In Silence

I wrote another song. This is becoming a habit. Only this one’s a little different, because it’s just voices.

It’s called In Silence, and this post is about how it was written, arranged, and produced. Let’s start with the song (on Youtube, because it embeds so nicely in a Substack page).

As always, it’s also available on Spotify, Apple Music (and the usual streaming services) for your listening pleasure (or tolerance).

Solaria, the soprano singer here, is also the lead on most mothershout songs - Plainsong and Tell Me What You’re Looking For are really good examples of her voice.

The Origins

I really like writing vocal harmonies. I like it so much that in some recent songs, I’ve forced myself to write for a single voice, or maybe only two voices, so that I get more of a musical challenge. But this time I decided to be self-indulgent, because (a) life is short and (b) I wanted to see if I could write a song for voices. Musicians gonna musician.

My self-imposed rules for this song were simple:

• Four voices. Soprano, alto, tenor and bass.

• No instrumentation.

That’s it.

The Words

The world can be an unstable and scary place. Especially lately. I sometimes find myself awake in the night, thinking about many Bad Things happening. But I try to remember that nobody knows what’s really going to happen, that most of the things I have worried about never came true and that I should shift my imaginings to a better place. I tried to capture that dark-hours-of-the-night feeling in a song.

Which all sounds very planned, but I had to fit the words to the music, because it came first. A couple of years ago I wrote an orchestral (wordless) piece as an exercise. It started with the same long, slow chords that start In Silence, played by an entire string section. Like this:

They’re wonderful chords, so I decided to recycle them. As I played around with them on the piano, the words in silence, in darkness, in moments assembled themselves in my head, and the rest of it followed, more or less.

In silence
In darkness
In moments
When sleeping fades away like
Sunlight in winter
I lie awake unfolded by the night within

And all I see
A circus made of nothing more than memory
And flying free
A rolling cloud of dreams of what might be
Storm over silent sea

Unfounded
Unspoken
Unbounded
I’m broken by the past
Enclosed in
This turning
Uncounted and abandoned by the day
To come

But time will bring
A shifting of the streetlight, some imagining
And on and on
The rolling seas of dreams of what’s to come
Stormcloud moving on
Bringing another dawn
In darkness
In silence

The Artwork

It’s a picture from Pexels.com, by Cliford Mervil.

It fits the atmosphere & feeling that I wanted for the track. I also tried out some images of moonlight, but since the last track was called Moonglow, I thought that would be too much moon. I also played around with images of four candles (for four voices), because candle-light has a sort of midnight quality, but they also gave the artwork a vaguely religious or even Christmas vibe, which I definitely don’t want. This song is not in any way religious; if anything, it’s humanist.

It Don’t Mean A Thing If It Ain’t Got That Sing

There are lots of ways to approach writing music for four voices. If you’re a proper qualified musician or composer who went to music school, you’ll have studied the rules of counterpoint and part-writing. I didn’t go to music school (I’m more of an improper musician) and my book on counterpoint felt too much like homework, so I took the easier route of using the instruments I can play to figure out the harmony.

First, I worked out a basic arrangement of the chords on the guitar. Then I took those chords and spent a few days at the piano, playing around with four-note chord voicings until they flowed. The way that changing a single note can affect the quality of a chord still fascinates me after decades playing around with music. A wonderful way to lose time is to sit at a piano and play around with gorgeous jazz-harmony chords and voicings. Especially when at least slightly high; there are good reasons that jazz and cannabis go together well.

The two main challenges I have when writing for voices are to keep within the singers’ ranges, and to give each singer an interesting, singable line. Although the synthesized voices I use are capable of interval leaps that might challenge a human, I still try to write for humans. Maybe some day a four-voice ensemble might sing this piece for real.

Eventually, I settled on writing two interlocking parallel pairs of voices. The soprano and alto are one pair, the tenor and bass are the other. The two pairs come together to form four note chords, but there’s still some independence.

I wrote the top-line melody first; that’s the soprano voice. Next I added a simple part for the lowest, bass voice, leaving a big space between it and the soprano. Then I went back and added in alto and tenor lines; the middle voices. Because I’d left space for them, I had a lot of scope to play around with the notes they sung, so I could create more interesting chords and harmonic movements. Finally, I adjusted the bass line, adding some more movement (to keep it interesting for the singer) and dropping it out in a couple of places where I wanted the upper harmony to stand alone.

The four voices are all from Eclipsed Sounds, played using Dreamtonics’ Synthesizer V. The soprano is Solaria, the alto is Nyl, the tenor is Saros and the bass is Asterian.

Building With Voices

To turn my musical scribblings into a real track, I started with a piano version of the whole thing. It’s played on an electric piano, because that happened to be the setup I had when I started writing, and I liked the sound:

Once that was done, I imported it into Synthesizer V and set up each of the four voices to sing their lines. There are no words yet, so they just sing “la”:

At this point, the voices are all kinds of wrong. There’s way too much vibrato (especially on the bass, who sounds like he’s drunk) and they’re singing with too much emphasis. The bass line isn’t quite right. And, of course, they need to be singing the right words.

The first step was to enter all the words. Synthesizer V has a built-in dictionary that translates English (or Spanish, or Japanese, or Mandarin) into syllables. Useful, but it often doesn’t get it right, so I carefully edited individual syllables of the voices to get the right sounds. For instance, on the line and all I see, three of the voices chose to sing the first and more like und, which is how English speakers often pronounce it. Those initial “uh” sounds needed to be “ah” sounds.

With all the words done, I then worked through the whole piece, setting up the right voice modes for each line or word. For instance, at the start, I want them to be singing softly.

Now I had something much closer to the final sound, so I went back again through it all, and adjusted notes so that the voices fit together better. Writing for voices on a piano is great, but it’s only when I hear notes sung that I can tell whether they’re right.

Once all of that was done, they sounded like this:

That got me most of the performance I wanted, although they sound like they’re singing in a dusty, muffled cupboard. So it was time for audio processing. The first stage is subtle - I use a vocal leveller, which keeps each voice at more or less the same volume, and a dynamic equalizer, which gets rid of resonant, harsh frequencies:

Better, but there’s more to do. I added another stage of frequency equalization for each voice, finding the right place to boost so that all the voices could be heard separately, yet still blend together. And I added compression, which is a second way of levelling out audio so that it stays at the same general level:

The overall effect is pretty subtle, but I think that if you replace the last three audio clips, you should be able to hear the voices get clearer and more “separated”.

But they’re still in the cupboard. So I added a very small amount of “room reverb” to make it sound as though they’re singing in a medium-sized room:

We’re nearly there, but what I’m really after is the sound of a vocal quartet singing in a big open space, like a church or a hall. To get that, I add a second reverb, which models a big, echoing space. The problem with reverbs like this is that they tend to clutter up the vocals with echo, so I set up the reverb so that it’s quieter when the voices are singing, but comes back in between the words. This also gives the impression that you, the listener, are close to the singers, in the same space.

And that’s the final result. Four voices, that can be heard individually, but that also blend together, in a large open space.

Harmony Harmony Harmony

I took the basis of the harmony more or less directly from the orchestral piece that was the basis for In Silence. It’s really just a series of movements from V to IV to I; since this is in G, the chords are D, C and G. Extra tension comes from moving the root notes down before the rest of the chord, so we get D D/C C and C C/G G. There’s a lovely sorrowful sound to those movements, which comes from #11 notes. For example, keeping the D and F# from D in the D/C chord gives it the sound of a C(9,#11). I love how that F# sounds over a C. In fact, it also shows up in Plainsong (in the bridge). I admit that it’s a direct steal from the largo of Dvořák‘s New World Symphony. There’s an arrangement of it for a brass band that was around when I was young; it’s at 1:08 in this Spotify link, played by the Grimethorpe Colliery Band. I took the feeling of those #11 and 9th notes from that section.

As the song continues, the same progression, V to IV to I is repeated, but with increasing amounts of variation, with more passing chords added to descend from IV to I. With four voices I can use extended chords with sevenths, ninths and elevenths, to vary the chords through the song. But underneath, it’s just good old functional harmony. And it ends on a plagal cadence because I grew up in Wales, where the people sing as easily as breathing, and that cadence feels like closure to me.

If you liked this, you might also like a whole other bunch of my posts about songs. Or you could even subscribe. All the cool kids are doing it.