postheadericon Drum programming

How to create a realistic-sounding drum track – drum programming crash course

First I’d like to point out that the tips here will focus placed on the principles of programming drums describing imitation of “live” drummer playing, however, these rules can/should be successfully applied even in genres with little in common with the so-called. playing live.

In almost every text on the subject of programming drums (and you will find a lot on the Internet) you can find similar advice: always remember that the drummer has only two arms and two legs, and therefore is able to play up to four sounds at once ! This is a fundamental principle to be followed if we are to realistic drum programming.

Basic examples

Let me start by describing the simplest example – the basic form of groove in the time signature 4 / 4 where the bass drum plays first and third quarter note of bar, snare drum plays on two and four and a hihat playing all eight notes of the bar 4 / 4:

This movie requires Flash Player 9

Nice “grooving” drummer will rather always play hihat from this example, accenting some of the notes. Basically, drummer will play accents on every other, odd eight notes: 1-st, 3-rd, 5-th and 7-th note:

This movie requires Flash Player 9

or vice versa – 2-nd, 4-th, 6-th and 8-th:

This movie requires Flash Player 9

..or example 1-st, 4-th i 7-th:

This movie requires Flash Player 9

Other sounds are quieter. In fact, it looks like the accented notes are assigned a “normal” level of velocity while those unaccented have lower velocity of say 30, so if “normal” level of velocity = 100, then a quieter note will have velocity, say 70. These are approximate values and are intended solely to illustrate the direction in determining the volume levels of individual sounds. Equally, it may happen that the hihat will sound good at velocity level 80 as an accent and say 25 as no accent – here much depends on the type of samples used and/or configuration settings for the analog sampler or drum machine if you work on such instrument.

Here is a midi file that contains the above examples of rhythms (General Midi):

  midi_1.rar (218 bytes, 2,769 hits)

Hereby, we have explained the first, absolutely essential question.

Sound of drums

Another very important issue is the sound of drums. Sounds of an acoustic instrument played with varying intensity sound different – this also applies to drums. Simply put, you can say that the quiet notes sound darker, or whatever other than a loud notes, and at this point it doesn’t matter why this happens. Also sounds played one after the other with an equal volume will never be the same on an acoustic instrument. For this reason, sometimes poorly programmed drums so easily distinguished from live-played drums – we all know very well the “gun-effect” that appears when you try to program a busy snare drum fill. However, you can fix this in several ways:
– for example, having only one hihat sample you can run low-pass filter in the sampler. The filter must be adjusted so as to affect more on the quiet sounds, leaving untouched the loudest – then the velocity parameter affects the filter cutoff frequency. Exactly the same applies to drum machines. Filter and tunning of the sample can also be slightly controlled by LFO – scope for action is the greater, the more of this type offers the possibility of our sampler or drum machine.
– you can also prepare a few versions of the same sample in an audio editor using a filter, tunning, timestretching etc. Depending on what software and ..imagination we have, results can be very interesting. Generally, the idea is to get several versions of “just like” sounding samples.
These methods are particularly effective when planning for the use of electronic sounds and are a little less useful to the acoustic sounds. For a reliable programming of acoustic drums is needed relatively large set of samples. There must not be a multi-channel “Drumkit From Hell” or “BFD”. However it is important that each element of the drum kit consisted of at least several different samples: loud, medium and silent strokes, preferably in several versions in each range of dynamics. Below are posted two examples of programmed hi hat groove – trained ear will notice the difference right away. The first example is played with single hi hat sample, the second example is played with a set of several hi hat samples recorded in all levels of dynamics.

  • single sample hi hat:
    This movie requires Flash Player 9
    – you can easily hear the “gun effect”.
  • multisample hi hat:
    This movie requires Flash Player 9
    - check it yourself. If you do not hear the difference, you probably need to change your audio monitors 😉

  hh_mid.rar (428 bytes, 2,574 hits)

Timing of programmed drums

Matter of timing – probably the most important thing in the whole process.
Generally I can say with full responsibility – the equal to, the worse …  I will try to help you understand what’s going on – you will need a midi-editor which allows you to move notes in the timeline for small values, eg something like Cubase.

For starters, let’s make a simple beat in 4 / 4, with the eighth notes on the hihat – may be that from the examples above – the simplest rock beat. Let all the notes will be quantized to the grid.

Now select all the hi hat notes and move them gently back in time – let them play with a slight delay. Now listen to what is happening – how it sounds? It should be so that the drums still sound equally, but a little more fat – delay value should be chosen with hearing – different audio editors offer a different native resolution, so I do not give the only correct prescription, but also it is not needed – I show where to look, you’ll do the rest.

Next exercise: we will only delay the odd (high) notes of hi hat – ie 1, 3, 5, etc. Even notes we leave untouched or lightly move forward to make them playing earlier on the timeline – Both ways will bring something new and can be mixed together, and go further listening carefully to what is then happening. I give you just a starting point for further experiments, being necessary to mention the principle – the denser or faster the phrase – the more you can move the individual sounds in the timeline. That’s why I proposed to start experiments on hi hat, but of course we can move the bass drum and snare drum notes, but I recommend here a bit more cautious starting with moving snare and kick notes, however, we must all – not just fear.

For further guidance, I would like to draw your attention to the natural rhythmic tendencies of all the real musicians – it is worth remembering that people:

  1. playing softly naturally seek to reduce the tempo
  2. playing louder naturally seeks to accelerate the tempo
  3. playing the more densely naturally seeks to accelerate the tempo
  4. playing less densely naturally seek to reduce the tempo

Of course, most musicians trying to eliminate these tendencies in the process of learning to play an instrument (working with metronome), but this natural feature is always present and it is nothing wrong, just our nature. Good musicians can simply control, but they also know about the fact that the complete elimination of this natural tendency is by no means necessary.

Programming drums is also worth to remember – if the song contains a quiet part with no drums groove, sometimes sounds unnatural to keep a constant tempo in such part of the song. If your audio sequencer allows tempo-mapping (tempo change) within the song, I encourage you to experiments – in such a quiet part of the song you can safely slow down the sequencer tempo by 1-2 bpm (in the ballad) to up to a dozen of bpm in a fast song,by then, with the entry of rhythm section groove, go back to the original tempo and everything should sound more natural than if we keep constant tempo within whole song.


Leave a Reply