying to put a compressor on everything and trying to find the ideal settings. It’s very rare that an instrument can just be equally loud in all parts of the song. Usually verses are much quieter. You want that contrast, as it adds excitement and a cycle of tension<->release.
So lots of time is spent automating almost every instrument. Louder in chorus. Fade in/fade out at the right times. If I hear notes jump out or fall away, make a small tweak.
I guess it comes with experience. I would never have done this with earlier albums, mostly because I just didn’t listen for it. But now I always check if I can hear all the notes the instrument is supposed to bring to the table, instead of just … accepting that an instrument will sometimes stick out and sometimes fall away (due to whatever reasons).
In fact, I noticed I switched my time management from getting/playing the perfect performance to getting something decent and automating afterwards.
4. More reverb types
Yes, I was vaguely aware that there were different reverb types, but didn’t know what they were, why you’d use them, where to find them, etcetera. In fact, I couldn’t really hear the difference when I was just starting out.
Now that I have more experience and more “ear training”, I thought I’d consider using different reverbs, experiment more with that. I quickly found someone listing the most “conventional” pairings:
- Guitar = spring
- Piano = hall
- Drums = room (and overheads = plate)
- Vocals = plate
As a start, I decided to just follow this rule.
Yes, plate reverb on vocals sounds awesome. (In addition to a generous pre-delay. On more busy songs, though, I’d prefer adding a subtly delay to the main vocal over a reverb.) Less “muddy”. It allows vocals to stay at the front.
Room reverb on drums is also better. Other reverbs tend to make them so “washed out” that you might even completely miss there are drums in a song.
Besides that … I find it hard to really hear the differences. Perhaps that’s just inexperience, my ears not being well-trained yet. Perhaps it’s because my piano is a digitally sampled piano (played on my plastic MIDI keyboard), which obviously doesn’t sound/work the same as a real piano, recorded in a live room.
Anyway, I did realize that I’ve learned to apply reverb more and more generously … but at the same time try to hear it less and less.
This album has loads of reverbs on all the different instruments, tracks, etcetera. But if you compare this album to my first, you’ll see it doesn’t sound like that. My first album sounds quite muddy, washed-out, as if everything is in a room that’s too large. I think this album has a much better balance and application of reverbs, simply accomplished by tweaking parameters and volumes with more detail, and automation of the reverb.
Conclusion
Another album. Even more variety to the songs, to the techniques I tried.
I really focused on comparing it to professional tracks this time. (After all the previous albums, the basics are pretty solid. I know where to point a mic, how to add more instruments and play them on my keyboard, I know how to make things “decent” with plugins and volume management. So we’re looking at the next step.)
Tried to force myself to try different things, really experiment and keep recording/tweaking, until I got closer to that amazing professional sound.
And, of course, bought a few small additions to the studio which ended up being a great idea, as well as tried many new VSTs/plugins/effect chains.
(For example, my DAW comes with Ampire, which has a handful of electric guitar amps built-in. I used that for all electric guitar work on all previous albums. Only now did I realize that I could obviously download/shape other sounds, and only now was I ready to try that and actually use it. The downside to all this is that my electric guitar is a very cheap, old, broken piece of equipment that detunes too quickly to record any long takes with it. So that’s one of the clear upgrades in the future.)
Hopefully all this effort shows in the final product. As I said, I think the second song is the star song and I really hope it ended up as great as I think it should be.
Thanks for reading, until the next album,
El Troubadour