Recent events have given me cause to believe that professional audio is losing its way. It’s my concern that we are unconsciously heading in the direction of smeared indistinct sound as opposed to striving for crisp intelligible sound.

A major part of the difference between the two conditions of reality and soft focus is a sound system’s transient response. If the transient response is poor, definition and dimension will be missing, speech will lack intelligibility, musical instruments will be confused and the sound source will seem distant, culminating in what can only be described as audio mush.

The word transient as used in audio usually means a short period of peak level with a fast rise time, such as a snare hit. Transient information is the definition and excitement of audio and is as important a parameter as frequency response.

To my understanding, transient response is the ability of any part of the audio chain to reproduce the leading edge and accommodate the dynamic range of a given waveform. Attack is the nature of the leading edge and rise time is the time taken for the signal to go from minimum to maximum level. Incidentally transient response in amplifiers is known as slew rate and is measured in volts per micro second. Sounds are quite variable in their transient demands.

Percussion instruments are by their very nature demanding and audio components that cannot keep up will compromise definition of the sound. Therefore good transient response is fast acceleration and deceleration. Devices need to stop as fast as they accelerate to avoid ringing on and interfering with the next part of the audio signal.

Good transient response not only delivers definition, accuracy and excitement, but also contains a wealth of information about distance and direction of sound. To retrieve this information the human auditory system is capable of discerning and processing an amazingly fine-grained level of time resolution.

This is demonstrated (I am informed by Dr Peter Lennox of Derby University) by work done to determine the minimum detected angle of vector shift. In other words, how much a sound source has to move before human hearing detects a change in location, using the difference of arrival times between the two ears. This turns out to be as little as one or two degrees in the horizontal plane which represents a timing arrival difference at the ears of 13 to 18 microseconds, which is 13 to 18 thousandths of one thousandth of a second. Contrast this with a movie frame rate of 28 fps or 36 milliseconds, which is over two thousand times slower.

I make this point to illustrate the highly refined nature of human hearing, which begs the question as to why it is so often short changed in favour of the less refined visual sense. Why are we so ready to accept poor sound when nopne of us would ever accept grainy or smeared visuals?

To get back on track, this amazing sensitivity to time information is naturally focused on the boundary conditions of the leading edge or attack of the source, which is where all the action is, including information about distance.

Compromised transient information has the perceived effect of placing the sound source at a distance and removing its dimensional qualities, which in turn impairs the feeling of involvement. Therefore the ability of a chain of audio equipment to follow a given audio signal in all its detail is paramount if we are to achieve reality. This ability usually falls over when we get to loudspeakers.                                           
Loudspeakers being electro-mechanical devices have a tougher time than electronics in keeping up with the dynamics of waveforms. Loudspeakers are the traditional bottleneck, which is the reason why my partners and I have spent 30 years researching and developing fast speakers.

Slow speakers compromise the perceived audio by slowing the rise time to the point where the transient can be over before the loudspeaker ever reaches the intended peak level. In this way the dynamic range is also compromised as well as resulting in the previously discussed problems.

For good transient response it is essential for the loudspeakers to be individually fast. However, when combining speakers to make a larger system if their configuration introduces multiple arrivals resulting in time smearing this also compromises transient.

System configurations that produce multiple arrivals will have the effect of smearing the attack and even affecting the perceived frequency response. I have found this to be very evident in the case of line arrays, particularly in the voice range where the obfuscation of defining consonants renders entire vocal performances unintelligible. Transient smearing is a different mechanism to slow rise time, but the effect nevertheless compromises the perception of definition.

In the last 15 years, line array has become the paradigm of professional loudspeaker systems. By and large they are all inspired by Christian Heil’s development of the classic line array arrangement and mostly employ multiple direct radiators and wide dispersion HF devices which, ideally, collectively reinforce each other by constructive interference although, of course, they also suffer from destructive interference. Direct radiators have inherently low efficiency because of the substantial impedance miss match between the moving diaphragm and the air. Although the low efficiency improves by the mutual coupling it is at the expense of introducing multiple arrivals.

The difference in arrival time between the top and bottom boxes can be as much as five milliseconds, but more disturbingly, the difference between adjacent boxes of 60 microseconds is at least three times greater than the lower limit of human time perception. To make matters even worse, any frequency which has a half wavelength smaller than the distance between the centres of the drivers will not couple at source and so will be prone to off axis comb filtering which is evidenced by the large amount of phasing and instability that occurs when line arrays are working in even slightly windy conditions.

Phase shift is measured in degrees of arc around a circle which looks flat, but mathematically the situation is in fact three dimensional with the circle of phase angle being an end on view of a spiral or helix, the side ways view giving us the familiar sine wave as shown in the diagram. This means that phase angle also represents a very small increment of time which is why corrective EQ is not only an admission of defeat but adds further confusion to the all important leading edges of sound. In fact, line array resolution is so bad that it has resulted in a generation of audio engineers using very blunt tools, which has provided the perfect smokescreen for all sorts of dreadful digital equipment to gain a very questionable hold on FOH.

There are political and psychological advantages to this, in that indistinct and two-dimensional sound can be likened to soft focus photography for smoothing out the blemishes. This reduces the dynamic range between the good and bad engineers, preventing good ones from excelling and letting bad ones get away with it. This combination of line array and inadequate digital equipment is what has brought pro audio to its current sorry state.

The sad outcome of all this is that performance audio has not properly progressed in decades and the audiences (the people who actually pay for the music industry) are being short-changed in audio quality on top of ever more draconian sound level limits and miserable weather. There are many reasons, but no valid excuse for mediocrity and so, to those professionals that actually care and want to take pride in the quality of their workmanship, I ask: do we want reality or are we more comfortable with boring undemanding two dimensional soft focus?