The Feel Factor
A Guide to Programming Music with Soul
How often have you heard synthesized, sequenced music described as 'cold' and 'mechanical? All of us want to compose a tune with good 'feel', but few people know exactly what makes for a good 'feel'. Yet it's not a mystery - as this timely article by Michael Stewart, designer of the Kahler Human Clock, reveals.
"Cold." "Mechanical." "Sounds like a computer." How often have you heard synthesized, sequenced music bad-mouthed like this - and, sadly, how often has the criticism been true? All of us want to compose a tune with good 'feel', one that bops out of the grooves instead of just lying there flat; but few people know exactly what makes for a good 'feel'. Yet it's not a mystery - as this timely article reveals.
There is a biological foundation for all of this that comes under the rubric of 'psychoacoustics'. The brain releases specific chemicals when it is surprised, and other chemicals when events come as expected. Furthermore, if the events keep coming when they are expected, the brain will release a 'boredom enzyme' to block out the sensation altogether. This last effect, the blocking of sensation, is what I call the 'jackhammer effect': that is, if someone started up a jackhammer next to you, the noise would initially make you jump out of your skin, but after a while you would tune it out - just as you don't notice the pressure of the seat you're now sitting on until someone mentions it. I have read that nature does this so that you're not constantly bombarded by information and miss some important new stimulus, like a truck coming at you. This interplay of the expected and the surprising is at the very foundation of music.
Johann Sebastian Bach was a master at this, repeating a phrase or sustaining a pedal tone to create tension and then introducing a surprising new element, such as a sharp accent, to create a release. The subject of music and psychoacoustics is at once fascinating and overwhelming, but often elicits a "so what" response in people who don't see how to use this information in a practical way.
When I was producing, I would endeavour to pick the take with the best feel. Often an arranger had written out all the parts, so on each take the notes played were essentially the same; yet there were differences between takes - differences in the feel the players gave the music.
In the studio control room, trying to get the feel I wanted, I used to tell my artists that there was no 'feel' button on the mixing console. But, in a way, now there is. Thanks to computer-aided music, we have been able to explore and quantify these effects a bit more precisely than in the past. I have experimented and catalogued a good deal about feel and related phenomena, and would like to share some of this research with you, the readers.
Like any gift from the gods, machine music has both given and taken away. Certainly the number of sounds, and the predictability of performance now available to us, are an advantage. The luxury of putting a record together at home and 'dumping it' at a studio can make the process much cheaper and easier.
On the debit side of life, machines are frequently accused of taking away spontaneity, dynamics, and feel. (This is not strictly true, of course. They make things feel like machines.) But although machines do sound this way, it's because they're machines - computers - and can be programmed, that they offer us the potential of putting the feel back into the music in a predictable way. Specifically, any sequencer software that has even remotely sophisticated editing functions will let you move events around in time. Sometimes this can be a somewhat time-consuming and primitive procedure, but it's powerful enough for you to put back into the music what I believe is a very important missing element.
Keep in mind that we're not talking about a giant difference in the music here; you'll hear a difference comparable to the one reverb makes. But if we removed all the echo and delays, robbing us of our front-to-back spatial dimension, that would be a definite drag. Putting feeling in a track is like adding reverb; it adds a dimension that can make your demo sound more like a master. By shifting our notes in time, we can even make all the non-drum instruments cook and groove all by themselves. This is a vital advantage that lets you put less on your records to make them say more. Then, when you do add the drums, your stuff will be romping - just like the pros.
Let's look a little deeper into the Guerin/Tutt effect I mentioned earlier. I'll need to use some subjective terms that I'm used to here, like 'lighter' and 'heavier', but since we are talking, after all, about using specific, objective, techniques to create subjective, emotional effects, then subjective terms are necessary for descriptive purposes. (You can always replace them with terms that are more familiar to you.) Taking the bass drum as the constant then, and the snare as the variable, it transpires that the more the snare is behind (following) the beat, the 'heavier' it feels. The more 'on top' of the beat (preceding the beat), the 'lighter' it feels.
There are also identifiable gradations of these general feels. If a snare is placed just slightly behind the beat, it is 'grooving'; a little more, and it is 'in the pocket' (R&B style); still more delay and it is 'pulling back'; and when taken to the extreme, it is 'dragging'. Remember, too, that there is also an overlaid perception of heaviness the further behind the beat you go. Is it possible that you can really give your music the feeling of a particular style just by where the snare is placed in time? You bet it is.
Going the other way, if the snare is put more on top of the beat (earlier) relative to the bass drum, it is first experienced as having a 'snap' on it; placed further forward, it is perceived as 'driving' or putting an 'edge' on the sound; even further ahead and the feel gets 'nervous'. Further than that: go look for a day job.
"What translates into 'feel' is the difference between when an event comes and when we expect it to come... this interplay of the expected and the surprising is at the very foundation of music."
Steve Perry, of Journey, likes to have his drummers 'lean' the snare ahead of the bass drum to drive the beat. He says it feels "more rock and roll" to him. Quincy Jones' snares are always a little on top (perhaps an influence of his jazz roots). Notice also that with older Motown recordings the snare goes in the 'pocket' position. Does this mean that you could take a Michael Jackson recording and, by moving the snare back in time, make it feel a little more like a Motown record? Try it, and see for yourself.
Don't restrict your use of this technique just to the drums, either. The effect is most obvious there, but it applies to all other instruments as well, and the effect can be just as worthwhile.
When I was developing The Human Clock, a device that makes drum machines and sequencers follow the tempo of a live drummer [distributed by John Hornby Skewes & Co, (Contact Details) - Ed], out of necessity I had to look at a lot of bass drum patterns. What I found was that even with the best players playing the simplest patterns, the bass drum would almost never arrive where it was predicted to be - there seemed to be a variance of at least one millisecond (1/1000th of a second) earlier or later. After beating my algorithm to death, I was convinced that this was not a software bug. This is the way things are.
Let's relate this to a studio experience. Tell your local 'drummer of doom', who is great at playing in time with a drum machine, to play the same part as the drum machine. Compare the drummer and the drum machine, and you will find that the part played by the human will not have that 'android vibe', but the drum machine part will.
Practically, this means that we can prevent the mind from releasing that 'boredom enzyme' (caused by constant predictability) by varying, in small amounts, the points where the expected events come. I've been able to randomise between 50% and 75% of the notes in a bass drum part to good effect. We don't have to move them around a lot to lose the robot vibe; differences of one to two milliseconds are noticeably effective.
This fine a resolution may not be currently available on all sequencers, but do what you can; even small degrees of randomising can make a difference. Note that digital delays can also be used to tune a part in time, although generally this limits you to a constant amount of delay.
Using a device such as The Human Clock lets a 'biological' drummer (whether you record the drummer or not) drive your drum machines and sequencer, and has an effect that has been described to me as making the music "come alive". This can be analysed on two levels: first, the music is being randomised to prevent the boredom syndrome and, second, the human is pushing and relaxing entire sections of the music-verses, choruses, bridges - thus causing even larger overlays of emotion (more on this later).
Silence, of course, is not an event but the absence of an event. Yet its effect on the notes it surrounds is so powerful that we need to discuss it as well.
"Is it possible that you can really give your music the feeling of a particular style just by where the snare is placed in time? You bet it is!"
Large feel = Amount behind the beat + Amount of space
These are predictable effects that you can use with predictable results.
If two sounds occur closely together, your mind will tend to pay attention to the sound that comes first. As an experiment, try putting a digital delay on a guitar with the dry signal panned left and the delayed signal panned right. Your mind will tend to gravitate to the original (first reflection) rather than the later echo. This, as you'll see, is a permutation of the 'more on-top/drive the beat' effect discussed under the 'Placement' heading above.
This effect is useful in a couple of ways. Let's imagine that you have a low string part pedalling on a single note (which soon becomes unnoticed and creates tension) and you do an accent-stab to another note. The effect of this will obviously be to draw attention to the accent note. This effect is similar to the piano note in the acoustic chamber: because the mind tunes out the pedal, the pedal acts a little like silence. You can heighten this effect even further by using the 'First Come, First Served' effect and placing the stab a little ahead (on top) of the beat.
Tricks like this let you get the most bang for your buck out of your notes, and can sometimes have an overall snowballing impact on your music: you'll use fewer notes to achieve your desired emotion, which means you can use more silence, which in turn makes all the notes more important, and so on.
In the same way, if you have a low rhythmic string part, you can place it on the beat or slightly behind and it will groove there and act as a complement. However, if you advance it and put it slightly on top of the beat, the 'First Come, First Served' kicks in and the line will become 'insistent' and noticed. Also, it will begin to 'drive' your track as if it were another rhythmic element.
Another use of the 'First Come, First Served' effect is in the opposite direction. Suppose you have a part that is musically busy, eg. a hi-hat part. Slide this part back in time slightly and it will tend not to be annoying. We are sliding into what our mind perceives to be the 'groove area'. The hi-hat will still be there, but the track will feel a little more 'down and dirty' as opposed to 'nervous'.
In a study by speech synthesis researchers, certain musical phrases were played to a wide range of peoples, from white Anglo-Saxon college students in New York to aborigines in the outback of Australia. It was found that certain of these phrases evoked the same emotion in all of the listeners.
Consider a musical phrase such as an arpeggiated eighth-note synth bell that goes from low to high to low to high, etc. If you slide the notes back in time as they descend, the phrase will feel sadder. I don't know why, but I can guess. There is a common musical thread in all peoples and perhaps this retarding of the downward notes in the arpeggio emulates the sigh-like sound (shape) that all of the listeners in Rubenstein's study perceived as sadness.
Then there's the Electrifying Drum Fill. Suppose you're cruising along in your tune and there is a section coming up where you want the track to really 'take off'. Make the drum fill start slightly on top (ie. ahead of the beat); it will command the attention of the listener, and deliver you into the coming section (assuming, of course, that the section is worth listening to; if not, this will be like pole-vaulting into a ditch).
If you want the next section of the tune to groove, set it up with an eighth-note style drum fill, but gradually lay (say) the last four drum hits back in time. If the following section has an effective groove element (bass pumping, but laying back), you will put your listener in deep Detroit.
Each one of these is an example of what I call the 'elastic phrase'. Slow down or speed up a portion of a phrase, and perhaps bring it back to its original tempo; a very human, emotional effect results.
"...if you want to render a group of notes meaningless, surround them with a lot of other notes. This is why 'cluttered' is one of the worst terms that can be applied to a piece of music."
These effects are just the beginnings of how to add the feel dimension to your music. Remember, though, that they're just tools and must be used in context. If you make the whole song feel a certain way, feel might begin to work against you. So, if one section feels a little nervous, for example, it might be a good contrast to what is happening in the rest of the tune, or might be just the thing to go under (say) a guitar solo. This is the Human Clock effect again: you can make whole sections of the tune breathe and feel different from others. Use these changes in feel instead of adding another instrument and avoid the trap of adding more and more notes that do less and less.
Combine effects by moving different instruments around in different ways in a section, and these processes will begin to interact. Add the concept of space as an amplifier and treat whole sections as entities (drive section, groove section, nervous section, etc).
It's possible to go quite deep into this feel phenomenon. In the future, I hope to design a device that will make playing with this dimension easy, just as digital delays have made it possible to play with acoustic time effects.
I was talking with one musician about this feel business and he showed me his Macintosh sequencer. When I saw the visual patterns with which the sequencer program represents notes (bars whose length represents duration and whose height represents pitch), I was struck by how familiar they looked. I realised that we see the same patterns over and over in nature and in our own creations. What you see in the swirling cream in your coffee cup comes from the same forces that shape the galaxies. I remembered too that the distance from the nucleus of an atom to the orbits of the electrons are the same intervals as the musical intervals of third, fifth, octave, and so on.
Perhaps these and other patterns in nature - magnetic fields, Jupiter noise, crowds leaving a concert - could be quantified and then used as 'driving patterns' for parts in something like a sequencer program. Or maybe it's possible to use nature's own designs to modify the parameters of feel. It should also be possible to analyse our favourite feels and store them in some yet-to-come device - "Add a little more Jeff Porcaro please." It's a great age. I love it.
Using feel will have the effect of making your music sound like it was played by high-class studio musicians who, consciously or not, use feel extensively in their work. You can get your cousin to play drums on a date or you can get a top-notch session drummer. Given identical notes, the session drummer will make it feel great, but you may have a long afternoon with your cousin. Stated as simply as possible, feel can be the dimension that moves your music from the realm of a nice-sounding demo to a rompin' and stompin' master. Investigate its properties and your music will never be the same again.
© Electronic Musician magazine (2608 Ninth Street, Berkeley, CA 95470) and used with the kind permission of the Publishers.
About the author: Michael Stewart wanted to be a radio astronomer but got hooked on music at age 12 and eventually formed his own band. After extensive touring he became an LA session guitarist and later produced upwards of 20 albums, including Billy Joel's Piano Man. After teaching himself to programme computers, he designed the Human Clock for Kahler with advice from musician/programmer Vance Gloster.
Feature by Michael Stewart
Previous article in this issue:
Next article in this issue:
mu:zines is the result of thousands of hours of effort, and will require many thousands more going forward to reach our goals of getting all this content online.
If you value this resource, you can support this project - it really helps!