Content uploaded by David Benyon
Author content
All content in this area was uploaded by David Benyon on Dec 18, 2013
Content may be subject to copyright.
Proceedings of the 12th International Conference on Auditory Display, London, UK, June 20-23, 2006
WORKPLACE SOUNDSCAPE MAPPING: A TRIAL OF MACAULAY
AND CRERAR’S METHOD
Iain McGregor, Alison Crerar, David Benyon and Gregory Leplatre
Napier University, School of Computing
10 Colinton Road, Edinburgh, EH10 5DT
{i.mcgregor, a.crerar, d.benyon, g.leplatre}@napier.ac.uk
ABSTRACT
This paper describes a trial of Macaulay and Crerar’s method of
mapping a workplace soundscape [1] to assess its fitness as a
basis for an extended soundscape mapping method. Twelve
participants took part within 14 separate environments, which
included academic, commercial and domestic locations. Results
were visualized and subsequently collapsed to produce typical
responses to typical environments, as well as specialist
responses to a shared workplace.
1. BACKGROUND
Macaulay and Crerar’s method [1] was method was chosen for
this study as it addresses the mapping of auditory environments
from a human computer interaction perspective, rather than the
more commonplace acoustic ecological perspective. The
original authors identified a ‘gap in the research agenda of the
auditory display community’ and attempted to ‘utilize
ethnographic techniques’ rather than the traditional ‘cognitive
science model’ in order to fill this ‘gap’. The method takes the
form of a ‘context of use’ through ‘activity’ in the form of an
‘analytical tool’ where each sound event is classified according
to its sound type, information category and acoustical
information providing a form of metadata (Table 1).
It also goes further than a traditional Gestalt figure ground
approach by having a third contextual dimension which can be
utilized as level of listening. This third layer provides an
insight into auditory monitoring, that is a sound event which
provides information about what is happening around a listener
without actively eng aging with it, such as is required when
monitoring the output of a computer printer while actively
listening to music.
Sound Type Example
Music Any type of identifiable music, radio/stereo
Abstract Unusual sounds not normally experienced, video recorder chewing a tape
Speech Conversation
Everyday Identifiable recognised sounds.
Information Category Example
Visible entities and events The phone ringing
Hidden entities and events The photocopier round the corner being used
Imagined entities and
events
Something big is happening on the political desk (it has gone quiet).
Patterns of events/entities Someone is batch copying a large document
The passing of time It’s nearly deadline time (because the shift change is happening)
Emotions The sports desk sub-editor is unhappy (tapping)
Position in Euclidean/
acoustic space of entities/
events and of the listener
The editor is at the foreign desk behind me (can hear his voice)
Acoustical Information Example
Foreground Computer beep to attract your attention.
Contextual Door opening (Help you orient to the nature of your environment.)
Background
Whine of disk drive providing reassurance or information about the state of the
world.
Table 1: Macaulay and Crerar’s Workplace Soundscape
Mapping Tool Questionnaire
This tool was originally designed to be utilized by
fieldworkers and designers, in order to preview the ‘workplace
context’ creating a ‘rich picture’ prior to the introduction or
development of an auditory interface or system. It was
developed during a yearlong ethnographic study at The
Scotsman offices in Edinburgh, and is based on the work of
Ferrington [2] for the acoustical information as well as T ruax
[3] and Chion [4] for the sound types and information
categories. The resultant map could then be used ‘to add
auditory aspects to ethnographic vignettes, as well as providing
a ‘shared language’ which would facilitate ‘comparative
studies’.
One of the key elements not addressed by Macaulay and
Crerar was th e end user or inhabitant. Each individual
experiences a unique soundscape, based on their previous
experiences and interests, and as such will provide unique
responses, what can also be termed as a Rashomon phenomenon
[5]. Maps created by multiple inhabitants can provide a further
insight into the typical versus the individual experience. The
designer’s perspective can be compared to that of individuals,
or a typical response for a specific environment, or a typical
response to a typical room. This would allow an
anthropocentric approach to the design of auditory systems
suitable for shared auditory environments. An evaluation of
their method was required, as no examples had been included in
the original paper, or subsequently published.
2. METHOD
This preliminary study took the form of fourteen different maps,
using thirteen participants in twelve individual locations, these
were divided between regular, intermittent and new inhabitants
(Table 2).
Participant Location Environment Type of inhabitant Duration No. Events Notated by
SL Library University Intermittent 15 mins 8 Participant
GC Computer Room 1 University Intermittent 15 mins 8 Participant
JN Computer Room 2 University Intermittent 15 mins 7 Participant
GA Computer Room 2 University Intermittent 15 mins 11 Participant
MB Staff Canteen University New 40 mins 16 Participant
EM Staff Common Room University New 30 mins 21 Participant
MP Computer Room 3 University Intermittent 15 mins 17 Participant
KM Computer Room 3 University Intermittent 15 mins 8 Participant
CR B/W Darkroom Photographic Lab Regular 55 mins 25 Observer
RW Colour Printing Photographic Lab Regular 85 mins 35 Observer
FD Reception Photographic Lab Regular 20 mins 20 Observer
DK 50's style Diner Diner Regular 180 mins 59 Observer
IM Kitchen Domestic Regular 180 mins 53 Participant
IM Study Domestic Regular 60 mins 46 Participant
Table 2: Subjects and locations for trials of Macaulay and
Crerar’s method
Regular denoted that the participant regularly spent many
hours in that environment, either due to it being their workspace
or in the case of the domestic environments it formed part of
their home and their familiarity was therefore high. Intermittent
denoted that participants had visited the space periodically, but
it did not constitute their main work or home area. New
described a participant who was new to the environment, having
never entered it before. This method of describing was not part
of Macaulay and Crerar’s method, it has been added in order to
notate a participant’s familiarity with an environment.
The first iteration involved eight individual participants in
one of six different University environments. Each participant
was asked to list in written form all the sounds they could hear
in their local environment within a minimum of a 15 minute
ICAD06 - 285
Proceedings of the 12th International Conference on Auditory Display, London, UK, June 20-23, 2006
period. All of the eight participants spontaneously closed their
eyes and stopped what they were doing in order to consider
their responses. They subsequently opened their eyes in order
to confirm what they had heard in each instance, before re-
closing them in order to continue. The majority of the
participants missed at least a few of the sound sources or events.
One major consequence was that participants stopped creating
any noises themselves in order to listen more carefully, thereby
omitting a major contribution to their personal soundscapes.
Each sound event was later classified according to Macaulay
and Crerar’s method and then visualized as detailed in section
2.1.
To compensate for the problem of participants dimishing
their usual auditory environment, by stopping what they were
naturally doing and closing their eyes, a further four participants
were observed while working. Notes were made of all the sound
sources/events by a trained observer (sound engineer), with
classification taking place after the prescribed period. During
questioning, which took place in the same environment, if the
participants w ere not aware of any of the sounds it was notated
as ‘not aware’ and omitted from the subsequent map. Subjects
were also asked if there were any sound events which they
wished to add due to possible omission by the observer, no
additions were suggested.
The final two studies, were conducted in a home
environment, with the participant making a record of each n ew
sound that was heard, ov er a specified period of time, as normal
work was conducted. This enabled a longer time period to be
studied without the need for an observer, which might
potentially alter their behaviour or working practices, although
this constant monitoring of the sounds was still an artificial
condition.
2.1. Visualization
In figure 1 each concentric circle represents the acoustical
information with foreground being located in the centre. The
seven segments on the left represent the information category,
as labeled.
Figure 1. Pictorial Representation of data, based on
an original map by Macaulay and Crerar (unpublished)
The sound type, was notated by the labeling of each
‘bubble’ with a symbol. Music was a couple of notes ,
abstract a series of numbers 123, speech a series of letters abc
and everyday by an everyday item in this case an apple which
appears to have had a bite taken out of it . Sound events were
cross-referenced to letters within each ‘bubble’ to help prevent
the image becoming too cluttered by confining the contents to a
letter and a symbol, rather than a textual description of the
source and event. The visualization did not use colour for
individual maps, this was confined to maps with aggregated
responses, allowing easy differentiation between the two
different types. The individual colours in the latter case either
represented different participants’ responses, or the quantity of
responses for each sound event. Each source or event was
placed into the diagram according to the participant’s responses.
2.2. Environments
Locations studied included a library, computing rooms, a
commercial photographic processing laboratory as well as a
50’s style diner, domestic kitchen and study. The intention of
this diversity was to establish if the specified categories were
appropriate for a variety of auditory environments, and to
establish if any additional categories or layers would come to
light through the questioning.
2.2.1. Academic Environments
The first type of academic environment chosen was one of
Napier University’s libraries. This environment was a mix of
traditional library, and computer workstations. In order to test
the method to create a typical response to a typical room, three
separate 50 seat computing rooms, which were occupied for
practical work, were mapped by five individuals, each
contributing once. The last two academic environm ents studied
a the staff canteen and common room, these was chosen to
contrast with the previous work orientated environments, as
well as to elicit responses from new inhabitants, as both
participants had never previously visited either location before.
2.2.2. Photographic Processing Laboratory
A city centre photographic processing laboratory was chosen to
represent a shared workplace environment in which a wide
variety of tasks take place. The layout within the p rocessing
area was open plan, with small internal rooms leading off for
specialist tasks such as printing and film development. The
reception area adjoined the processing area through a
communicating door, and the main entrance to the lab was via
large glass doors, which opened onto a quiet lane.
2.2.3. 50’s Style Diner
A busy 50’s style diner located on a main road in Edinburgh’s
city centre was observed in order to represent a completely
different type of workplace environment. A central kitchen and
bar area was surrounded on three sides by customer seating for
up to 80 diners. Typically up to five people worked in this
environment, sharing all of the tasks from food preparation
through to washing dishes, serving drinks and calculating bills.
2.2.4. Domestic Environments
Two domestic environments were included to evaluate the
method’s suitability for notating domestic as well as workplace
environments. The first was a kitchen over a busy three hour
time period preparing a Christmas dinner. The kitchen was
located on the top floor at the rear of a five-storey building, with
few exterior sounds penetrating the double-glazing. The second
environment was a home office or study within the same
building, this contained many of the elements found in the
University computing rooms.
3. RESULTS
The sounds listed by the participants were classified during
interviews with the first author using the Macaulay and Crerar
model. All of the responses fitted easily into the categories
supplied, all of which were applied. Within the sound types,
music and speech were readily understood, and applied,
ICAD06 - 286
Proceedings of the 12th International Conference on Auditory Display, London, UK, June 20-23, 2006
although sometimes if the speech was a background sound then
if was classified as everyday. Abstract and everyday were not
consistently applied, there were three main interpretations
applied, the first was that of abstract representing a sound that
the participant was familiar with, but thought was unusual, or
differed from the norm for that particular object. With everyday
being applied to a sound which was closer to what they had
expected. The second was in terms of natural or artificial,
everyday representing natural, and abstract representing
artificial or man-made. The final interpretation was that of
other known for everyday and other unknown for abstract.
Within the information category visible was applied to
sound sources which could either be seen, or where a source
could be identified, even if it were not immediately visible.
Hidden was referenced for both sound sources which were not
visible as well as when the specific sound source was visible but
could not be isolated. Imagined was applied when an estimate
was being made as to the source rather than trying to interpret
the meaning of a sound, or lack of sound. Pattern denoted
either a series of connected sounds over a short period, or an
irregular long-term sound. The passing of time was applied
only once, and referred as a reminder that it was time to go
home. Emotions were routinely used when referring to speech
when it was a contextual or foreground sound event. It was also
applied to actions which informed others about someone’s
mood, when referring to impact sounds. Position was applied to
moving objects rather than marking where a stationary object
was located.
The application of acoustical information did not match
with the original aims of the Macaulay and Crerar paper [1], in
that this information was intended to illustrate the richness of
the information being gathered, a foreground sound provides
very little about what is going on in the world around you, such
as a ‘beep’ whereas contextual informs you about what is going
on contextually in your acoustic environment, and background
provides reassurance about what else is occurring in the
vicinity. The results more closely represent levels of listening
as suggested by Amphoux [6], where foreground sounds were
actively monitored and interpreted (sonic symbols), contextual
sounds told the participants about the place they were inhabiting
(sonic ambience) and background was applied to sounds that
were not paid attention to.
3.1. Typical Environments and Responses
Within the twelve environments studied, three were computer
labs which five respondents experienced. These were combined
to form a typical soundscape map of a typical computer lab
(Figure 2 and Table 3). A note was made of the number of
responses and then the descriptions were collapsed and
sequenced according to the amount of responses in descending
order. Colours were added in order to give a precise figure of
the number of responses per sound event, this started at the red
end of the spectrum for the highest value of six, and worked
through to yellow for single responses.
The combined results show a fairly even mix of acoustical
information, which would be expected in a shared environment.
There were similar amounts of visible and hidden sound events,
although all of the hidden were background, and most of the
visible were foreground. Emotions figured next highest, all of
which were associated with people either talking, whispering,
laughing, coughing or sighing. Only the laughter was
foreground, everything else was classified as either contextual
or background. Illustrating that the variety of forms of
vocalization did not intrude upon the shared acoustic
environment to any real extent. Within the sound type everyday
formed the largest group, which showed familiarity with the
environment. This was confirmed through only a couple of the
events being imagined (‘computer alert’ and ‘paper
movement’). All of the sounds created by the computers were
classified as abstract, compared to the sounds created by
inhabitants, which were either emotional or everyday. This
highlights the perception of the computers despite their
familiarity as being artificial, whereas the constant road traffic
was referred to as an everyday sound event.
Figure 2. Visualization of a Typical Response to a
Typical Computer lab.
AKeyboard Typing HComputer alert O
Purse velcro
B
Chair movement IMouse movement PLaughter
CComputer Fan J
Air conditioning QCoughing
DDoor opening & closing KPaper movement RSighing
EPeople talking LWhispering SFoot movement
F
Mouse clicking MComputer hard drive
G
Traffic NBag being zipped up
Table 3: Key for Figure 2
With the photographic processing lab it was possible to
create a map which represented the entire auditory environment,
rather than just the specific areas studied in isolation.
Responses from the individual regular inhabitants of the three
main areas were combined, only a single sound event was
shared, that of the telephone, which both FD and RW classified
as everyday, visible and foreground (Figure 3 and Table 4). This
approach reifies the soundscape from the perspective of an
individual who inhabits the space most regularly.
The map clearly showed the v arying control over the sound
events within this environment. CR had most control ov er his
environment, which resulted in 69% of the sounds being
foreground, whereas FD has least control w ith only 30% being
foreground, the inverse was true for background sounds with 7
and 35% respectively. Visible formed the largest group within
the information category at 62%, followed by patterns (18%)
and then hidden (12%). There were no instances of imagined as
all of the sound events were easily identified, nor were there
any references to time. Everyday formed the largest portion of
sound type at 74%, illustrating the familiarity due to repeated
exposure, compared to 19% abstract. There were surprisingly
few instances of speech (5%) as most of the work was
conducted in isolation, even, surprisingly, in the reception area.
ICAD06 - 287
Proceedings of the 12th International Conference on Auditory Display, London, UK, June 20-23, 2006
Figure 3. Visualization of Combined Responses to the
Photographic Processing lab.
A
Telephone ringing AA Paper door closing BA Slide Mount
B
Doors opening to lab/computer room AB RA4 paper processor Temperature beep BB Hairdryer (drying prints)
C
Coffee machine AC Closing paper insert lid BC Trays banging
D
Radio AD Paper bag rustling BD Radio 4
E
Customer’s mobile phone AE Footsteps on floor BE Paper towel
F
Customer AF Trapped air in paper box BF Mixing chemicals while measuring temperature
G
Chair noises AG Light switch BG Ilford 2150 RC Processor fan
H
Door alert AH Bremson enlarger focusing switch BH Processor warming up
I
Telephone ringing AI Bremson fan BI Processor ready for next print
J
Telephone hands free dialling AJ Keypad buttons BJ Brochure for timings
K
Fax ringing AK Keypad confirmation of settings BK Tapping bottom of processing tank
L
Modem dialling AL Locking lens into position BL Splash of fluid
M
Keyboard tapping AM Racking enlarger up and down BMThrowing empty canisters into metal bucket
N
Till beeping AN Aperture selection BN Handling plastic/paper bags
O
Cash drawer AO On/off switch (Bremson) BO Air canister
P
Switch receipt AP Inserting film carrier BP Easel adjustments and opening/closing
Q
Conversation AQ Enlarger easel BQ Running water cleaning film
R
Traffic AR Air buster BR Enlarger on/off switches
S
Fan heater AS Staff knocking to warn approach BSEnlarger fan
T
Printer AT Revolving door BT Timer confirmation
U
Keys AU Water pressure gauge (omnipro) BU Timer countdown (seconds)
V
Stereo AV Print finished beep (omnipro) BV Click of light switch
W
Enlarger controller buttons (Buick)
AW
Paper handling
BW
Printing paper box opening/closing
X
Exposure transport x 4 AX Door banging BX Water run off from tanks
Y
On/off switch (Buick) AY Nitrogen Generator BY Squeegee film
Z
Fan (Buick) AZ Trimming prints BZ Cleaning squeegee
CA
Knocking excess water off reels
Table 4: Key for Figure 3
4. CONCLUSIONS
Macaulay and Crerar’s mapping tool proved very easy to use,
with the combination of categories covering every perceived
sound event. Participants uniformly found it useful as a starting
point for analyzing their auditory environment, but all of the
individuals wished to contribute a greater amount of
information than the classification requested.
A number of omissions became evident through the study,
which could be split evenly between quantitative and
qualitative. The first was quantity, a room could have twenty
inhabitants, but speech was only detailed a single time. There
was no indication if only one person was talking or everyone
was talking. There was also no indication of how often these
conversations took place, whether they were continuous or
intermittent, as well as if they were concurrent with other
sounds or were isolated events. In addition, location was
omitted, in some cases the sound sources were equally spaced
around the inhabitant, whereas on other occasions they were
clustered. Directivity would have provided information about
whom the sound was intended for, with speech it is common to
direct the sound toward the intended recipient by facing them,
whereas ubiquitous sounds, such as computer fans, tend to be
omnidirectional. The last two quantifiable om issions were how
loud and how high or low in terms of pitch or frequency sounds
were. Whilst it would be extremely unrealistic to expect
responses in terms of decibels and hertz, simple terms like
loud/quiet and treble/bass might have provided consistent
responses.
This method clearly showed the relative percentages of
type, category and acoustical content, but was poor at
representing the original sound event. Recording the time of
each and every instance in real-time proved beyond the
capabilities of the researcher, outwith unusual events. In order
to produce an accurate record for data gathering, an auditory
recording was required of the location and a highly skilled
listener to decipher the recording; no software currently
available would be capable of automating this task. It was also
apparent that obvious sounds predominated, foreground and
contextual sounds w ere notated first, whereas background
sounds were notated last. This conforms with the way in which
individuals interpret the world around them, but it does allow
omissions due to perceptual masking, where a sound event is
being established for notation and a quieter less intrusive sound
is ignored, only to be notated if it is repeated after the
predominant sounds have been detailed. This problem can be
alleviated through recording the time period and notating the
complete set of events from the recording.
In qualitative terms, the classification of emotions did not
provide information about the mood which was being
expressed, whether it was anger, frustration or relief. There was
also no indication of what the type of interaction was, whether it
was produced by air passing through an object or an impact, this
was partially achieved through detailing the event, but not fully,
as the requirement is to represent the sounds which were
perceived rather than just a list of objects and actions. Bill
Gaver’s classification [7] of interacting materials by the
focusing on the simple sonic events could easily be added, as
the participants found it easy to recall the tonal qualities o f the
sound even an hour after the acoustic event. An understanding
of the sound’s information content would allow an insight as to
how the sound event was interpreted by the participant, such as
defined by Delage [8], wh ether it was an error alert or
confirmatory sound or even unwanted. Further detail about a
sound’s perceived aesthetics would also prove useful in terms of
communicating to designers the listener’s preferences such as
detailed in Gabrielsson and Sjorgen’s method [9].
5. REFERENCES
[1] Macaulay, C. and A. Crerar, 'Observing' the Workplace
Soundscape: Ethnography and Auditory Interface Design.
ICAD 98, International Conference on Auditory Display,
1998.
[2] Ferrington, G., Keep Your Ear-Lids Open. The Journal of
Visual Literacy, 1994. 14(2): p. 51 - 61.
[3] Truax, B., A coustic Communication. 2nd ed. 2001,
Norwood: Ablex Publishing Corporation.
[4] Chion, M., Audio-Vision: Sound on Screen, ed. C.
Gorbman. 1994, New York: Columbia University Press.
[5] Altman, R., The Material Heterogeneity of Recorded
Sound, in Sound Theory/Sound Practice, R. Altman,
Editor. 1992, Routledge: New York. p. 15-31.
[6] Amphoux, P., L'identite sonore des villes Europeennes.
1997, Cresson/IREC: Grenoble/Lausanne.
[7] Gaver, W.W., How Do We Hear in the World? Ecological
Psychology, 1993. 5(4): p. 285 -313.
[8] Delage, B., On sound design, in Sto ckholm, Hey Listen!,
H. Karlsson, Editor. 1998, The Royal Swedish Academy of
Music: Stockholm. p. 67-73.
[9] Gabrielsson, A. and H. Sjogren, Perceived sound quality of
sound-reproducing systems. Journal of the Acoustical
Society of America, 1979. 65(4): p. 1019-1033.
ICAD06 - 288
