Seeing that psychedelia is all set to make a comeback, this month's Advanced Music Synthesis is all about putting life back into music or, rather, getting music out of life. Actually, 'advanced' is a misnomer, because we'll be looking at waveform generators that were around for thousands of years before the first sine wave oscillator crawled out of the primeval electronic sea into Stockhausen's lap - namely, Man's own biological rhythms. Before leaping into this quagmire of quirks and quacks, we'd do well to analyse the rhythmic connection from the musical starting-point. This less contentious angle has been the subject of a fair amount of work and reveals some pretty interesting insights into how our bodies react to music.

The plot thickens

As well as music being the food of love, it also goes to parts of the body that the other Arts don't seem able to reach. This is particularly true when one gets down to measuring such pointers to one's well-being as blood pressure, pulse rate, respiration and other autonomic functions controlled independent to the conscious mind. Reference points are all important if we're looking for music-induced changes and Figures 1 and 2 show, respectively, the respiration plot (amplitude vs. frequency of breathing) and electrocardiogram (ECG) of a normal, resting subject. All this nice, reassuring regularity flies out of the window as soon as music enters the picture. Figure 3 shows the influence of a snare drum roll, split equally into crescendo and decrescendo portions, on a particular subject's pulse rate. What's rather interesting is that this type of pulse plot is fairly consistent when the same piece of music is played several times to the same subject, i.e., a particular musical 'device' is capable of repeatedly switching some aspect of the body's physiology away from the norm. Furthermore, certain rhythmic devices - in particular, syncopation - desynchronise the regular heart action illustrated in Figure 2 by causing an extra-systole ('systole' being the phase of the heart's action when contraction sends blood to the body) or premature beat. In the case of Figure 3, one actually observes a 'driving' of the pulse rate from dynamic changes in volume of the snare drum. The incessant pounding of African tribal drums or, on a more banal level, the 125 bpm (beats per minute) of disco singles produces a similar sort of synchrony between an external 'pacemaker' and pulse rate, thereby inducing a state of blissful ignorance, euphoria, or divine awareness, depending on which escapist camp you care to belong to.

An interesting point about the preferred speed of the disco beat is that this is the fastest external pacemaker that will reliably synchronise the pulse rate of the majority of the bopping population. Anything faster than 125 bpm actually has the opposite effect and slows down the heartbeat. So, there's obviously a fine balance between causing intense physical excitement and inducing a trance-like state when it comes to laying down a pacemaking drum track - something that disco producers found out more by error than trial!

Music is also found to have a pronounced effect on muscle activity. Tapping one's feet is one more or less subconscious manifestation of this, but, more generally, there's a low level ticking-over of muscle twitches throughout the body when you're awake - even though there may be no obvious movement of the arms or legs. If electrodes are placed on the skin above a muscle or set of muscles, this activity appears as an electromyographic (EMG) plot (Figure 4), and one's actually observing the electrical activity of the thousands of fibres that make up each muscle in the body. This activity occurs at a fairly high frequency (100 to 500 Hz) and, for measurement purposes, it's usually integrated into a more convenient signal. Figure 4, then, shows the change in rate of muscle activity in the forehead region and in the legs in response to various musical conditions imposed upon the luckless subject. At the start (marked 'Beginning'), there's an equal balance between activity in the two parts of the body.

Moving to dance music, activity shifts to the legs (hardly surprising, perhaps, but remember that foot-tapping was verboten); with maths, the opposite occurs. Normal level Bach (Brandenburg Concerto No. 6) again predisposes the body towards activity in the leg muscles, but loud Bach sends the leg muscles off the scale into hyperspace - and this was still with no actual movement of the leg. Figure 5 shows both muscle activity (EMG) and pulse rate (P) coincident with different aural stimuli ranging from the nerve-grating to sublime. Once again, Bach scores high in terms of muscle activity, but Kalinka drum music wins outright with its devastating effect on pulse rate. These measurements can be extended to cover other parameters that the body gives up easily. We then end up with the multi-plot or polygraph shown in Figure 6, something that's very similar to those used in the U.S.A. for lie detection.

It's obvious, then, that music is capable of eliciting profound physiological changes - especially if certain rhythmic or harmonic 'tricks of the trade' are used to engineer a response from the listener. One of the national newspapers has recently been encouraging readers to submit their nominations for passages of music with high 'tingle quotients' (TQs), i.e., those that send a delicious shiver up the spine. Some of my favourites are the development section of the first movement of Rachmaninoff's Third Piano Concerto, the entry of the chorus at the start of Bach's St. John Passion, the "Maman" sung to a descending 4th at the end of Ravel's L'Enfant et les Sortilèges and the climactic end to David Bedford's Star's End. These TQs owe their effect to the craft of the composer and the alarming susceptibility of our bodies to the ultimate psychotropic drug - the manipulation of common-or-garden sound.

Feedback

The sure-fire test of a 100% guaranteed TQ would be to equip an entire audience with buttons and get them to fire them at the appropriate spine-tingling moment of a piece of music. If each button then triggered a light above the person's seat the orchestra would get a pretty immediate feedback on how they were doing as regards TQ communication. Not, perhaps, the most meaningful of exercises! Another approach might be to wire up the conductor with sensors so that his physiological reaction to the music he's conducting could be passed on to the audience via suitable displays.

In a paper published in 'Music and the Brain' (see acknowledgements), two psychologists suggest that the artist's capacity to relay an emotional impact may be actually more important than artistic perfection as far as success and acceptance by the public are concerned. So, maybe there's some sense in providing the audience with a short-cut to appreciating what's going on inside the mind (and body) of the performer.

Legitimate use of the body's autonomic response to music could also include light shows driven by sensors attached to each member of an onstage rock group. Doing this might reveal quite a lot of information about how rock musicians interact with each other: for instance, a good drummer might have the ability to drive the pulse rates of the other musicians to an equal extent and therefore ensure synchrony of their autonomic responses as well as the basic rhythmic fabric of the music. Conversely, the average keyboard player might find himself stuck out on a less responsive limb because of the restrictions a multi-key board set-up imposes on freedom of movement.

Putting plots into practice

It's rather fortunate for those interested in experimenting with these ideas that the majority of autonomic functions can be measured with fairly simple apparatus. Figure 7 provides some information on the origin of these music-influenced functions, the changes that are observed, and how they can actually be measured. Temperature is probably the least useful function because any changes are slow to occur and rather dependent on the prevailing climatic conditions. However, surprisingly large temperature fluctuations have been observed - especially if biofeedback assistance (pairs of thermistors, bridge and centre zero meter) is used to make you aware of differences as they're happening. Transcendental meditators actually developed this temperature offset trick to a fine art and were able to lower the temperature in one hand, with respect to the other, by as much as 2°C!

Galvanic skin response, as measured by skin resistance, makes much more sense for the type of applications we're interested in. Generally speaking, a relaxed body is reflected in a high skin resistance (which may go as high as 500k) and any anxiety or stress rapidly lowers it. This is easily proven with the simple circuit in Figure 8 built around a unijunction transistor. Reasonably effective electrodes can be constructed by wrapping wire around two fingers, though it's preferable to use proper skin electrodes attached to the palms of the hands with conductive jelly. The preset adjusts the initial pitch of the oscillator and this will be further lowered the more the resistance between the electrodes increases. Battery operation is essential in this and all other cases where electrodes are in direct contact with the body. I take no responsibility if you choose to do otherwise! Tone generators are all very well for biofeedback, but the galvanic skin response can be used much more effectively to control modules in synthesisers by means of the sort of voltage-processing circuit shown in Figure 9. This can be used quite successfully to superimpose a music-induced galvanic skin response on VCF tracking, LFO speed, or whatever.

Temperature and skin resistance are examples of aperiodic autonomic changes, i.e., there's no clear evidence of any short-term cyclic characteristics. In the long-term, though, humans (and other warm-blooded animals) do show small fluctuations of body temperature throughout the day and especially at night, but warm-blooded physiology demands that this diurnal rhythm is kept well in step and certainly not too disturbed by something as frivolous as music!

Biorhythms more reminiscent of musical waveforms can be found in the other three autonomic functions described in Figure 7. Heartbeat provides a low frequency waveform that is by nature highly unsymmetrical owing to the biphasic contraction of the heart. Respiration produces a waveform that is more regular in shape but about a fifth the frequency of heartbeat. Muscle activity, on the other hand, gives relatively high frequency electromyographic signals.

Provided that one's content with a simple on-off pulse, the easiest of these three to measure is respiration. This is because Honeywell have recently introduced an ultrasensitive air pressure switch that is activated by an air pressure of only 0.02 psi (equivalent to a gentle puff from a distance of a few inches). The only drawback of the switch is its limited current rating of 10mA DC, which means that, in most applications, external buffering will be required. To use the switch, one merely has to tape a length of narrow gauge tubing (e.g. aquarium tubing) under a nostril and attach the other end to either the high (0.05 psi) or low (0.02 psi) port of the air switch. And, if you're clever, two of the switches can be used in parallel with one triggering on expiration and the other on inhalation! These respiration-derived pulses could be used for triggering lights via a SCR, switching FX units on and off, stepping through a sequencer, and so on. Remember also that you'll then be putting yourself into a musical biofeedback position - the faster you play the greater the respiration rate the faster you play - until you collapse in a heap in the middle of the stage suffering from exhaustion!

This was, in principle, the basic idea behind a piece written by the oboist Heinz Holliger for oboe and electronics. Here the player uses his own amplified heartbeat to provide a click-track for his performance - faster playing raises the pulse rate and vice versa - hence the title, 'Cardiophonie'! This spiralling biofeedback situation is an example of positive feedback - the same stuff as chain reactions are made of - and clearly control elements have to be introduced into the loop to make it more meaningful and practical - from both the musician's and audience's point of view.

This brings us to the two conventionally "difficult to measure" autonomic functions, heartbeat and electromyographic activity. Fortunately, several electronics magazines have recently published good designs for monitoring both pulse rate and muscle activity. In clinical situations, cardiac electrodes are used to detect the electrical activity of the heart, and this gives an ECG (as in Figure 2) showing detailed information about how the heart is functioning. However, many isolation precautions may have been taken, there's always the outside chance that such electrodes could momentarily engage in a handshake with the mains and thereby cause cardiac arrest. Take heed from the tales of electrocuted guitarists and leave this technique to medically-qualified personnel!

A sensible way of picking up the basic characteristics of the heatbeat is to use the infra-red light-absorbing properties of blood. Since blood flows through capillaries close to the surface of the skin reflectivity at IR wavelengths is inversely proportional to the blood influx, and therefore changes cyclically with each heartbeat. Suitable optoelectronics in a light-excluding band attached around a thumb will extract this information painlessly, and then it's just a matter of conditioning the signal into a state suitable for human and/or musical consumption. In the case of the heartbeat, this involves filtering out noise (low-pass filter), removing mains hum (twin-T filter), and then subjecting the signal to high gain amplification and further filtering before it can be output to the waiting world.

Collecting electromyographic activity involves applying rather similar signal treatment, but here one has no option other than using skin electrodes and conductive jelly as these signals are fractions of a microvolt in amplitude. Figure 10 outlines the salient features involved in processing these types of bioelectric signals. Having extracted these signals, it's then necessary to consider how they can be used in our pursuit of getting music out of life.

Some clues as to a suitable approach can probably be gained from looking at the Bio Activity Translator (see Figure 12) produced by Jeremy Lord Synthesisers. This unit was originally designed for the somewhat dubious role of translating bioelectric potentials present on the surface of plant leaves into 'musical' sounds, though it's also perfectly feasible to use it for looking at muscle activity. Figure 11 shows the block diagram of the Bio Activity Translator. The conditioned bioelectric signal is used to increase the gain of a VCA via a simple diode envelope shaper. It also turns on a pulse generator, which in turn triggers a sample and hold to produce a stepped control voltage that tracks the bioelectric signal. The addition of a VCO in the chain results in pitch, speed and volume of the output all being controlled by the original input into the translator.

It would take a pretty fertile imagination to see anything remotely musical in the translated sounds of a plant, or my pectoralis major come to that, but that's really because taking pitch information from these bioelectric signals is bound to be a haphazard affair. What makes much more sense is to use these very rhythmic bioelectric signals in a way that complements the musical events that caused them to change in the first place. Controlling VCFs, VCAs, sequencers and light shows seems a less contentious application of all this bioelectric potential. A future bioelectric concert might consist of a trio wired up to banks of preprogrammed sequencers, synthesiser modules and lighting controllers, but without a single instrument in sight. Whether or not anything emerges that could be described as musical must depend on what musicians choose to do with the newfound self-awareness of how their bodies react to the music they're playing. Just remember that, in the words of that song, "the rhythm of life is a powerful beat"!

Sources: skin electrodes/conductive jelly from Wye Valley Electronics, (Contact Details); Honeywell PSF100A air pressure switch (cat. no. 41,623, price $7.00) from Edmund Scientific, (Contact Details)

Acknowledgements: Heinemann Medical Books for permission to use/adapt figures from a paper by Harrer and Harrer in 'Music and the Brain' (1977), pages 202-216; Jeremy Lord Synthesisers for loan of the Bio Activity Translator.