Noise Health Home 

[Download PDF]
Year : 2003  |  Volume : 6  |  Issue : 21  |  Page : 51--62

Signal perception during performance of an activity under the influence of noise

CA Sust1, H Lazarus2,  
1 ABoVe GmbH, Wettenberg, Germany
2 Bundesanstalt für Arbeitsschutz und Arbeitsmedizin, Dortmund, Germany

Correspondence Address:
C A Sust
ABoVe GmbH Dresdener Str. 11, D-35435 Wettenberg


Usually the perception of acoustic signals is investigated under conditions where the subjects pay full attention to the signals. It can be assumed that the threshold of signal perception is much higher if the attention has simultaneously to be focused on the execution of any kind of other activity. In the following experiment subjects have to perceive acoustic signals while solving different arithmetical tasks at the same time. The results (number of perceived signals, number of arithmetical tasks solved, reaction time, and solving time) show that the threshold of signal perception rises while other tasks are being performed simultaneously. Consequences for the recognition of warning signals in occupational safety and in traffic conditions are discussed.

How to cite this article:
Sust C A, Lazarus H. Signal perception during performance of an activity under the influence of noise.Noise Health 2003;6:51-62

How to cite this URL:
Sust C A, Lazarus H. Signal perception during performance of an activity under the influence of noise. Noise Health [serial online] 2003 [cited 2020 Sep 24 ];6:51-62
Available from:

Full Text

Introduction to the problem

In many situations of everyday life, in particular at the workplace and in road traffic, there is often a need to pay attention to non-verbal, acoustic signals because they contain important information on general environmental orientation. For example, they can symbolise hazardous situations or indicate special operating conditions in machines and motor vehicles to which attention has to be paid and which require a reaction. Their function is then to draw attention from the performance of an activity or task to certain information (orientation reaction). Frequently, however, the perception of signals is hindered by unfavourable conditions. In addition to social (pressure of time) and person-related aspects (individual constitution), such conditions are mainly environmental ones, such as noises which can mask acoustic signals. That is why, with special regard to the often noisy situation at the workplace and in road traffic, the perception of signals is regulated by relevant regulations (ISO 7731) to ensure their function as "attention drawer". Such regulations only give rules for adequate signal design and require an adequate signal-to-noise ratio, however, which only ensures perception and recognition of the signals if full attention is directed at them. This does not take into account that the perception of signals is not only dependent on the acoustic conditions. For most working individuals, the primary purpose of their presence at the workplace, is the performance of their respective working tasks. Although there is a general need to pay attention to signals, it is in particular the performance of the task to which concentration is directed, because it is absolutely necessary for success to focus attention on the various steps of the action. Attention as a "state of directed alertness" implies precisely that "recognition objects are, as a matter of priority, perceived, highlighted, fixed and more clearly registered" (Clauss, 1986, 57). The narrowing of attention is then of functional use for dealing with the activity (see also Neumann, 1985, Chap. 6).

Signal perception performance during performance of arithmetical tasks

It is known, that the threshold of acoustic signals is influenced by a number of acoustical and nonacoustical parameters. Such parameters are:

the acoustic features of the signal,the background noise (level, spectrum),environment (reverberation time) andthe potential listener's condition (hearing loss, wearing of hearing protectors, knowledge about dangers in the situation, probability of the signal).

The influence of the parameters on signal perception and recognition is well known (Zwicker and Feldtkeller 1967, Coleman et al. 1980, Patterson et al. 1982, Wilkins and Martin 1982, Abel et al. 1985, Wilkins and Martin 1987, Wickens 1989, Edworthy 1994, Lazarus 1998, Malter and Guski 2001). But what happens if, during any kind of activity, signals have to be perceived and recognised.

It is assumed that the perception of the signals is essentially influenced by two aspects: on the one hand by the acoustic quality of the situation, in which the signal should be perceived, and on the other hand by the intensity of the attention directed to this perception task. In other words the thresholds for perception of the signal need to be higher when attention is bound to the performance of an activity. Therefore danger signals should have the quality for interrupting the focusing process in their respective acoustic situation. In everyday-life signals might be missed because the thresholds are not high enough for identification although the signal alone would be perceived, if a person would not be involved in his or her usual task. Obviously the signal needs to be much clearer - that is, needs a higher signal-to-noise ratio - in order to be perceived. One can assume that the perception of the signal worsens, the more somebody is involved in another activity, just like performing arithmetical tasks. That means the time-relationship between signal perception and task solving determines the attention which can be spared for the signal perception: The more a person is involved in solving the task - which absorbs more attention - the less probable is the perception of a signal which occurs shortly before or just during this critical phase. This is especially true, if in addition the signal is difficult to recognise because of a medium signal-to- noise ratio.

The present experiments were used to investigate the perception of signals in a noisy environment, while being involved in an activity, in this case performing arithmetical tasks. The acoustic quality here was given by a simple signal more or less disturbed by noise. The degree of difficulty was controlled via the signal-to-noise ratio (L SN ). The direction and intensity of attention was controlled via the time-related relationship between the signal and the task (DT).

It can therefore be assumed that the quality of signal perception is reduced, especially if the signal is difficult to recognise and/or the work on the task is in a critical phase.


Acoustic signals and noise

In this test, the subjects have to react to acoustic, randomly presented signals of varying intensity by pushing a button. In a number of test series they have also to solve arithmetical tasks.

The tests are conducted in an interfering pink noise at a level of L NA = 91 dB. The sound pressure level of the third octave band (L Nt ) at 1 kHz is L Nt = 77 dB.

The signals are sine tones of 1 kHz. The sound levels of the signals (L S ) are L S = 66 to 94 dB. The signals are presented in level steps of 4 dB, so that 8 signal-to-noise ratios are available. The signal-to-noise ratios are expressed in third­octave sound pressure levels as L S - L Nt = L SNt = 17, 13, 9, 5, 1, -3, -7, -11 dB. The signals have been recorded on a tape for a duration of approx. 8 s and are presented to the test subjects via a tone control for a period of 1.5 s. The interval between the successive signals is, in the experiment, between 10 and 16 s (signal-task phase).

Signals and noises are presented to the subjects through a loudspeaker in a anechoic chamber.

Arithmetical tasks

The experiments are conducted with two kinds of arithmetical tasks, single arithmetical tasks (test I; single AT) and with triple tasks (test II; triple AT).

The single AT (test I) is a multiplication task "a * b = ?", where 3 a , b 9 is. The multipliers are selected from a random sequence, the [Figure 3],[Figure 4],[Figure 5],[Figure 6],[Figure 7],[Figure 8],[Figure 9] being approximately evenly distributed. The tasks are presented to the subject for 2 s on the screen. After that the subject has 4 s to solve the task and to enter the solution on a keyboard (from 0, 1, 2 ... 9) by pressing the relevant keys. The triple AT (test II) includes four different types (ATT):

1) a * b = e

e - f = ?

Operations: * / - / -

c - d = f

2) a * b = e

e - f = ?

Operations: * / + / -

c + d = f

3) a * b = e

e - f = ?

Operations: * / * / -

c * d = f

4) a * b = e

e + f = ?

Operations: * / * /+

c * d = f

where 2 d, e > f.

With these triple tasks the test subject has to note the interim results e and f and process them further with the third operation. Only this last result from the third operation has to be indicated. The arithmetical task is displayed on the screen for 4 s and the test subject has 7.5 s to solve the task and enter the result on the keyboard.

Experimental design

The subject had to solve the arithmetical tasks. If a signal occurred, the subject had to react as fast as possible to the signal by pressing a button. No priority instruction was given. Both tasks - signal perception and arithmetic - should be done as fast and carefully as possible. The perception of a signal was accepted as correct if a button was pressed within 2.5 s of the beginning of the signal. The possibility of reacting to a signal was limited so as to largely avoid possible, purely random answers. The criteria for the quality of perception were the percentage of signals perceived (signal responses, SR) and the reaction time for this perception (signal reaction times, RT).

The criteria for the quality of the calculation performance were the number of tasks correctly solved (correct answers: rAA) and the corresponding solution times (solution times: ST). The solution times were the times measured from the beginning of the tasks appearance to the input of the solution (input of first figure) see [Figure 1]. In the triple AT only the final result counted, not the interim solutions.

The microprocessor registered the responses to the signal (SR), the correct and incorrect solutions to the arithmetical tasks (rAA) and the reaction and solution times (RT, ST).

The test lasted approximately 1.5 hours and was broken down into four sectors each lasting 15 minutes. Before the first test series, the subject worked through a brief practice phase of 6 minutes with 20 signals and tasks. Between every sector, the test subject had a break of 5 minutes.

A test series consisted of 80 signal-task phases. During a signal-task phase, which lasted about 10 to 16 s, only one arithmetical task and/or one signal was presented to the test subject in every case including a short variable break. In eight signal-task phases, the arithmetical task was presented without a signal. In further eight phases the signal was presented without arithmetical task (one for each of the eight signal levels). In 64 signal-task phases one signal and one task appeared. The presentation of the signals was via a tone control and the display of the respective arithmetical tasks were controlled with a microprocessor in such a way that, in relation to the beginning of the acoustic signal, a part of the tasks could appear on the screen before, during and also after the presentation of the signal see [Figure 1]. The time between the onset of the signal (S) and the beginning of the arithmetical task was defined as time differential (DT AS )

The appearance of the arithmetical task is:

. before beginning of signal:

DT AS = -0.7 / -1.4 s

. during signal:

DT AS = 0 / 0.7 / 1.4 s

. after end of signal:

DT AS = 2.1 / 2.8 / 3.5 s.

This is repeated for all eight signal levels, so that, in each of the 64 signal-task phases, every combination of the task time-point (8) and signal level (8) occurs only once. The 80 different signal-task phases are distributed at random within a 15-minute test series and their order is different in each test series.

Test I was conducted with 20 male and female students aged between 20 and 30, and test II was conducted with 8 students of the same age. All test subjects were paid. A hearing test was conducted and only subjects with normal hearing (hearing loss lower than 25 dB at frequencies at 0.5/1/2/3/4 kHz) participated.

Presentation of results

The investigation in each case was to establish the influence of the main variable - the signal-to­-noise ratio (L SN ) and the time differential (DT AS ) - on the signal perception performance (SR, RT). Additionally the influence of the arithmetical calculation performance (rAA, ST) was examined.

The data were checked statistically by means of a 2-factor variance analysis separately for test1 and test II [Table 1].

For the single and triple task (test I, II) the influence of the independent variables (L SN ,DTAS) on the perception performance are significant (SR, RT; [Table 1]), as well as the effect of interaction for SR.

Since the test with triple tasks (test II) was only conducted with eight test subjects, the results for this test can only be discussed in terms of trends. In spite of a small number of test subjects, there was in most test results a significant influence [Table 1] from the independent variables signal-­to-noise ratio (L SN ) and time differential (DTAS).

The effects of the independent variables on the arithmetical calculation performance (rAA, ST) for both task types (test I, II) were only partly significant [Table 1]. The influence of the four task types (Test II, ATT) was calculated in a single­factor variance analysis: The signal perception performance is not significant, but the arithmetical calculation performance is. False alarms - pressing a button with no signal present - were smaller than 1 %.

At first the results of test I are described. The signal-to-noise ratio influences the percentage of signals detected and the signal reaction time (SR, RT) as anticipated [Figure 2]a, [Figure 3]a. Signal perception is low with low signal-to-noise ratios; it is therefore only possible to determine signal reaction times (RT) for the highest five signal-to­noise ratios (L SN > 1 dB, [Table 1]). The percentage of signals heard (SR) show the known dependance of signal-to-noise ratio [Figure 2]a

The masked threshold (L SN for which 50 % of the signals are heard) is in relation to the third-­octave sound pressure level (L Nt ) of the noise in accordance with [Figure 2]a L SNt = L St (50 %) - LNt = -5dB. This figure corresponds approximately with that given by Zwicker and Feldtkeller (1967).

The signal reaction time decreases with increasing signal-to-noise ratio [Figure 3]a.

Perception of the signals and the signal reaction time are also influenced by the time at which the signal occurs in relation to the arithmetical task [Table 1], [Figure 2]b, [Figure 3]b. If the arithmetical tasks are presented before or near the onset of the signal (DT AS 1 s), the difference between the beginning of the tasks and the beginning of the signal (DTAS) has no influence on the signal responses and the reaction times (see [Figure 2]b, [Figure 3]b). The values do not differ significantly from those with no arithmetical task (k.A.) and are influenced only by the signal-to-noise ratio (L SNt , [Figure 2]a, [Figure 3]a).

In addition, the combination of both factors - signal-to-noise ratio and signal-task relation - has an effect on the perception of the signal (s. [Table 1]). As expected the perception of the signals is impaired most strongly by an additional task when the percentage of signals is relatively high (> 95% for DT AS >1 s). With very high (SR » 100 %) and low signal-to-noise ratios (SR AS = -1.5 to 0 s; [Table 1]). The signal reaction time increases by about 40 %: at high signal-to-noise ratios (L SNt = 1 to 17 dB) without an arithmetical task the RT is 0.8 to 1 s, and if the arithmetical task appears before the signal, the RT is 1.1 to 1.3 s [Figure 3]b.

Further the results of test II are described. The results for signal perception [Figure 4],[Figure 5] correspond approximately to those from the test with the single arithmetical tasks.

The influence of the signal-task time relation (DT AS ) on signal perception (SR, RT) obtained for single arithmetical tasks (test I) is also evident in triple tasks (test II) as is shown in [Figure 4],[Figure 5]. Since the solution times see [Table 2] are longer because of the degree of difficulty of the tasks, the range of maximum changes in perception performance (reduction in the percentage of signals perceived) is here DT AS = -1.4 to 0.7 s [Figure 4]; with the single tasks it was -0.7 to 0 s [Figure 2]b.

The average of the arithmetical calculation performance - percentage of correct solutions and solving times - are summarized in [Table 2]. The percentage of correctly solved tasks is about 95 % in the case of the single tasks, and with the triple tasks it is 55 to 90 %, depending on the type. The solution time for single tasks is 1.5 to 2 s and for triple tasks 2.5 to 3.5 s. Although the percentage of tasks solved and the solution time only changes slightly as a function of the two independent variables (DT AS , L SNt ), it is possible to highlight some changes.

If the task appears before the signal, the percentage of tasks solved is lower (rAA = 91.7 to 94.3 % with DT AS 0 s) than when the task is presented after the signal (rAA = 95.0 to 96.0 % with DT AS > 0.7 s). The solution time - when the arithmetical task appears before the signal - is about 10 % longer than in other cases: ST = 1.9 s with DT AS 0 s as against ST = 1.8 s with DT AS > 0.7 s.

The solution of the triple tasks (test II) is only slightly influenced by the presence of the signal. The percentage of correctly solved tasks and the solution time differs for the different types of tasks [Table 2]. By far the most difficult task is that in which the three different types of calculation are used (*/+/-), and the easiest is that with only two different types of calculation, where multiplication only occurs once (*/-/-).

Discussion of results and summary

The main result is as expected that the perception of the acoustic signal is influenced by a simultaneous activity, arithmetical tasks in this case. Results are first discussed for the tasks separately. Signal perception as well as handling of arithmetical tasks both are influenced by the respective degree of difficulty.

Eight steps of the signal-to-noise ratio (L SNt = - 11 to 17 dB) give the degree of difficulty of signal perception. The reaction time to the signals thus increases from the high to the lower signal-to-noise ratio from RT = 0.7 to 1.5 s [Figure 3],[Figure 5]. The results of the perception task (SR) were, according to the degree of difficulty, i.e. signal-to-noise ratio, SR = 0 to 100 % [Figure 2],[Figure 4].

The degree of difficulty of the arithmetical task was only changed in that on the one hand single tasks (test I) had to be dealt with (a * b, 3 a, b 9) and, on the other, triple tasks. These triple tasks (test II) had four different difficulty stages. The solution times were, according to degree of difficulty of the task, ST = 1.7 to 3.5 s. The percentage of correct solutions is according to the degree of difficulty of the tasks, rAA = 55 to 95 % [Table 2].

The results confirm the known relations. But the main results show reciprocal influence, if signal perception occurs while the arithmetical tasks are being dealt with.

The percentage of signals perceived (SR) depends very much on the level of the signal-to­noise ratio and the call to solve a single arithmetical task (test I, [Figure 2]a). With a signal perception of medium difficulty (L SNt = 1 to 9 dB, SR = 90 - 99 % without arithmetical task, [Figure 2]a) the percentage of signals heard is reduced by up to SR = 15 to 20 %, if an arithmetical task has to be solved at about the same time (DT AS SNt SNt = 9 to 17 dB, SR » 100 % without arithmetical task) the influence decreases only very gradually. Even with a signal-to-noise ratio of more than L SNt = 15 dB, the number of signals heard is still reduced by up to 10 % by the arithmetical task [Figure 2].

The signal perception (SR) shows with the triple task (test II) similar results although this task is more difficult than the single task (test I; [Figure 2]b, [Figure 4]).

In the perception of signals, an further effect is clearly evident: the reaction time increases by approximately RT = 0.3 s, if an arithmetical task (test I, II) has been indicated shortly beforehand (DT AS AS SNt > 1 dB) by approximately 10 %, if DT AS 0.5 s) through the call to solve arithmetical tasks to 0.4 s (DT AS SNt > 0 dB otherwise it is 0.4 s at L SNt AS ) is from -1.5 to 1 s, that means about DT AS Consequences

From the results shown it is possible to draw the following conclusion. Clearly even this relatively simple signal perception task can be sensitive to interference from an accompanying activity. This applies essentially regardless of how clearly or reliably the signal can be heard or recognised. The probability of perceiving a signal is lowered by being involved in an activity, in particular in those cases where the signal is not that easily perceived, that means at low or medium signal-to-noise ratios.

The practical consequences for the recognition of acoustic hazards and warning signals in traffic and in factories are obvious. Such signals have usually to be perceived during the performance of another activity, frequently with a high degree of concentration or attention directed to this activity. At the same time the noise levels in traffic and factories are frequently so high that the signals are only just audible.

There is some literature (see above) which deals with the problem how to guarantee a reliable, safe and fast perception in noisy condition (Wilkins and Martin 1987). The question is, how to design the signals that meets this high quality in each situation.

The main feature to ensure a safe and fast perception is the signal-to-noise ratio, which is discussed in the literature to range from L SNA = 5 to 25 dB. The impairing influence of other activities on signal perception has to date been disregarded. The discussion in the literature and experiences of experts have lead to the result that in ISO 7731 the A-weighted signal-to-noise-ratio should not be lower than L SNA = 15 dB or the masked threshold in octave bands should not be lower than L SNoct =10 dB. It must be noted that the perception of signals is frequently made difficult because of a low signal-to-noise ratio. This means that the level of the signal-to-noise ratio in factories and traffic tends in many cases to be L SNA = 0 dB rather than L SNA > 15 dB. But it is precisely in this range that an activity being performed has a very negative effect on the result of signal perception. Missed danger signals and accidents sometimes caused by this may be due to the concentrated performance of other activities.

Since the type of activity cannot usually be influenced, the signals and their sound pressure level (compared to the noise level) must be designed in such a way that they can be reliably and quickly recognised in spite of any activity, especially with regard to warning signals. So the above mentioned required quality of ISO 7731 must be met.[13]


1ISO 7731: Ergonomics - Danger signals for public and work areas - Auditory danger signals
2Abel, S.M., Kunov, H., Pichora-Fuller, M.K. and Alberti, P.W. (1985). Signal Detection in Industrial Noise: Effects of Noise Exposure History, Hearing Loss, and the Use of Ear Protection. Scand Audiol 14, 161-173
3Clauss, G.; Kulka, H.; Rosler, H.-D.; Timpe, K.-P.; Vorweg, G. (ed.) (1986). Worterbuch der Psychologie. Koln: Pahl-Rugenstein
4Coleman, G. J., Leamon, T. B., and Drayton, I. D. R. (1980). Auditory Communication in the Mining Industry - Final Report on CEC Contract 6245-11/8/019. Edinburgh: Institute of Occupational Medicine, 1-116
5Edworthy,J. (1994). Urgency mapping in auditory warning signals. In: N. Stanton (ed.). Human Factors in Alarm Design. London: Taylor and Francis. 15-30
6Lazarus, H. (1998). Noise and Communication: The Present State. Noise Effects '98, Congress Proceedings 7th International Congress on Noise as a Public Health Problem, Sydney, 157-162
7Malter, B.; Guski, R. (2001). Gestaltung von Gefahrensignalen. Forschungsbericht der Bundesanstalt far Arbeitsschutz und Arbeitsmedizin, Dortmund FB 935. Bremerhaven: Wirtschaftsverlag NW
8Neumann, O. (1985).Perspektiven der Kognitionspsychologie. Berlin: Springer
9Patterson, R. D., Nimmo-Smith, I., Weber, D. L., and Milroy, R. (1982). The deterioration of hearing with age: Frequency selectivity, the critical ratio, the audiogram and speech threshold. J.A.S.A. 72, (6), 1788-1803
10Wickens, C.D. (1989). Engineering Psychology and Human Performance. Columbus Toronto: Charles E Merrill Publ.
11Wilkins, P., and Martin, A. M. (1987). Hearing Protection and Warning Sounds in Industry - A Review. Applied Acoustics 21, 267-293
12Wilkins, P. A., and Martin, A. M. (1982). The Effects of Hearing Protection on the Perception of Warning Sounds. In: Alberti, P.W..Personal Hearing Protection in Industry. 339-369
13Zwicker, E.; Feldtkeller, R. (1967). Das Ohr als Nachrichtenempfanger. Stuttgart:Hirzel 1967