Help with multiple voice harmonization

Question

I recently started composing using software, before that I only used guitar and memory, so this opened up lots of new possibilities with voices, intervals, ranges, etc. I started composing a chorale based on a children's song's melody, which is EFEFGG x2 AGAGAFF GFGFEE in a rhythm of 4 eighths and 2 quarters.

In the third part my idea was to have the 2 altos, one doing the quarter note pitches (G, G, F, E) and the other doing a chromatic ascend (B, C, C#, D) in half notes while the soprano alternates the eighth note pitches in quarters. I want the bass and tenor to have several voices constantly overlapping each other and creating a wall of sounds with m2,M2,m7 and M7's, or just sharp dissonances, flowing in and out of whatever seems appropriate.

I've never harmonized >2 voices, so I don't know how to approach this. In the current state (sheet music and flac file {+embedded player}) in part III bars 1 and 5 are examples of where it's decent, the others kinda of work on their own, as in, tenor and bass separated, but not tog

Answer 1

This is quite opinion-based, so these are not universal rules, but here are some principles that match how I think about harmony. I apologise if some/all are either too obvious or completely nonsensical!

1. General principles, and complexity

   1. It's all about how easily the listener's brain can intuitively understand the chord progression. Too easily understood and it's boring, but too confusing and it just dissolves into a pile of notes instead of music. The optimum amount of challenge/complexity is a matter of personal preference (and the listener's skill/practice at hearing harmony), but it's usually good to keep the level consistent to keep it interesting but not impossible.

   2. The "complexity" a progression sounds is a result of both the chords themselves, and also how smooth/jumpy the transitions are. You can get away with more discord if the transitions are smooth, and you can get away with a more drastic transition if the chord you're transitioning to is more simple/harmonious.

2. Constructing chords

   1. The most common chords (simple to understand) are majors and minors. Next-simplest are variations/additions (e.g. 2nds/7ths), which can be viewed as borrowing notes from other chords. Using one of these extra notes hints towards some chord that uses it, so it's smoothest to use notes that are also used in your other chords.

   2. Evenly-spaced notes in a chord will sound rich instead of sparse. It is better to have the notes slightly crowded together at the top end than the bottom end (because small intervals at low pitches sound "muddy").

   3. Think about the "centre of gravity" of your chord in terms of pitch. For example, if your soprano line is jumping up or down, then you can counterbalance that motion using the tenors/basses to keep things neutral (a good default), or deliberately move the entire chord to a different register for effect - but it's good to know which one you're doing.

   4. If you are going to have multiple parts singing the same note (or octaves), then it matters which note it is. Doubling up the root note of the chord will sound the most balanced, doubling up the fifth is OK, and doubling up other notes can sound less "stable". This is basically a way to hint to the listener which note is most important, so their brain can start forming a chord around it more easily.

3. Transitions

   1. If your next chord has a clear root note, then your bassline can jump to this root note from any distance and it won't sound awkward. If your bassline is not going to the root note (because you're using an inversion, or your chord is too weird to have a clear root) then it is best for the bass part to not jump a long way, or the listener will have trouble tracking the change.

   2. If two parts are an octave or a perfect fourth/fifth apart, moving them in parallel can sound clunky (e.g. if your tenors and basses are an octave apart, then they should not also be an octave apart next time they move). Octaves and perfect fourths/fifths are (mathematically) the simplest interval, and so moving them together can make the transition sound too simple.

   3. Having some notes kept constant makes a transition smoother. The shared notes keep the transition more comprehensible even if the chords themselves are bizarre. (Remember - the smoother your transitions, the more discord you can get away with - see point #1.2.) For chords with few/no notes in common, you can artificially produce this by moving some notes early or late (which can also add some extra discord/interest for the short period until it is resolved).

Hopefully, these guidelines don't limit the actual chord progressions you use - instead they're about how to spread out the notes from your chords across multiple parts in a way that makes sense to the listener.

While I personally work on the piano/keyboard, and I find it very convenient for planning out harmonies, I don't think it's necessary. If you work out what chords you want to play (e.g. using the guitar), then you can look at the notes that are actually in that chord, then lay the notes from the chord out as a separate step (and that's how I used to do my harmony work at school). :)

Answer 2

Do you have access to a music keyboard, either that can play its own sounds or that can generate sounds via your software? I find trying out different sequences and making notes to myself about the ones that 'work' is the best approach to a task like this. Since my primary instrument is piano, that's where I work it out.

Sounds to me like you're on the right track so far. I do think some study of 16th century counterpoint (Fuchs) might help, though you need to realize that that is built around avoiding or resolving dissonances rather than emphasizing them, which it sounds like you would like to do. (I don't think 18th-century style would help though.)

The tune itself is simple, and while you say you want to avoid the simple in working out your composition, I think you need to make it at least SEEM simple in order to maintain the overall atmosphere. What follows is NOT advice: you need to work it out your own way, in your own musical language, in your own style. But what I would do if it were me would be to make the accompanying rhythm and harmony gradually more and more complex, and more and more dissonant, until a climax is reached, and then go back the other way, until more and more simplicity is reached, ending as simple as the tune itself. But that's just me. You need to be you.

If you're into classical at all, I'd suggest studying the works of William Walton and William Schuman, as they both excel at working smoothly from assonant to dissonant and vice versa.