Why do some arrangements of notes make a good melody and some don't?

Question

A note by itself is not always particularly pleasant to hear. Why do some specific combinations of notes make a good, pleasant melody?

Why are some people good at coming up with such pleasing arrangements?

What skills are being used in that process?

Answer 1

Some notes sound good together. This is an example of what we call consonance. Some notes do not sound good together. We call that dissonance.

In simple terms, certain notes blend well together because of the way the sonic frequencies merge together and complement one another. Our brains will instinctively have a desire to gravitate towards complementary frequencies that will blend together to form pleasing sounds. The relationship between the sonic frequency of two notes is described in music theory as an "interval" which is how far apart the sonic frequencies are - commonly measured in what we call semitones (with one semitone being the smallest step in a Western Music chromatic scale).

Different sounds produce wave forms in different frequencies. A particular note will produce a particular and unique sound print based on how fast the waves move up and down which is measured as “frequency”. The mathematical relation of these frequencies to one another, account for the fact that some sets or groups of notes are harmonious with one another (sound good together) while others sound dis-harmonious (don’t seem to go together).

To further expand from math to biology and physics, we need to understand that a sound is heard because of sound waves which travel through the air to our ears. Sound waves are created by vibrations and these vibrations are detected by our ear drums or more particularly - the basilar membrane in our inner ear.

When two notes are consonant and the peaks of their sound waves (frequencies) blend together harmoniously, the corresponding vibrations in the basilar membrane are balanced and pleasing to the “processing center” in our brain. If the two notes are dissonant because their frequencies overlap instead of blending together, they create an uneven (offbeat) vibration inside our ear and the brain feels unsettled.

Here is an example to help illustrate the tendency of our brain to prefer an evenly spaced rhythmic flow. Imagine walking down a path on square pavers (like stepping stones) that are evenly spaced about two and a half feet apart. Your pace is steady and even and you don’t even have to think about it. You could say your gait is natural. Now imagine walking on a similar path – only now the pavers are irregularly spaced. Maybe one foot apart, then three feet apart then two feet apart then two and a half feet, then one foot 8 inches – you get the picture.

You don’t like this because there is not a natural even flow – like there is when you walk or run down the sidewalk (unless you are trying to avoid the cracks). Two notes that blend are like the evenly spaced pavers. Maybe one wave hits every five feet (every second block) and the other hits every two and a half feet(every block), but they come together and create a harmonious flow of motion. The uneven blocks are like two notes that never blend together.

If you have ever rolled two windows down at the same time in a moving car you may have experienced the effect of air waves that collided together and created an unpleasant vibration. So in simplified terms, two notes who’s sonic frequencies do not blend together, will create an unpleasant irregular vibration in the inner ear. And our brain does not like it.

Music in Western Culture is divided into scales and each of these scales is commonly divided into seven notes which makes it a “diatonic scale”. Any given diatonic scale is based on subdividing an octave into seven notes. An octave is the note that is exactly double the frequency or half of the frequency of the starting note. From C to C is an octave and the ratio of the sound waves from the root note and the corresponding octave note is 2:1 The octave is the most consonant sounding interval and is common to music from every culture.

To over simplify it, the notion of dividing the octave into seven notes is based on mathematical principles that determine that dividing an octave into seven notes yields the most harmonically pleasing ratios between the intervals that can be formed with seven divisions.

Hope that answers your first question "why some specific combinations of notes make a good, pleasant melody and others don't?"

Now to answer your question about what skills contribute to the ability to write a good melody and expand into how to go about writing a pleasant melody.

To compose a pleasing sounding melody, you must first determine what key (or what scale) you want to play the song in. Then you will know what seven notes you can use in your melody. They will be the seven notes in the key or scale you are composing in. Of course if your melody spans two octaves you have 14 notes but with only 7 different names.

The skills to compose a good melody can come from a study of basic music theory to learn about chord progressions, modes and keys, as well as just listening to or performing music on an instrument. The more involved you get with music that features melodies similar to what you want to write, the easier it will be to hear melodies in your head. Learning to play these songs on an instrument will ingrain a deeper understanding of how the melodies on these songs fit together.

There are also good books on composing music and on-line newsletters and blogs. It is beyond the scope of this site to outline everything you need to know to compose great melodies - but maybe we can get you going in the right direction. See below.

For a more in depth discussion on the approach I take to write good melodies - click this link: composing melodies the easy way

Answer 2

If your question is about coming up with good melodies, then I think there are a few starters.

1. Out of the 12 notes in an octave, any permutation and combination of notes can generate music. However, there are pre-defines scales(in western) or Ragas(in indian classical) that define a specific set of notes. If you limit yourselves to those notes, then there is a chance that the resultant product may be better than the one where you used all the 12 notes. So one needs to have a grasp of the scales or Ragas to compose using them.

2. A knowledge of western harmony will also help. Imagine three notes that when played together would make chord. These notes, when played individually, can be a melody too. You need to know those set of notes to compose using them.

There are a multitute of such techniques and experienced composers may be able to tell you more. In a nutshell, coming up with melodies takes a good deal of practice, imagination and pure emotion(Yanni). Some people can pull melodies out of thin air, whereas others need a great amount of effort.

Answer 3

This is a great question with a VERY broad answer and a lot of opinions on the matter, but I'll sum up my view:

When we talk about how notes work together, we can discuss them in two dimensions: the harmony (all the sounding notes at any sample point in time; the "vertical" dimension) and the melody (A rhythmic sequence of notes across time; the "horizontal" dimension).

Most musicians decide what notes to put in these dimensions based on various scales, which are collections of notes. The C major scale, for example, consists of C, D, E, F, G, A, and B; or all the white keys on the piano. Because of the varying distances between notes in the scale (some have black keys between them, some don't, etc.), different notes take on various levels of importance. What constitutes harmony is a bit out of the scope of this answer, but melody is certainly worth exploring.

The best melodies are the ones that have embedded themselves in the fabric of society by virtue of their structure. Let's look at "Happy Birthday", in the key of C major. Note that the name of the scale is the note that sounds the most stable, which means it is a good note to end on. I have included the name of the note in parentheses after each syllable:

      "Hap(G)py(G) birth(A)day(G) to(C) you(B)" 

Try singing just this part to yourself (don't worry about what key you may or may not be in, that does not matter, just sing this part of the song as you have heard it belted out at many a birthday party (: ). Notice how if you stop on "you" the song sounds incomplete? That's because you have stopped on the "leading tone", B, which REALLY wants to go to our stable C, which is just one note away. Let's see how the song takes care of this in the next line:

      "Hap(G)py(G) birth(A)day(G) to(D) you(C)"

Now, instead of going to C on "to", we go up to D. D is also one note away from C (but not as close as B; if you look at a piano you'll see there is no black key between B and C but there is one, C#, between C and D), and this time we get to end on C. Try singing the first two lines of happy birthday; notice how much more conclusive it sounds now?

Now, let's see what the melody does to keep itself interesting on the last two lines. Here is the whole melody:

    "Hap(G)py(G) birth(A)day(G) to(C) you(B)"

    "Hap(G)py(G) birth(A)day(G) to(D) you(C)"

    "Hap(G)py(G) birth(G)day(E) dear(C) some(B) name(A)"

    "Hap(F)py(F) birth(E)day(C) to(D) you(C)"

On "birth" in line 3 we jump up an octave (to a higher G), which serves to freshen up the melody. We're halfway through the melody, and this jump grabs the listener's attention. This is a very common trait in the best melodies: the highest note occurs ONCE somewhere between halfway and 3/4 of the way through the melody. This allows tension to build and then gradually fall off.

Remember those charts from elementary school about plots consisting of an exposition, rising action, climax, falling action, and resolution? Well, the best melodies have that SAME structure!

After we go down to an A at the end of line 3, we wrap up the song on the last line by, once again, going from D back to our stable C. While there are many other factors that determine the quality of a melody, like its contour and rhythm, it is this basic structure of tension/climax/resolution that makes the best melodies work. While this example is rooted in Western musical traditions, these basic qualities are found in melodies across many cultures.

As for the last part of the question, some people may be musically gifted, but anyone who is not completely tone-deaf can pick up basic musical skills. The skills to arrange and compose music are difficult to exactly nail down, however there is a saying in music that composition is 5% inspiration and 95% perspiration. If you are willing to put in the effort to learn, then you will very quickly pick up the skills you need and will begin to intuitively see the patterns in music that make it tick :)

Answer 4

Survival in the wild, and pattern-matching
Why do we have ears in the first place? It's for survival. The ears (and the rest of the auditory system) are supposed to tell us what's going on in the world. Rockin' Cowboy mentioned the Basilar Membrane, which is able to do a real-time spectral analysis of the sound. However, a full spectrogram would be a lot of information to process and we're busy doing other stuff (seeing, eating, moving). So the auditory system is a big pattern-matching machine - it does some clever stuff to match the different frequencies it's receiving to each other and try to divine some information about 'things that are happening'.

The harmonic series
Sound is vibration, and in nature, if something is vibrating at a frequency, it will also tend be vibrating at multiples of that frequency. So one of the things the ear does is look for frequencies that are multiples of each other, and if it finds them, it tells us that this is 'one sound' - one 'thing' in the world that is vibrating. So if the ear hears vibrations at 1000Hz, 2000Hz, 3000Hz... all at the same time, it sees that as one 'note'. And as you say, it's not particularly surprising or special - it's just the ear hearing what it's expecting to.

Showtime!
What happens, though, if we play two notes that are related in frequency? Say, one at 1000Hz (with harmonics at 1000Hz, 2000Hz, 3000Hz...) and another at 1500Hz (with harmonics at 1500Hz, 3000Hz, 4500Hz...)? All of a sudden the ear's pattern-matching mechanism goes a little bit wild. Wow, what's happening? It's two sounds (sets of related harmonics), but they also kind of merge into one (as the frequencies of the harmonics of both are closely related to each other). It's a bit like the aural equivalent of an optical illusion.

If we're playing these notes at the same time, then we're talking about harmony rather than melody. But as we have memory and the ability to 'pattern-match' things that happen across time, similar principles apply even when we are talking about two notes that happen one after another.

Scales
So, if we want to come up with interesting combinations of notes, why not come up with a set of notes that have interesting frequency relationships, so we can choose to play them one after another to make up a melody? Great idea - and that's what a scale is.

Try having a play on http://labs.dinahmoe.com/plink/. Can you see how it pretty much always sounds 'good'? That's because it's using a pentatonic scale that has simple frequency relationships between all the notes (http://www.phy.mtu.edu/~suits/pentatonic.html). It's called 'pentatonic' because there are 5 notes to each octave.

A very popular scale in western music is the Diatonic scale. This allows for many simple frequency relationship between notes, but also some that are not so nice. This gives rise to the idea of 'tension' and 'resolution' - the way that you can create interesting melodies by mixing nice consonant intervals (frequency gaps with simple ratio) with more harsh, dissonant ones.

Another scale that would be worth reading about if you are interested is the 12-tone even tempered scale - another western scale that is a clever way of squashing a few more notes into the diatonic scale such that you get 12 evenly-spaced notes in an octave.

And there are many other scales in use around the world (some of them seem designed to work with instruments that don't have harmonics that are integer multiples of a base-or fundamental- frequency : the Gamelan and its associated scales is an example of this.)

And there's more
Of course our ears can match patterns other than frequency ratios. We can perceive rhythms - ratios in time between sounds. Again, we perceive regular rhythms in a special way - you could again consider this might go back to nature as they are found in things like walking, breathing, dripping, chewing, heartbeats, many animal songs, and so on. But the brain will even match patterns that aren't 'natural' (think of letters, and words, for example) and rhythms are just that: patterns in time.

Also, remember the way that the ear likes to try to correspond frequencies with a 'base' frequency at the bottom of a harmonic series? It likes to do that with whole songs as well, so there is a 'root' note that it's satisfying for the melody to return to.

And the brain will find satisfaction in recognising higher-level patterns, too. Think of a song like London's Burning. There are two groups of two notes followed by two higher notes, and then two of the same note, then a stepwise downward pattern occurs twice as the melody goes back down to the root. And there is melodic correspondence with the rhythm - the first syllable of 'burning' is the strong beat of the rhythm, and it's no coincidence that that's also the root note.

Music doesn't exist in a vacuum - we experience it while we are doing and seeing other things, and we may tend to enjoy melodies that we associate with other good times or good things. We may love our national anthem despite it perhaps not being the kind of tune we usually like, for example. Happy Birthday is another example of a melody that we associate with good times!

And there's plenty I still haven't considered here - lyrics, the timbre of the sound that's playing the melody, how the melody relates to the harmony...

So, does any of this help us write good tunes?
Even if we consider melodies to be only sets of pitch and rhythmic relationships, they're complicated. 3 notes have 3 relationships between them. 4 notes have 6 relationships. 10 notes have 55 relationships (these are triangular numbers, by the way : http://en.wikipedia.org/wiki/Triangular_number). And this is before we start analysing the relationships between groups of notes.

Also - people get bored. The nursery rhyme-like tunes we loved when we were small may not seem so exciting now. As we grow up, we get bored of nice relationships between notes - Some of us want to hear something new - so we all go off and listen to jazz where they use all the notes at the same time, or classical music that changes key every half a second, or rock music with the sound so distorted that it generates a whole load of other frequencies not related to the original notes...

And of course, people like different melodies (and harmonies) for reasons that are as hard to explain as why they like different foods, sports, books.... There many, possibilities in a scale system like 12-tone equal temperament, and within those, you'll find, for every person, many things that annoy them, many things that bore them, and many things that excite them. Different styles of music tend to use melodies (and harmonies) with different characteristics, so that melodies in heavy metal songs are not the same as boy-band songs. That group of friends hanging around outside a concert may just be together because they happen to like the same melodies. And do they like them because they have learned to like them, or because of something to do with their basic brain chemistry? Who knows?

Answer 5

Music is not just a melody (unless we are talking chant, but even that works on a harmonic framework) but also consists of harmonies which structure the notes into subunits. Those harmonies are a bit layered: you can even have things like an organ point where one aspect (bass note) can stick around for a long time and form the root for a harmonic framework.

Now a nice melody has a progression of harmonic relations that tends to fall into place with the harmonic framework either at the main rhythmic places or in some meaningful relation to them.

But a melody is just a potential for harmonic frameworks: one of the famous themes is that given by Friedrich der Große to J. S. Bach as a challenge for improvisation, and Bach used that theme for creating the whole of the "Musical Offering".

As a melody, that thing leaves a bit to be desired... Bach's mastery of making music from it and dragging it back into harmonic contexts and classic compositional theory might, in a state of suitable inebriation, be counted responsible for triggering Schönberg, Zwölftonmusik and serial composition: if you could not leave the baroque time behind my messing up melody, it obviously became necessary to mess up everything else as well.

Now a "good" melody carries strong hints at the harmonization it is intended for: it basically carries half of its accompaniment with it. And it tends to have a solid relation to its tonic: you can often sing it splendidly while a vacuum cleaner is droning on (I tend to let it take over the dominant rather than the tonic). If you don't have a vacuum cleaner at hand, you may achieve the same effect with bagpipes.

Answer 6

It all depends on context.

In some cases, when there is no instrument playing chords to provide harmony, the first few notes of the melody will imply it, and this will dictate how a particular note would sound. If you drastically change these notes so they imply another harmony, the first seconds may sound funny, just before it all makes sense as a change. It's really your musical ear trying to fit what you hear into a familiar canvas.

Music is all about movement and creating/resolving tension.

Instead of pleasant or not, you can think of these notes as dissonant or consonant. Dissonant notes want to be replaced as soon as possible in a melody. Consonant notes want to stay sustained. It would also help if you think of dissonant and consonant as extremes in a scale, not as possible values for a binary variable related to the notes.

A very dissonant note within a melody can be very consonant in another context.

Western Music Theory has some ways to predict how dissonant a note will sound over a certain key and chord. You could study the modes of the major scale and their "avoid notes" and tensions. This is extremely useful if you do it paired with some ear training.

Also, what you really hear as a melody are the intervals between the notes, not the absolute notes (after all, you can transpose them), so it would be wise to study intervals.

The people that are able to create good melodies have a good understanding of how these intervals sound. A lot of them work outlining a harmony. They can create and dismiss tension at the right moments. I'm pretty sure they also experiment a lot with it.

Don't forget that rhythm plays an essential role in music. It can really transform what a melody has to say.

Answer 7

Intervals that can be expressed as small integer ratios, such as 2:1 (octave), 3:2 (perfect fifth), 4:3 (perfect fourth), 5:4 (major third), 6:5 (minor third), usually do not sound bad together or in a sequence.

But you cannot rely just on this when creating music.