Peter Eckersley (06-01-1892 - 18-03-1963) was the first Chief Engineer of the BBC.
He knows whats what, he knows hi-fi; Is not a true FideliIy . (Adapted)
What a term! How can fidelity be high? Lack of it can stink to high heaven but that hardly justifies an opposite. Perhaps it is something raised up, usually in volume! But let that pass, 'we know what we mean'. In my young day we used to speak of quality, good and bad, or, if faithful is the code word, of faithful, reproduction.
But do we know what we mean? I know that perfect reproduction would be that which caused a loudspeaker to create a field of sound around a listeners ear identical to that existing around the ears of an individual situated in the auditorium, studio, or whatever, where the reproduced sound originated.
Fig. 1. Illustrating the obstacles to fidelity of reproduction.
The diagram of Fig. 1 helps the understanding of the definition and is an aid to an explanation why truly faithful reproduction, according to any means known to me, cannot be achieved.
For the sake of example we postulate an orchestra spread around one end of the auditorium and a microphone (M) facing it. This microphone is connected by a single channel to a loudspeaker (L) placed opposite to the listener, in the room where he listens.
Apart from any distortion that may be created by the transducers and in the channel connecting them the principle inherent artificialities, which militate against perfect reproduction (hi-fi to you) are:
- The acoustics of the room in which the loud-speaker is situated are superimposed on those of the auditorium.
- The source of the sounds impinging on the listeners ears is a point source, the sources of sound in the auditorium are spread over a relatively wide area.
- A minor cause of distortion is produced because the microphone, not being the shape of a human head (and not having two ears) must in some degree alter the composition of the sound field from its form as it would be created around the ears of one listening in the auditorium.
Neglecting for a while the problem of super-imposed acoustics (paragraph 1 foregoing) the artificialities of a point source of reproduction and a single microphone (as distinct from two ears) it has been suggested (and the suggestion taken up in modern equipment) that more faithful reproduction would arise by the use of so-called stereophony, consummated by the use of, typically, two microphones, two channels and two loudspeakers. Dr Leakey has more than adequately discussed the possibilities in a recent article (April and May issues).
While it may be, and often is, claimed that two-channel reproduction is an improvement we must nevertheless appreciate that it cannot achieve the ideal of true fidelity.
I recollect, and this, though it is related to 'binaural' rather than 'stereophonic' listening, may be of some academic interest, that in the very early days of broadcasting, when we transmitted opera from Covent Garden, H J Round set up two microphones spaced feet apart, among the foot-lights, and connected each one to each earpiece of a two-earpiece headphone. It was remarked that as a singer moved across the stage parallel or at some angle to the line joining the microphone he (or she) appeared to us wearing the headphone to move not from side to side but in an arc above our heads. Thus if one kept ones eyes shut one looked upwards!
One of the more dramatic effects of stereophony is the verisimilitude of movement of a sound source. Properly located in relation to the loudspeakers the listener hears an aeroplane flying over his room diagonally or a speaker appears to move from side to side. Orchestra players, however, sit still, but it is claimed that two-channel broadcasting adds realism in the sense that, for instance, the fiddlers do appear to play to one side, the wood wind to another; is there a claim for depth?
An American friend of mine, writes to me and starts a paragraph with the words 'Why Stereo?' and goes on 'Dr Harry Olson wrote an interesting paper . . .' on 'the psychological response to monaural' (sic Mr Editor) 'low-fi' (sic), 'monaural hi-fi' (he is unrepentant) 'with several spaced speakers . . . playing the same record; binaural fringe-channel two-speaker reproduction' and 'filled-in' binaural hi-fi, three channel, with speakers respectively playing the left-ear channel, the right-ear channel, and (at intermediate position for the speaker) the mixed left and right-ear channels.
'The results indicate successive improvements between each of these and the preceding but very unequal steps. The big jumps were to hi-fi and spaced speakers whether two . . . or single channel.'
I envy Dr Harry Olson, he must have had a lot of fun; I would join in it more thoroughly, however, were I better acquainted than I can be, without a sight of his paper, with the meaning of some of the terms he uses.
It is time for a confession - quite simply I do not find any real improvement between any single- and any multi-channel reproduction I have heard demonstrated and I hasten to say that many, whose powers of observation and whose integrity in expressing their opinions I respect, hold different views; I must add that some equally competent people agree with me. There is, without doubt, a difference between the two systems but to me it is no more than a difference; it is not an improvement.
I would characterize this difference as giving one a feeling that the sound from two-channel reproduction is more diffused, it is fuller than when the single channel system is compared. But, with a limited number of observations, I have remarked this same improvement when two or more loudspeakers are energized from a single-channel source. Moreover I would say that those loudspeakers which are designed to be facing the corner of a room and are responsive to single channel energization give much the same effect. It is, I repeat, a difference that I observe, and a pleasing one in some instances, but it is still artificial, reminding one of reality rather than copying it.
No! What I believe is that the primary need in improving reproduction is not so much a perfect copy of the original but rather a result, limited in certain respects as it may be, which pleases because it is free from the intromissions of the several types of non-linearities, and is unaccompanied by extraneous noise.
Art and Artificiality
This is where I mount my hobby horse and discuss art and artificiality. There is some rather involved phrase implying that the object of the artist is to conceal art. It is doubtful if artists are objective, but if the sense of the statement is that art produces emotion in those who appreciate it without the means to this end being obvious (and certainly being artificial and distinct from realism) then one can agree with a supposed meaning.
When it is seen how a two-dimensional picture can represent a three-dimensional subject, or how in statuary without loss of the value as art, dimensions are shrunk or expanded below or above those of reality then the artificiality of these forms of art are its obvious characteristics. The artist paints not what he sees but from a sub-conscious which tells his hand to register the emotion a scene conjures in it. This is not realism but it can be good art. I must say, in passing, that it would help if some painters would issue a guide to the operation of their sub-conscious; it is not always easy to join in as it were. Is it, as a final example, necessary to call attention to the artificiality of the theatre and to its impact as an art?
This may seem to have wandered a long way from hi-fi, but surely not. Surely in broadcasting there is on the one hand an artist creating a programme and upon the other the person upon whom an effect is produced and between the two a medium, a means, indeed, an artificiality, namely the technique of 'reproduction'. In television, as in the film, it is again the two-dimensional image that creates a three dimensional impression; in sound it is more usually the point not the diffused source which stimulates the listeners sensibilities. And provided always that these artificialities are such as to convey reality without precisely reproducing it, and provided in so doing at least some of those who look or listen are moved, and provided, in other words, their sensibilities are awakened, what more is required?
But if the artificiality of reproduction has added to it the distractions of dissonances and the peevish introduction of irrelevant noise then its value as an art is at least reduced, in some cases destroyed. So in discussing hi-fi, I would count it of greater importance to consider chiefly the effect of the generation of harmonics, and combination tones and the presence of noise than what, in a cynical mood, I describe as the sales gimmick of stereophony.
Fig. 2. Fletcher-Munson curves of the upper and lower limits of hearing, with superimposed (dotted) response curves of a poor-quality receiver at two different volume levels.
Based on Fig. 70, p. 141 of 'Speech and Hearing', Harry Fletcher, (Macmillan and Co., 1929).
In discussing the kind of distortion I have in mind it helps to consider the graphs of Fig. 2. In this figure the ordinate represents pressure (a scale of decibels is also shown) and the abscissa: frequency on a log scale. The upper full-line graph shows sound intensities at which the ear experiences pain, or which stimulate the sensation of feeling; the threshold of feeling is therefore the intensity at which the listener starts to experience painful sensations.
The lower graph delineates the threshold of audibility, intensities less than those shown by the graph are inaudible. It is remarkable that the maximum pressure at the threshold of feeling is some 130dB greater than the minimum of audibility. The ratio of fortissimo to pianissimo of a symphony orchestra is of the order of 80dB. The contrast ratio of hi-fi reproduction is of the order of 40dB at maximum; if it were more faithful to the original with respect to contrast ratio pianissimo passages would be masked or spoiled by noise. Fi cannot be so hi as some would believe.
The upper broken-line graph in Fig. 2 is drawn to illustrate the performance of a poor quality receiver; typically one designed for medium wave reception or perhaps one of those little snarlers that use transistors and pick up anything anywhere without visible means of aerial.
In drawing the graphs of Fig. 2 we have assumed 'attenuation distortion', ie 'distortion due to variation of loss or gain with frequency'; in jargon terms the frequency characteristic is not 'flat'. Another assumption, alas by no means unjustified, is that there is considerable mains hum (frequency 100Hz).
The lower dotted graph illustrates a contrast ratio of 20dB, may be a pessimistic value, but not greatly so for medium-wave broadcasting.
We notice that the reproduction of the upper and lower frequencies varies with the volume knobs and this may explain, if it does not excuse, why the user of an inferior type of set turns up his volume. It is seen from Fig. 2 that as volume is increased so the frequency gamut is increased but now the middle frequencies must approach nearer to sensation level. The result may well be to overload the audio amplifiers with a consequent introduction of amplitude distortion 'the lack of constancy of the RMS value of the output of the system to that of the input at different amplitudes of the input', also of harmonic distortion 'the production of harmonic frequencies at the output by the non-linearity of a network when a sinusoidal voltage . . . is applied at the input'. In other words turning up the volume produces a harsh and unpleasing result marred by the introduction of spurious frequencies not existing in the original.
Noise and Bandwidth
It is also remarkable that as the volume is increased the bandwidth of reception is also increased and so any noise picked up is also increased - 'the wider you open a window the more dirt that comes in'. Maybe this noise is masked by the greater intensities of speech or music, while this is transmitted, but during quiet passages or during pauses noise is annoyingly audible.
Perfect quality would be represented by points lying within the lozenge-shaped area, indicated by the full lines of Fig. 2; it would demand a frequency characteristic, including the loudspeaker, which was flat between, say, 30 to 16,000Hz, freedom (to, say, 80dB) from harmonic or amplitude distortion and a contrast ratio without the introduction of noise of, say, 80dB.
There is another form of distortion which may or may not be audible, namely, phase distortion 'distortion due to variation of the group velocity of the system with frequency' and, as explained later on, a distortion associated with a Doppler effect in the loudspeaker. There is also the effects due to hangover of oscillation of the loudspeaker diaphragm. We know that if the frequency characteristics of a system is flat then the group velocity of waves transmitted through it is constant; phase change is then proportional to frequency. If, however, the effects of reactance are present, causing a variation of the ratio of the output to the input of the system with frequency, then inevitably phase distortion appears. This is why some argue that the frequency characteristic of the, amplifiers in a receiver should extend to, say, 100kHz and then fall off gradually. In common practice cut-off is allowed just above the highest frequency it is desired to reproduce. Whether this effect, other distortions being eliminated, is audible, I do not know.
The Doppler principle is that which makes the frequency of waves appear to change when there is a relative velocity between the wave source and the observer. Thus if a loudspeaker diaphragm is moving as a piston at a low frequency, and if it is simultaneously reproducing a higher frequency than there is, so far as the higher frequency is concerned, relative velocity between observer and source; the higher frequency is thus frequency-modulated by the lower. The degree to which the effect is audible is not known to me; it may well be negligible.
But there are more things in transducers and amplifiers than are dreamed of in some philosophies; flattening the frequency characteristic, as judged by audio oscillator and output volt-meter (electrical or acoustical) is a step on the way, but there are other side effects which this simple test cannot remark. For example, there is the hang-over of the diaphragm of a loudspeaker which, given a steep wave front, continues to oscillate long after the impulse which sets it in motion has died away. By the same token it will not immediately and therefore properly respond to a steep wave-front. These are effects which subtract from good quality by robbing music of its attack, its crispness, which when present is an engaging characteristic of good reproduction.
There are other distortions which defy analysis; recounting an experience may illuminate my meaning.
A friend, whose judgment of quality is of a very high order, installed an allegedly hi-fi single channel radio-gramophone which, at first hearing, pleased him. The same impressive housing beautiful wood, discreet lid cushioning into place, contained tuner and turntable while a set of loudspeakers, contained in what might be described as a cupboard, radiated their output through an elegant grill.
As time passed my friends satisfactions diminished until, thoroughly disillusioned, he decided to install a separate loudspeaker in substitution for those boxed in what I have described as a cupboard.
A change-over switch allowed a comparison. It only needed to be operated once to demonstrate excellent as compared with indifferently good quality. Incidentally, the single loudspeaker which gave the improved quality was one which faced into the corner of the room.
What can one conclude and what more when it is recounted that the substitution of the tuner and the gramophone pickup by others of different design made a further improvement? Perhaps all is not fi which is described as hi; except the latter abbreviation did truly qualify the cost of my friends set. I can hear my critics saying, 'What after all have you said, that the elimination of harmonic, amplitude and phase distortion is essential? We hi-fi experts are quite aware of that'.
I reply 'Yes! But if you are why can someone buy an expensive hi-fi equipment and find it lamentably wanting and why do I and others feel that stereophony is no more than a gimmick, not a fundamental improvement?'
Another critic might exclaim, 'Are you so simple as to neglect the cost factor? Its all very well to ask for a wide frequency gamut, amplifiers free from distortion, elimination of mains hum, but have you considered the cost?'
'I am', I reply, 'quite aware of the cost factor and that is one reason why I have continuously and persistently (without making much impression) argued the merit of wire-broadcasting'. Let me once more, in the light of the foregoing, point out its advantages with respect of reproduction.
Essentially, given a conductor, joining programme source and loudspeaker, the receiver can be simpler than when radio is used. In audio frequency technique the receiver is no more than a loudspeaker; if a carrier frequency method is essential, then the received level is not a few, but hundreds of milli-volts and the receiver is consequently cheaper and gives better reproduction in spite of its decreased cost.
In sum, while I respect those who believe that stereophony represents a major advance in the art of reproduction, my own ears fail to notice more than a difference, not an improvement. I am not alone in this belief.
The greatest step towards hi-fi would be that which concentrated on removing distortions due to non-linearities and the effects of noise from the average receiver.
A wider application of wire broadcasting would be a major advance towards hi-fi. It would also perhaps be easier, because of its facility to provider more channels, to introduce stereophony with wire rather than radio broadcasting.
As a final and possibly 'tantrumistic' contribution to the subject I must air a grievance. What is it that turns ordinary decent folk, once they get their hands on the steering wheel of a motor car or the volume knob of a loudspeaker into sadists demonstrably hating their fellow men? As one who suffers from my neighbours ever-louder speaker, I pray that the designers of reproduction equipment should limit volume output and should not give the user the excuse to increase it by the limitation of the frequency characteristic. If I were in charge of a wire broadcasting system I would deserve the thanks of many, because I would make it impossible for the reproduced sounds to exceed a certain level. I am aware that a reduced level may subtract from realism, but then I deny the need for realism. I would and do accept limitations both of contrast level, frequency characteristic, and volume, but I cannot abide the invasion of spurious tones; I want clean reception and the crisp reproduction of transients.
Good quality, as I define it, at a lower volume than may be theoretically desirable is, as I prove whenever I listen, satisfactory, but the quality must be good quality, and the operator of the set, like me, a good neighbour.
The following are extracts from Wikipedia and illustrates Eckersley's interest in delivering entertainment into the home by wire as opposed to the monopoly and morality represented by the BBC - known as Aunty. At the time there was no commercial broadcasting in the UK and Pirate Radio was a few years away.
Background: In 1929 Eckersley began an affair with Dorothy 'Dolly' Clark, estranged wife of the BBC programme planner and conductor Edward Clark and then divorced his own wife. Reith (head of the BBC), a deeply religious man, forced him out of his job. This affair prompted a public enquiry into the BBC's personnel practices.
For a time, after leaving the BBC, Eckersley was engaged in working to build a broadcasting station based in continental Europe which could be received in the United Kingdom. Captain Leonard Plugge, who became a Member of Parliament, also set about building his own International Broadcasting Company by leasing transmitters in France and other countries to beam commercial radio into Britain. The venture was very successful and, because Reith had banned Sunday light entertainment on the BBC, the IBC stations gained as much as 80% of the Sunday listening audience by 1938.
From 1930: Peter Eckersley had sought other ways to bring the signals of the IBC stations into the living rooms of Britain. Rather than relying upon a receiving set licensed by the General Post Office, he began to wire parts of England for an early form of cable radio but was stopped by the intervention of the General Post Office.