For better or worse most arranging is done on a computer, which means that composers rely increasingly on how things sound on playback. While there’s nothing wrong with this approach (it’s what I do), the computer has a way of making certain things sound better in playback than they will in real life. When I get asked to review arrangements, these are the most common mistakes that I see, and I have a feeling the computer might be to blame.
I always encourage young arrangers to think about the type of sound they’re trying to create. Tight, close voicings have a different sound than big, tall chords and depending on the range of the melody one or the other might be the better choice. But moving back and forth between one and the other in the same phrase can create an uneven sound. An even bigger problem is when there is an unusually large gap (greater than a sixth) between one of the inner voices—for example, if you had the soprano and alto notes close together, the tenor and bass notes close together, and a large gap in between.
Ignoring the Text
Always write your vocals to fit the lyrics, not the other way around. As a general rule, the more divisi you have, the harder it will be to understand the words—especially with a rhythmic or busy text. Follow the rhythms of natural word stress whenever you can and remember that the “sound” of a word lends almost as much to the music as a word itself.
Use the characteristics of vowels to your advantage as much as you can. An “AH” or an “OH” vowel will be ideal for tall chords. An “OO” vowel is harder to sing loud than the others. Notes in extreme high or low registers can be harder to tune on an “EE” vowel. None of these are hard and fast rules—my point is that when you have the choice you shouldn’t write anything difficult on accident.
Unintended Parallel Motion
Parallel 5ths get a bad rap in music theory classes but there are LOTS of examples—especially contemporary music—when parallel fifths are used to great effect. The problem is when they’re written accidentally (nothing you write should be by accident but that’s a different conversation). Still, parallel fifths are hard to sing in tune, so unless you’ve got a compelling reason it’s still a good rule of thumb to avoid them.
That being said, parallel fifths can be really hard to avoid when making arrangements of popular songs if they’re built into the relationship between the melody and the chord progression. Generally speaking, the best way to use them is to create a rising or falling effect with all of the voice parts moving together. In this case the overall effect prevents the sound of that interval from becoming jarring. Parallel octaves are always bad unless it’s the melody or part of a stacked triad. Parallel fourths can also sound weird in certain scenarios but are mostly fine.
Thinking Vertically Instead of Horizontally
This one is hard to define, but I know it when I see it. Sometimes (especially when composing at the piano) there is a tendency to fill out every note of every chord which can result in some awkward voice leading. If there’s ever a conflict between the flow of the line and the notes needed to fill a chord it’s usually better to focus on singability. It’s a good idea to sing through parts and check for things that feel uncomfortable.
Doubling the Third of the Chord
There is just something about the overtone series or the laws of physics or maybe a little bit of dark magic that makes chords sound weaker when the third of the chord is doubled. Even in divisi situations when there are 6-8 vocal parts it’s better to not double the third unless it’s part of a melody line and you can’t get around it. But certainly, on sustained notes the best practice would be to avoid it. This rule should be strictly observed in first inversion chords (when the third of the chord is the lowest note).