I've researched my question on Google before and came across a few threads which I've read a couple of times. So far I've got into music theory and understand scales, how chords are formed, I know what intervals are in terms of half steps apart (e.g. maj2nd - whole step) but I can't seem to tie the information to create my own melodies.

I've seen the same terminology so am learning the relevant material but just can't seem to really know how to create a melody.

To me, a melody is a series of notes put together which sounds pleasing to the ear. The fact there are 12 notes in an octave and people make endless amazing melodies really inspires me, I've been reading a lot into music theory and would love to start creating music now. Any advice is always appreciated.

The one rule I'm going to stick to in terms of simplicity sake is to remain within a chosen scale so I know what notes to press but how do I learn what notes sound good together - is there any theory there which I can apply?

I've also read that making melodies is something that cannot be learnt - so how does music theory apply to it? I want to focus my efforts on creating melodies and building instrumentals from my knowledge so far and to build skills.

  • 1
    Creating melodies can be learned, so that bit of information is wrong. You do however, need to provide more information. Do you play any instruments or are you trying to create music using some sort of piano roll program?
    – MrTheBard
    Jun 20, 2014 at 14:18
  • Ok thanks, I've read on forums such as gearslutz attacking people asking this very question. Sorry for my lack of background -- I've got a midi keyboard and want to create instrumentals and use Cubase as my DAW
    – MJohnson52
    Jun 20, 2014 at 14:19
  • Well we're not here to attack we are here to help, but that bit of information is wrong. I ask again though, do you play any instruments? I do need this bit of information before I can answer.
    – MrTheBard
    Jun 20, 2014 at 14:20
  • Keyboard :) Thanks, I've browsed through some relevant threads and there's good information on this forum.
    – MJohnson52
    Jun 20, 2014 at 14:26
  • 6
    The creation of melodies is learned in the same way that any skill is learned, by practicing and experimenting. One does not learn an instrument simply by reading theories on how to play the instrument, you need to get the instrument under your hands and practice. Same thing applies for melodies, you can read about theory which I recommend and will greatly assist you, but you have to practice. Creating melodies can certainly be learned.
    – MrTheBard
    Jun 20, 2014 at 18:53

12 Answers 12


As you listen to music analytically (especially music you like), there are various patterns and things you'll discover, which will help you learn how to write melodies.

For example, it's obvious that in many pieces, a melodic phrase spans a number of bars that's divisible by the number 4. If you can pin down the 4 parts of some such melodies (that you like, of course), you can figure their role in the melody. I noticed rather early, for example, that I really favored melodies where the first and the second part are the same, except for the last note. You may find some such patterns that work for you, as well.

Also, as you might have noticed already, movements in semitones are great for conveying sadness. Use that!

Or, if that's not your cup of tea, throw away those and stick with the pentatonic scale. It has a very unique and recognizable style and songs that use it are quite often described as "awesome".

Work out how the chord progression (and the expectations it creates to the listener) works together with your melody. For example, the same melody has an entirely different effect if it ends in the I chord (expected) than the same melody would if it ended in the VI chord ("deceptive"). And a melody can raise expectations if it rests in a V chord for a while - it often implicitly promises an equal amount of melody coming soon.

Perform mutations on shorter melodies. For instance, try repeating the chords of the melody, but in strategic places pick a different note (from the same chord) than you did last time. For example, in a V chord, where you played the II note in the first iteration, now play the V note instead. This trick is common to let the listener know that this particular section of the song is reaching or has reached its climax.

Alternate the mood as you go. If your melody has a minor feel, try following it up with something that has a major feel. Pieces that don't do that run the risk of sounding monotonous. In my experience, I've found that the human mind notices and appreciates transitions more than stationary situations. The happy part will sound happier if it follows a sad part and the sad part that comes after a happy one will sound even sadder and more dramatic.

Always keep in mind that, as it happens with all works of art, the point is not only the destination, but also the journey. A melody can be a journey to the other side of the scale and back. That doesn't mean that there aren't pieces where the melody stays around just a couple of notes. But usually, these kinds of pieces invest in something other than their property of being melodic - some kind of harmonic or even purely rhythmic effect, perhaps.

Of course, you can also just choose some chords (even at random, if you're willing to experiment), and write voices using the rules you know from studying music theory (if you have any background in that). There are some great melodies to be discovered there, while if you stick with more predictable chord progressions, your songs will tend to sound the same as other songs.

Try sticking to an unsual time signature (it may take a few minutes of practice beforehand) and explore what you can produce off that.

You can even just read a poem or some lyrics (which you've never heard music to) and try to notice how you instinctively read it in your head (or out loud), then write down whatever melody has popped up in your head. It's fun and some recognizable melodies often come out of previously written verses.

Or you can literally watch birds sitting on power wires.

  • 1
    Thanks so much for this!! I'll re-read all the answers again, but this is the melodic style I'm after, can anyone bring up these notes so I can analyse the intervals etc. afterwards please: This sounds like it is single notes being played out one after the other soundcloud.com/mjohnson52/piano-1 This is a chord progression with a bit of passing tones I think in between - this is especially the one I'm after so I can learn to play that sort of style, sounds really good to me.. soundcloud.com/mjohnson52/chord-progression-piano
    – MJohnson52
    Jun 22, 2014 at 11:09

Creating melodies or a melodic line is not much different than creating a figure in drawing.

I see two ends of a spectrum of choices with everything in between as possible.

1) derivative

2) original

I. Derivative melodies are a subjugation of environmental influences literally and figuratively. Examples include the following:

A melody based on songbirds.

A melody based on car horns in rush hour traffic.

A melody or texture based on the ocean surf.

A melody based on a woman talking (often used by Blues guitar leads).

A melody based on machinery, "Ballet Mécanique" by George Antheil.

A melody based on fire, wind, or rain, "The Four Seasons" (Italian: Le quattro stagioni)...a set of four violin concertos by Antonio Vivaldi, or "Sonic Seasonings" by Wendy Carlos formerly Walter Carlos.

II. Original melodic content is based purely on imagined sounds. Examples include the following.

Improvised on the spot and in the moment such as melodies by Classic North Indian Raga Singers, and many Jazz and New Music composers/performers.

A melody where sound is totally aleatoric. (John Cage 4' 33")

A melody that accomplishes a specific compositional goal, e.g. 12 tone atonal music.

A melody where technical skills permit pitch organization in a super human way requiring mechanical and/or electronic and/or computer engineering such as the electronic music of Edgard Varèse, Karlheinz Stockhausen, Maggie Payne, The HUB, David Behrman and many others.

III. Somewhere between these two extremes:

A melody that furnishes a good thread to connect harmonic content or based on a chord progression, most of Jazz.

A melody that defines a musical form such as counterpoint.

While the above is brutally simplified, it allows a wide set of guidelines to permit almost anything to be constructed into melodic content.

Music theory from any culture, study of scales, study of musical literature all have a place in assisting the composer in sculpting a melody, but are as confining as using a ruler or French Curve for the artist. These things are best used as guidelines and measurements for proportion and compositional form but can not be greater than the emotional content endowed by the composer and brought to life by the performer.

Regarding "melodies is something that cannot be learnt".

If one can learn to walk then one can choose to journey.

If one can learn to talk then one can be heard.

If one can listen then one can learn from others.

If one can read and write then one can compose letters and understand those sent.

If one can make art (music, dance, figures) then one can connect from one heart to all.

Composing is simply making choices whether you are looking at melodic content or a symphony. The more you learn what options you have to choose from the better you will become with making your idea real.

  • In what way is 4'33" aleatoric? Jun 21, 2014 at 1:37
  • 3
    @ filzilla - aleatoric is a dicey way of describing it. His follow-up, 9:06 , was twice as good...
    – Tim
    Jun 21, 2014 at 7:14
  • @ninemileskid 4'33" as you know is all 'rests', and there-fore the only sounds that occur are ambient, which is completely random.
    – filzilla
    Jun 24, 2014 at 17:45
  • 1
    @filzilla youtube.com/watch?v=HilGthRhwP8 Jun 27, 2014 at 2:32
  • 1
    @filzilla Why don't you consider it aleatoric? In a certain light, all performed music is aleatoric in that the performer is endowed with at least a modicum of choice in how to interpret the work. In Mozart this is manifest in the choice of tempo, vibrato, etc., and at the other extreme is 4'33". In my opinion there is a category of music nearer 4'33" than Mozart which is conventionally called "aleatoric" to which this piece by Penderecki belongs, especially because its "original melodic content" is a superposition of the semi-random choices of many performers arranged in a totally random way. Jun 27, 2014 at 21:34

You say you know scales. Most melodies, especially the more simple kind, are made up from the notes of a particular scale, be it major or minor. To start with, try using the pentatonics. They are the major/minor scales with the awkward notes missing. As in C major, C,D,E,G,A. Not F or B. Using just those notes, many good melodies can be made, as leaving out the 4 and 7 means nothing will clash - ever !! And don't think the tunes will turn out 'Micky Mouse'. Listen to the solo in Sir Duke (Stevie Wonder) - apart from two little passing notes, it's all pentatonic.

Once you are happy with this idea, start putting the other two in. I think you're bogged down with theory. Theory for me comes way after where you are. If a set of notes sounds good in sequence, then it'll be justifiable in the theory. Using theory to create a tune won't do anything.

Above all, stay in one key throughout a song, and use only notes that come from that key. They will sound good together, and at least give you a basis on which to start.


First things first: I believe you are confusing a melody with a full blown music piece. In simple terms a melody, or melody line, is a sequence of monophonic notes (one note played at a time) and a music piece is a melody line played with an harmonization to support it.

For example, imagine you write the following melody line:

enter image description here

The first thing to do is to determine which tonality it is in and properly change the key signature. In this case we have a D minor denoted by the last long note being a D and after looking at all other notes used they matched to the D minor scale. Especially the natural F (which gives a minor third) and the Bb gives it away:

enter image description here

Here is how it sounds.

Then comes the part which for me is the most difficult part of building a music piece. I am still trying to learn how to do it properly: you add chords to harmonize the melody. This was the first piece I ever composed and harmonised on my own. I tried to use a II-V-I cadence very common in jazz. If there is one thing I have learned about harmony was that it is far from being an exact science (just like music or art in general) and it takes lots of time to learn:

enter image description here

Here is the sound of if.

Finally, by adding rhythm to the harmony and by adding a base line and drums I got this. (It is the full piece and not the first 16 measures as before) This last step can be easily done by using software like Band in a box or iReal Book on the iPhone.

I hope it helps.

  • 1
    In jazz, V-II-I is common? Maybe you meant II-V-I (or ii-V-I)?
    – Tim
    Nov 8, 2016 at 8:49

Other posters have mentioned the key resources for you to check out. I won't rehash them. I will, however, give a more unconventional answer to balance them. By all means, soak up as much as you can re: theory, composition, etc. Let your curiosity really take you into the weeds. Examine the melodies of your favorite tunes with peculiar scrutiny, as if you were decoding them in search of some grand theory of melody.

Then sit down at your DAW and let all of that go. If you have zero inspiration in that moment, take the first hard step, and put down a chord progression- even one that doesn't particularly move you. Your mind will rebel because it sounds so plain, but as you said, everything comes from those 12 simple tones.

Then, let the magic of improvisation take over. How do you do this? Your mind will rebel against the very notion that you're capable of doing anything. "I'm not a musician. I don't know what I'm doing." Nonsense. Anyone and everyone can be musical. A baby can hum along to a tune, playing its first "instrument," if you will, before it knows anything of theory. Just listen to the progression playing in your DAW. The compulsion to immediately lay down an entire melody will make that difficult at first. Ignore that compulsion. Just listen to the progression. Then listen to yourself. You likely won't hear an entire melody. You will probably hear a one or two note embellishment to the chord progression. Fine. It's an idea. Loop it in your DAW. You now have the chord progression with a couple notes embellishing it. Listen to it. Listen to where your ear wants to take this little loop. Does it want you to extend this rudimentary idea you already have down? Does it want you to respond to it with another idea, another embellishment? You'd be surprised what you can build that way. Before you know it, you might have a more complex musical idea that calls you to forget that initial chord progression and give it another one altogether.

Think of yourself as a jazz musician soloing and let yourself go--whatever helps. The notion of jazz improvisation DEFINITELY helps. And if that means you can compose a great jazz solo a month from now, but want to create more focused, conventional melodies, then begin to limit yourself to the pentatonic scale and see how it goes.

If you have some time, check out Victor Wooten's book THE MUSIC LESSON. It's a little new-agey, sure, but it hammers home the notion that there is more to melody than theory, and more to music than melody (or simply "notes" as he might call them).

Don't get me wrong. I work in a quantitative, analytical field, and love getting really into the science of melody in its most nuanced form. But for every such foray into theory, I think you need another foray into dropping it all, letting it go, and letting yourself create.


If you've already worked out part of a piece of music- say the chord sequence, then a melody might "suggest itself" as you listen to the chords being played.

Also the set of notes in the scales used by the chords will of course narrow down the notes that are sensible. Sometimes this works in your favour in that making a melody is hindered by there being too many options. Using the chords as a limiting factor might help a bit.

Another way of going about this is to write a melody that you like then fill it out with accompanying chords.

Regarding "what notes sound good together" - well there's the alchemy.

It's subjective in that some tunes might sound great to some people, not to others. That doens't help you make melodies but it does hopefully help you understand that the "Universally lovely tune" is something that's very hard / impossible to make.

When it comes to writing music and melodies, you might do well to consider what mood you want to get across, and follow on from that.

For example major keys are generally considered a "happy" sound, minor keys "sad" but that that's just a guideline notion, there are plenty of examples where the rule is broken.

I'm not sure there are really any hard and fast rules to follow. In fact I'd argue the most interesting tunes (to my ear) are ones which seem to break percieved rules.

However I do think the skill of making a pleasant melody can be learnt, perhaps by learning how to build tension and release (cadence) and different moods (major/minor).

My experience is that experimentation while playing is probably the best way forward.


Almost any book on counterpoint offers advice on creating melodies. I would recommend taking that advice even if you're not interested in writing contrapuntal music (fugues, canons, etc.) If it's a good melody for singing it's probably also a good melody for most musical instruments.


There are a couple of outstanding answers here already, I'm just going to address the "cannot be learnt" fallacy, because it just might be well-intentioned.

I understand what is meant by "cannot be learnt", and disagree, I think it would be better to present the argument as "cannot be taught". Like, nobody taught you to walk, or ride your bike, they just gave you endless encouragement until you got there yourself. So here I go...

You wish to create something completely original, easy, anybody could do it, nobody needs to teach you to do that, so... there are probably a few additional requirements:

  • You probably would like to create something that you would like to hear again.
  • You need to be able to remember how to play it.

I would like to hear that again!

Why? So, I assume it wasn't a collection of random notes. Why wasn't it random to you? Even if all the notes belong to a key, scale, mode, even if they harmonize with a chord progression, they can still be random, just random within those constraints.

When you heard those notes for the first time, something unusual happened, they made sense, even at the abstract level, you somehow understood that thing you just did, as if they told a story, and you liked it. Which is fantastic! That's the desired result. I wish I could tell you how to do that, at will, and prolifically pen meaningful melody.

You can part with good money on books and courses that claim to teach just that, be my guest, but I think such material would be best-sellers if they really worked.

Now, how did I play it?

Enter music theory. Especially scales, keys, modes, chords, harmony, rhythm... based on centuries of observation of musical phenomenon.

This will help you make sense of the mechanics of your melody more easily, it's not impossible to do this without any explicit musical knowledge, many people do, but it can reduce the time to find things you would like to hear again, and help you to mentally structure the information about your music so that you can remember how to play it. Excellent.

However, did music theory teach you how to create a melody? It certainly didn't hurt, but in itself, I'm not sure it provides the 'creative spark', the imagination, the ideas. Of course, it could have, since anything can, but in my experience, creativity 'just happens', there is no formula, no secret to success, no hard and fast rules, you just do it.

If you cannot naturally mentally conceptualize an original melody (for example, if you are reading candidate lyrics and cannot 'hear' the melody you think would fit it well), it's nothing to worry about, that's only one technique, there are countless others (e.g. free play).

Music theory provides a framework, but don't expect it to be a portal into creating original melody. Music theory can give you constraints for your music, and in certain situations, being constrained can be a creative boost (e.g. try making a melody from 4 random pitches - you'd be surprised how many people can pull that off).

Conversely, removing constraints is another method to boost creativity, studying a musical concept that is unfamiliar to you can broaden your horizon and draw in new inspiration. Any shift to a new perspective can contribute to creativity, even non-musical changes, an emotional event in your life, for example.

Many artists have confessed at some time that "the melody just came", as if it was a divine experience. There is also the aspect of when trying to force creativity, it becomes blocked. Apparently, even boredom can help creativity!


It seems that you have at least a basic understanding of music theory. Understanding the concept of chords and scales is fundamental for music composition. I too find it so interesting and beautiful that with the 12 notes that are found in western music, there are so many possibilities.

There are a lot of complexities in music composition and the deeper that you get into music, the more that it will continue to surprise and delight you, as well as occasionally frustrate you.

What I would suggest to you is look up some common chord progressions, namely a I, IV, V, a ii, V, I, and a I, IV, vi, V. Stick to those chord progressions for your first few compositions and come up with melodies to them. A lot of melodies tend to follow a sort of question, and answer format. I think it's not only great practice, but also perfectly acceptable to keep your melodies simple and utilize this question and answer format.

Those chord progressions, are used frequently in all genres. The I, IV, V works so well because it contains all of the notes of a scale! Therefore if you stick to only one scale as you proposed above, it won't sound wrong. Now whether it sounds good, is up to how clever you can be with the notes of the scale. Ultimately you will have to be the judge of whether or not what you're writing sounds good.

Although theory exists, it is not a formula for successful music. Art is extremely subjective, if theory provided a formula for how to write good music, then technically every single song in the world that was written by someone knowledgeable in music would sound good, and as I'm sure you realize, this is simply not the case.


You might consider looking at Jack Perricone's Melody in Songwriting. It's designed for, obviously, songwriters, but the basic principles of melody it sets forth are applicable to any kind of music.

  • Thanks I'll check it out, beside from that I've decided to really learn my scales, chords and get the most from music theory and apply it with practising other songs then making my own compositions
    – MJohnson52
    Jul 16, 2014 at 12:25

An old-fashioned but still relevant book on melodic creation is Percy Goetschius's "Exercises in Melody Writing" which is available as a free download here: https://ia802605.us.archive.org/16/items/exercisesinmelo00goetgoog/exercisesinmelo00goetgoog.pdf

It's a basic but very detailed description of parts of melodic structure. There are also lots of exercises.



If you really want to create melodies, sing. And listen. All the time. Sing with bits of music you over hear or only half hear. Sing with sound tracks. Sing with the radio. Sing your favorite tunes. When you talk to yourself, sing the words. Make up songs like you did when you were a kid. When you hear yourself sing something cool, record it.

Sing riffs, short repeating melodic fragments 2-4 bars long. Sing variations of each riff.

When you get back to your instrument of choice, play against the songs and riffs you've recorded. Learn to accompany them. Sing with them. Sing contrapuntally.

Finally, start mashups of songs. If you're unsure how to do this, listen to Umphrey McGee's "Zonkey" album. Or if you're of more a classical bent, listen to Charles Ives.

Compelling melodies don't come from theory, they come from another place somewhere in our minds. Theory helps support the tunes and sand off the rough bits

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.