Douglas W Robinson (22 July 1920-12 July 1999)

R Hinchcliffe 
 ,

Correspondence Address:
R Hinchcliffe
,

Full Text

The always eagerly awaited writings of Douglas Robinson will appear no more. Douglas' immortality ended on the 12th July of 1999.

Douglas was born and bred a Londoner. He studied electrical engineering at the City and Guilds' Engineering College of London University's Imperial College of Science and Technology. Douglas graduated with First Class Honours. He also took the Henrici Medal for Mathematics. Aside from a brief spell with the Western Electric Company further academic and professional training was disrupted by World War II.

Having been commissioned into the Royal Air Force Volunteer Reserve, he was posted to the Far East where he had the misfortune to be taken prisoner. His experiences left him with "lasting scars shared with no one but himself".

Douglas spent the major part (1950-1981) of his working life at the UK Government's prestigious physics institution, the National Physical Laboratory, near London. From 1959 he was responsible for all the NPL work in acoustics. Prior to joining NPL he had worked for a time for Standard Telephones and Cables which work had stood him in good stead for his subsequent researches.

I first met Douglas forty years ago when he attended his first International Congress on Acoustics (The Third ICA, in Stuttgart, in 1959). He presented papers on "The Loudness of Directional Sound Fields", and the "Limits of Precision in Pure-tone Audiometer Calibration." The two papers symbolized the two main research themes that occupied his working life, i.e. the subjective magnitudes of sounds, and the auditory threshold together with the factors (particularly, noise) influencing it. The first paper was a shortened version of a longer report with his co-worker LS Whittle which was published in the following year in Acustica. The second was published in full in the Annals of Occupational Hygiene as "Variability in the Realization of the Audiometric Zero". He said "The problems of pure-tone audiometry can be conveniently grouped into three distinct phases. The first and most fundamental of these is the determination of the absolute threshold of hearing for normal listeners. The second is the transfer of this information to the dials of practical audiometers, which we refer to as the realization of the audiometric zero. Finally there is the actual technique of hearing loss measurement, which is principally though not exclusively to be thought of in terms of clinical application�The importance of an internationally agreed set of standard values for the normal threshold of hearing, as a step towards interchangeability of clinical information, is widely recognized, and the task of formulating such a standard has been undertaken by the Acoustics Committee of the International Organization for Standardization (ISO)." The standard appeared as ISO 389.

Douglas' papers illustrated the precision of his approach to acoustical measurements and his dedication to calibration and standardization that dominated his professional career. It was therefore inevitable that he be asked to sit on, and often chair, one or other national or international standards committee or working group.

By the time that he presented his papers in Stuttgart he was well established on the staff of the National Physical Laboratory. With RS Dadson he had already undertaken a re-determination of the equal-loudness relations for pure tones. This had been a prodigious task which had occupied him for four years. The new results covered a range of frequency of from 25 Hz to 15 kHz and of sound pressure level up to about 130 dB SPL; these data were used to derive ISO 226: 1961 (Specification for normal equal-loudness level contours for pure tones under free-field listening conditions).

In 1957 he had published the results of his studies on loudness scaling, confirming the contemporaneous Harvard studies that loudness increased as about the 0.3 power of sound intensity (0.6 re sound pressure). The sone scale became incorporated into an international standard (now ISO 532: 1975).

In the late 1960s with LS Whittle he undertook studies on the standardization of the bone conduction threshold. Their design for an artificial mastoid for the calibration of bone vibrators was adopted by BSI as BS 4009: 1966. Douglas was to return to the subject of bone conduction measurements a quarter of a century later. With colleagues from MRC's Institute of Hearing Research he wrote on the relevance of a change in the standard for calibrating bone conduction audiometers BS 6950: 1988 (now superseded by BS ISO 389-3:1994) which specifies the standard reference zero for the calibration of pure tone bone conduction audiometers (equivalent to ISO 7566-1987). That standard uses a method of measurement different to that of the standard that it replaced, i.e. BS 2497, Part 4: 1972. "The new standard will typically yield BC thresholds that are about 3-4 dB less acute averaged over 1, 2 and 3 kHz as compared with the old standard, and about 5-8dB less acute at 1 kHz in particular, depending on the date of manufacture of the mechanical coupler used for calibration�Taking account of all the possible sources of error, only an air-bone gap averaged over 1, 2 and 3 kHz greater than 15 dB can be regarded as significant, when measured with an audiometer whose output is calibrated to BS 6950: 1988. When an audiometer has been used that is calibrated to the earlier standard (BS 2497, Part 4: 1972) only an air-bone gap greater than 20 dB should be regarded as significant."

Douglas also turned his mind in the 1960s to quantifying environmental noise. He showed that a single index, the "Noise Pollution Level", can accommodate survey results for aircraft and for motor vehicles. The index is based upon two terms, one representing the equivalent continuous sound level on the energy basis, and the other representing the augmentation of annoyance when fluctuations of the noise level occur. An NPL report on motor vehicle noise appeared as Appendix IX, and one on aircraft noise as Appendix X of the Wilson Report on Noise (Cmnd. 2056). Jet engine noise, helicopter noise and the sonic boom subsequently received particular attention. One of the last studies was when Douglas' group looked at an impulse noise correction to Effective Perceived Noise Level to take into account helicopter blade slap (Ac 93); this formed the basis of a draft ISO standard prepared at the behest of ICAO.

In an early 1970s essay he put forward a criterion by which the inevitable complications of a more general measuring scale may be balanced against the gain in precision over the existing formulation of the Noise Pollution Level. The Noise Advisory Council, which had been established in 1970, endorsed the use of the Noise Pollution Level for the assessment of environmental noise.

But the one subject that dominated his thoughts, work and life during the 1960s was the Medical Research Council/National Physical Laboratory investigation of noise and hearing in industry which had been commissioned in 1962 by the then Ministry of Pensions and National Insurance. This study, overseen with W Burns and TS Littler, provided the major source of British data relating hearing threshold levels to hazardous occupational noise exposure.

He and Williams Burns had recognized that there was not just one medical/scientific method to tackling an investigation such as this: "It is generally accepted that there are two different approaches to the kind of problem posed by this study�For convenience we will call the two the 'parametric' and 'incidence' philosophies, and it is the former to which we subscribe�the parametric approach is based on the proposition that the fundamental physiological characteristics of the hearing process are essentially the same for individuals. This basic similarity is overlaid by minor differences due to normal biological variability and within these limits therefore the stimulus-response characteristics are determinate and broadly alike for all persons. In practice, large departures from this state of affairs can, of course, arise due to disturbing factors including pathology of various origins. The experimental [strictly speaking all the epidemiological studies of this type are sub-experimental studies] approach is therefore to utilize for investigation only ears which are free from such disturbing factors so far as can be determined�The end-product of the immediate investigation can thus be envisaged as a specific relationship between a physical description of noise exposure and the resulting hearing level with, of course, statistical overtones."

This parametric approach examines hearing in terms of epidemiological/demographic factors that influence hearing threshold levels. It was thus inevitable that the first age/noise formula came from the National Physical Laboratory and included numerical values for noise level, exposure duration and age, factors which clinicians had noted to be relevant. It was also logical for a compact formula to combine noise level, expressed on the dB (A) scale, and duration using the equal-energy hypothesis which had been adopted by the US Air Force (1956) in AFR 160-3, into a single value, the "noise immission level". The "equal-energy hypothesis" means that equal energies (acoustic intensity � time) give rise to equal physiological changes regardless of the shape of the intensity time function. As has been pointed out by others, the equal-energy theory is the most attractive because of its simplicity. Halving the exposure time and doubling the intensity (increasing the level by 3 dB) would keep the level constant. A hyperbolic tangent function relating hearing threshold to noise immission level would, as he and JP Cook pointed out, be logically preferable (to a quadratic function) in that its limiting behaviour for zero and large immission is more plausible. A later model for the relationship between noise exposure, age and hearing threshold level discarded the equal energy hypothesis. Nevertheless NPL Ac 61 (usually referred to as "The NPL Tables"), remains a useful tool in the assessment of cases of alleged occupational noise damage to hearing, especially for retro diction. Data and concepts developed from the MRC/NPL Survey were used to produce BS 5330:1976 Method of test for estimating the risk of hearing handicap due to noise exposure, the HSE document Framing Noise Legislation, and the HSE's Code of Practice for reducing the exposure of employed persons to Noise.

In a 1976 paper discussing the characteristics of noise-induced hearing loss, Douglas said: "�in addition to the noise, numerous intervening factors also operate, all of which tend to obscure the principal relationships�General acoustic wear and tear on the organs of the internal ear, or socioacusis, in greater or lesser degree is a reality of modern life�so that neglect of non occupational noise is tantamount to a straightforward underestimation of the accumulated noise dose. Here, then at the simplest level of consideration is the possibility of quantitative error, perhaps significant, at the lower end of the occupational noise range�a dip in the average audiogram at 4 kHz is always the first feature to appear (we may nevertheless recognize individual deviations with dips at 3 or 6 kHz). This dip at first deepens and later flattens off due to the decelerating function; thereafter the hearing levels at other frequencies begin to catch up, beginning with 6 kHz because that is rising fastest with age. The dip then smears out or turns into a shape increasingly progressively with frequency, distinguishable from that of non-exposed persons only by the increased level at lower frequencies."

In the late 70s Douglas and his colleagues looked at hearing thresholds in a population from heavy industry. They reported: "Hearing levels for the whole group were compared with those of a sub�group having no auditory pathology other than that due to noise exposure. The difference in hearing levels is shown to account for divergences in prediction of the incidence of occupational hearing loss between ISO Standard 1999 and British Standard 5330. The numerical analysis highlights the care needed in comparing audiometric data from different sources and incidentally has significant implications for the setting of industrial noise exposure limits." Later in the same decade he continued (as he always had done) in a search for "pure and unsullied" hearing. Ac 89 found a "relation between hearing threshold level and deviations from otological normality as scored by a points system based on a questionnaire and otoscopic examination." Later, in a publication with Ben

Lawton, he endorsed the socio-economic factor: "�the value, being an average, must imply values both above and below 3 dB, up to 6 dB�"

During the mid-70s, the problems inherent in speech audiometry came to the attention of his group. The report (Ac 73) with the title "A Feasibility Study of Diagnostic Speech Audiometry" belied the searching and critical mathematical appraisal of the scientific basis for this form of testing hearing..

An analysis that he undertook with GJ Sutton in the late 70s examined the age effect on hearing. A formula was derived for predicting the changes in hearing threshold level of otologically screened groups with increasing age. The study provided the basis for ISO 7029-1984 Specification for the threshold of hearing by air conduction as a function of age and sex for otologically normal persons.

Working with the same collaborator he subsequently made an appraisal of the methods for estimating the effectiveness of hearing protective devices. They wrote "We consider the Botsford�type method in which the protection offered in a flat noise spectrum is used (with the mean attenuation corrected by 1 or 2 SD) is highly acceptable as the single-number procedure in terms of accuracy and simplicity. This rating is simply subtracted from the C-weighted noise level to give the closed-ear A-weighted level. Thus if we have a noise whose C-value is 109 dB (C), then we can be 95% sure that any protector whose single number rating (as defined above) is 19 dB or greater will protect 80% of wearers to below 90 dB (A)�It must be emphasised that a single-number rating based on this method must be used in conjunction with the C-weighted, not the A-weighted, sound level of the noise. The above method is thus very close to the EPA method except that here we include no "safety factor" correction-the need for this has been eliminated by the choice of the spectrum which gives the single number rating, and by a conservative adjustment of the percentage (84 to 80, and 98 to 97)."

After his retirement from NPL Douglas continued at the University of Southampton's Institute of Sound and Vibration Research the analysis of his MRC/NPL data. In one of his earliest Technical Reports (Audiometric configurations and repeatability in noise-induced hearing loss) from ISVR, he wrote : "Notch frequency and depth are highly correlated between right and left ears�the mean audiogram level (average HTL across 6 frequencies, i.e. 0.5, 1, 2, 3, 4 and 6 kHz) repeated on average to within � 3 dB�Positive and negative shifts were nearly equal in frequency of occurrence indicating that the principal cause was a random process, and this is attributed mainly to subjective uncertainty of the threshold."

A later report dealt with auditory impairment and the onset of disability and handicap in noise-induced hearing loss. The threshold of what was termed auditory inability (defined by the 2%ile performance on simulated listening tests) corresponded to a hearing threshold level of 38 dB averaged over 3, 4 and 6 kHz. However this threshold was less well defined than the corresponding inability threshold (30 dB) using the frequencies 1, 2 and 3 kHz. He expanded on these concepts in a Health and Safety Executive report three years later: "The commissioning of this study provided the ability to explore�the subjective consequences for the hearing ability of those affected �The term disability is not altogether appropriate in hearing conservation, however, and the writer prefers 'inability' to describe the beginning and lower levels of what is technically termed disability (this word generally conjures up the notion of more severe conditions). The existence of a disability may, in turn, give rise to handicap, the state of being at a social disadvantage with respect to one's peers�hearing conservation necessarily starts from the concept of disability�it is concerned with the preservation of intact ability rather than the loss of it. Hence the material concept is the 'threshold of inability'. Attempts have been made to identify this threshold by numerous experimenters�It is hardly surprising that the pure-tone level so identified (the so-called 'low fence') varies from experiment to experiment�A full treatment of this subject has been given by Robinson, Wilkins, Thyer and Lawes (1984) in the report of an investigation specifically aimed at identifying the threshold of inability and its relation, on the one hand to various audiological impairment measures, and on the other to self-rated handicap (though, again, handicap is perhaps too strong a term to describe people's hearing difficulties at the just-not-normal level)�The resulting value of HTL, average at 1, 2 and 3 kHz, was found to be approximately 30 dB. By coincidence this is the same numerical value as is used in BS5330 to describe the level above which a "handicap" is deemed to exist�50% of the possessors of this level of hearing loss are still just within normal limits of hearing performance for young persons. Much lower 'fences' have been canvassed by some authors, based on the level at which a test group shows a barely perceptible difference in performance from one with a smaller HTL. In the writer's opinion, such estimates are a case of chasing shadows." He also noticed the gradual submersion of occupational noise induced hearing loss within the ageing process: "The difference between the exposed and the non-exposed appears to be at its greatest between the ages of 45 and 50 years�the HTLs attained at the age of 60 or 65 years in the general population, when compared with those of noise-exposed persons of similar age, left little margin to account for the specific effects of noise."

At the beginning of the present decade he was a co-author of a proposal to use subjective magnitude scaling of auditory acuity to provide a basis for compensation. Such a scheme would provide a means for breaking through the barriers of the inability and disability thresholds together with the clinicians' and physiologists' range of normality to bring everyone into its ambit.

Around the same time he wrote that the concept of simple (linear) additivity (in decibels) of the various factors governing noise damage to hearing was no longer tenable (what might be termed "compression" occurs).

Later, with colleagues from University of Southampton's Institute of Sound and Vibration Research, he conducted a study of low noise level hazards for the Health and Safety Executive. A formula derived from the analysis showed a start level of 71 dB (A). However, the results were summarized as: "Curve fitting procedures�indicate a negligible effect at 75 dB (A). Above this level, but below 85 dB (A), long term exposure to noise has some effect but the amount of noise induced threshold shift is so small as to be practically undetectable in individual cases and only measurable in a statistical sense. Moreover, it is so small as to be overshadowed by the loss of hearing associated with advancing age, whether due to natural causes or the insults of daily living."

A final contribution was to co-author a glossary of technical terms used in audiology.

Towards the end of his life Douglas was coming round to the belief that the article of faith that he had enunciated forty years previously "The first and most fundamental of these is the determination of the absolute threshold of hearing for normal listeners" was indeed a quest for the Holy Grail of "pure and unsullied hearing."

Douglas' contributions were recognized by the receipt of the Wolfe Award in 1975 and the University of Southampton's Honorary Degree of Doctor of Science in Engineering in 1978. In 1965 he had proceeded to the DSc degree of his own university on the basis of published work on the subjective assessment of noise and the metrology of hearing. A useful overview of his many contributions to the environmental noise field was given by his successor at NPL in a seminar at the Institute of Sound and Vibration Research in November 1997 (Berry 1998). The Director of that Institute rounded off the discussion by saying that the lecture could have been sub-titled "The Douglas Robinson Story."

Notwithstanding his dedication to numbers, to data, Douglas had many other interests - his family, his garden and his music.

In his Eulogy at the funeral service, Southampton's Dean of Engineering said: "There were three main driving forces in Douglas' life, his work in acoustics, his love of music and his family. But most important of all was Joyce, who provided the framework within which he was able to work so tenaciously. Without her support the distinctions which have become the hallmark of his professional life may not have been so profound."

Douglas was an accomplished pianist. It was therefore fitting, when we all bade farewell on that summer day in Southampton that it was to the music of Johann Sebastian Bach, Johannes Brahms and Richard Strauss.

Symbolically, the service opened with Bach (Prelude and fugue) - "one of the greatest and most productive geniuses in the history of Western music" and "the supreme composer of fugues". Passed on to Strauss, "a master of composing for the human voice", with his rendering of Hesse's "Beim Schlafengehen", and closed, appropriately, with Brahms' Late Piano Music (Opus 119 and 79). One wonders whether Douglas saw Brahms' music in mathematical terms as did Vernon Blackburn when he wrote (in the Pall Mall Gazette of 28th of February, 1900): "You feel an undercurrent of surds (a quantity not capable of being expressed in rational numbers) of quadratic equations, of hyperbolic curves, of the dynamics of a particle�"

One suspects that Douglas, like Hermann Hesse, envisaged a philosophical Utopia under the control of a quasi-monastic elite. One could well see him endorsing Hesse's sentiments which were expressed in a private letter over forty years ago:

"Everywhere on earth there are people of our kind. That for a small part of them, I can be a focal point, the nodal point in the net, is the burden and joy of my life."

One could not fail to be impressed with the serenity and simplicity of that service in Southampton which was, at the same time, efficient yet exhibiting aesthetic values and intellectually satisfying. I find it difficult not to believe that it was Douglas' last throw - a draft standard prepared at the behest of our Masters in Brussels to govern such events. The inclusion of Bach, Brahms, Hesse and Strauss, together with Dylan Thomas poetry ("And death shall have no dominion"), will ensure the draft's easy and rapid passage through the various CEN committee stages as it goes on to receive its EN stamp.

All this will do much to commend Douglas to St Peter, who, having been described by at least one hagiologist as having an "impulsive nature", would see in DWR a kindred spirit.

Once inside the Pearly Gates, Douglas will no doubt make his way to the equivalent of the "Roter Igel", Brahms' favourite watering place in Vienna, to enquire of the whereabouts of JSB. We would then anticipate a sequel to Eli Maior's account (in his "e: The Story of a Number") of an apocryphal meeting between Bach and Johann Bernoulli:

JSB: Herr Professor, I am very glad to meet you at last, having heard so much about your remarkable work.

DWR: I am equally delighted to meet you, Herr Kapellmeister. Your fame as an organist and composer has expanded immeasurably since your lifetime. But tell me, are you really interested in what I have to say?

JSB: Yes. But first tell me, is this logarithmic spiral that Johann told me about so basic to hearing, to music?

DWR: Oh, Yes. We see it even in the structure of the ear. That part of the ear which deals with the detection and analysis of musical and other sounds, the cochlea, is itself a logarithmic spiral.

JSB : But what is it in the formula of the logarithmic spiral that makes it so special?

DWR: It is this magic number e, which is so pervasive in the world.

JSB: So did you use it in your formulae?

DWR: Yes, look at this which I have just scribbled on some paper

H' = 27.5 [1 + tanh{(EA + k - k+ u)/15}] + u + c (N - 20) 2

JSB : But I cannot see any e there.

DWR: No, it is hidden in the "tanh"

JSB: Ah, yes. Now, what can I do for you?

DWR: How about a little composition dedicated to the decibel?

We will all miss Douglas. We share the sadness of his family and friends. But he will forever be with us in his extensive published work.

Valete![36]

References