If you aren’t confident enough to record melodies from a MIDI keyboard or even your computer’s keyboard (a nice feature in Live for when you’re on the road with no MIDI controller), I find that the easiest way to write melodies with the pencil tool (Command + B for mac users, Ctrl + B for windows users) is to write in your melodies step by step. If you recall the formulas for major in tones (W = whole step, H = half step) (W W H W W W H) and minor (W H W W H W W), then you can use the Fold feature of Live’s clip view to hide the notes that are not included by one of these formulas. Notice in the first image, we have one octave of notes stacked up on top of each other in two different scales, F major and F minor.

All we need to do is create one of these stacks in a MIDI clip, and then duplicate it once or twice. Just select all the notes, then hold down option (mac) while dragging the notes up one octave. This should create a duplicate of your notes, but transposed up one octave. Do this again for the octave below. Now when you click the “Fold” button located at the top left of the clip view, all notes that are not in the clip are hidden. Note in this second image that at the very left, there is a stack of notes that form the scale of F minor. After that, I randomly double clicked to create new notes all over the grid. I set my synthesizer’s polyphony to 1 so that it can only play one note at a time. So no matter what notes I drew, they were all in key. As long as you have the fold view enabled, you can now draw notes anywhere and it will still sound pretty decent.

In most sequencers, there is a quantize function. This will take all of the note data on a track and aligns all of the notes to the nearest division. So if you play sloppily or you just have bad latency on your midi keyboard, quantize is your friend. Select the division most appropriate to what you’ve just played. 1/8th and 1/16th divisions are most common. Another cool thing about quantization is that it will sometimes snap a note to a division that you didn’t intend. But it sounds good anyway, minus a few odd parts. Then this quantized melody may inspire you to come up with a new melody altogether. Of course, not all music sounds good with quantization. If you’re going for an organic, human sound, then you probably won’t want to quantize 100%. There are options in most sequencers to quantize to a certain percentage. This is useful if you want to correct your timing, but not to the point of computerized perfection.

So, you don’t want your melodies to be boring? My first suggestion would be to come up with different variations on a melody that you can use as a loop. Then vary which loop you use in the sequencer. For example, if you wrote two melodies (we’ll call them A and B) and then made a variation on B (call that one C), you could sequence A B A C. This is a very common pattern. Or perhaps try A A B B, or A B A B, etc. My next suggestion is automation of several different knobs on your synth. For example, you can automate the cutoff frequency, vibrato, volume, modulation depth, etc. One trick I commonly use on my synths is to route an LFO to the filter cutoff frequency. Then I route the modulation wheel on my keyboard to the LFO’s depth (the LFO speed should be synced with the tempo at 1/8th or 1/16th). You could also route the modulation wheel to the LFO’s speed for an interesting combo-modulation. The modulation wheel on any midi keyboard is handy for live performances. But when it comes to sequencing, there are no limits to how many automation tracks you can use. I suggest automating many parameters at once to keep things interesting. For an example, listen to the solo synth on my track, “your best shot”. In this track, I am automating the cutoff frequency for those intense moments (on a low-pass filter), amount of vibrato (depth of LFO routed to cutoff frequency), and amount of delay (via the aux send knob). All of these combine to make for an expressive, almost human-like quality to the synth.

In this screenshot of my song, (made with Ableton Live), you can see the automation of the vibrato amount overlayed on top of the notes.
Listen to the melody shown in this image.

The vertical lines in a sequencer are what slice up musical notes in time. Depending on what resolution you are viewing your sequence at, there can be many variations between how closely the vertical lines are spaced. But before we can look at the divisions of time, we must talk about time signature.

The time signature tells you how many beats are in a bar and what note or rest is equivalent to one beat. Most time signatures are 4/4, but things can get interesting when you change the time signature. Whenever the second (or lower) number of a time signature is 4, this means that one beat is equal to a quarter note. When the second number is 8, one beat is equal to an eighth note.
Get a more detailed explanation.

For a 4/4 time signature:
1 bar = 4 beats (4 quarter notes) = 16 sixteenth notes = 32 thirty-second notes
As you can see, there are many different resolutions that you can split 1 bar into. Typically, resolutions of 1/32 and 1/64 are in the realm of microediting because any change of notes at these divisions will sound very quick.

Say this out loud to get an idea of what 16th notes are like:
“1 e and a 2 e and a 3 e and a 4 e and a”
This whole phrase is equal to one bar. Each utterance equals a 16th note. Count ‘em up, you’ll see that there are 16 separate divisions.

Vary your note lengths: If you want your melodies to be expressive and interesting to listen to, you must vary your note lengths. After drawing a sequence of notes, try altering the length of the notes and shifting them from the left or to the right in the sequencer.

Vary your note velocity: Same as if you’re programming a drum track. Unless you specifically want a track to sound mechanical and computerized, you should randomize your note velocities a bit. All the while, loop your sequence and listen to it as you go. You will get sick of it eventually, and that’s when you know that it’s time for a break.

Sometimes it’s just easier to draw melodies in a sequencer than it is to actually play the melody on a keyboard. Especially if your music is fast. Essentially what you will be doing is drawing a rhythmic pattern composed of note segments that can be organized within a scale. Most software sequencers have 2 dimensions: The horizontal, or time dimension, and the vertical, or pitch dimension. As the playhead progresses along from left to right, the notes that are drawn on the horizontal lines are played. The rest of this section will be divided into two parts, one for each dimension.