The practice and potential of urban soundscape design Space -Kang.pdf · 2009. 9. 3. · Outline I. SOUNDSCAPE ANALYSIS & EVALUATION - typical results z A large scale survey across

The practice and potential of The practice and potential of urban soundscape designurban soundscape design

Jian KangJian KangSchool of Architecture, University of SheffieldSchool of Architecture, University of Sheffield

SOUNDER SPACES: the forgotten side of quality?London Zoo Wednesday 14 March 2007

SOUNDSCAPESOUNDSCAPEon relationships between ear, human being, sound environment andon relationships between ear, human being, sound environment and societysociety

relating to acoustics, aesthetics, anthropology, architecture, erelating to acoustics, aesthetics, anthropology, architecture, ecology, cology, human geography, landscape, linguistics, media arts, musicology,human geography, landscape, linguistics, media arts, musicology,noise control engineering, philosophy, psychology, political scinoise control engineering, philosophy, psychology, political science, ence, religious studies, sociology, and urban planningreligious studies, sociology, and urban planning

From noise control to soundscapes … …More people from various sectors consider soundscapes ….

SchaferSchafer

science, engineering, social science, humanity and art

OutlineOutline I. SOUNDSCAPE ANALYSIS & EVALUATION I. SOUNDSCAPE ANALYSIS & EVALUATION -- typical resultstypical results

A large scale survey across Europe and ChinaA large scale survey across Europe and ChinaStatistical analysisStatistical analysisSemantic differential analysisSemantic differential analysis

II. SOUNDSCAPE DESIGN II. SOUNDSCAPE DESIGN –– a systematic approacha systematic approachSystem: factors to be considered: soundscape descriptionfactors to be considered: soundscape descriptionSounds:Sounds: Activities; sound elements: water, musicActivities; sound elements: water, musicSpace: Sound field simulationSound field simulationPeople:People: Social/demographic factors, aSocial/demographic factors, aesthetics & pre-perceptionEnvironment:Environment: Aural/visual interactions, etc Aural/visual interactions, etc Tools:Tools: Design guidance, auralisation, ANN modelDesign guidance, auralisation, ANN model

AimAim

Soundscape in urban open public spacesSoundscape in urban open public spacesFocusFocus

Discuss/explore some key issues in the analysis, evaluation, andDiscuss/explore some key issues in the analysis, evaluation, anddesign potentials for soundscape design potentials for soundscape through our recent worksthrough our recent works

AcknowledgementsAcknowledgements Researchers involved in the Sheffield acoustics group, and project partners

This talk

Part IPart I

SOUNDSCAPE SOUNDSCAPE ANALYSIS AND EVALUATIONANALYSIS AND EVALUATION

Soundscape field survey in EU and ChinaSoundscape field survey in EU and China

14 urban open public spaces of five European countries, 5 14 urban open public spaces of five European countries, 5 in Chinain Chinaoverall physical comfort investigation, including acoustic, overall physical comfort investigation, including acoustic, thermal, lighting, visual, and social aspects thermal, lighting, visual, and social aspects each site 600each site 600--1000 interviews, >10,000 total1000 interviews, >10,000 totalSound level and reverberation measured (Leq10, 50, 90), Sound level and reverberation measured (Leq10, 50, 90), together with other physical indices together with other physical indices

EU

Sheffield China

Relationship between the sound level and subjective evaluation of sound level

Sheffield, UK

The Peace Gardens0.843

The Barkers Pool0.896

1

2

3

4

5

50 60 70 80Leq, dB(A)

Subj

ectiv

e E

valu

atio

n

73

Thessaloniki, Greece

Makedonomahon Square0.452

Kritis Square0.817

1

2

3

4

5

50 60 70 80Leq, dB(A)

Subj

ectiv

e E

valu

atio

n

In urban open public spaces, the

background level, say Leq90, is an

essential index. A lower background

sound level can make people feel quieter,

even when the foreground sounds

reach a rather high level

Statistical analysis of the survey results Statistical analysis of the survey results

Some typical results …

Subjective evaluation of sound level and Subjective evaluation of sound level and acoustic comfortacoustic comfort

1, Very quiet (very comfortable); 2, quiet (comfortable); 3, neither quiet (comfortable) nor noisy (uncomfortable); 4, noisy (uncomfortable); 5, very noisy (very uncomfortable).

R2 = 0.7721

R2 = 0.5413

1

2

3

4

5

50 55 60 65 70 75 80 85Leq, dBA

Soun

d le

vel e

valu

atio

n

1

2

3

4

5

Aco

ustic

com

fort

eva

luat

ion

010203040506070

1 2 3 4 5Evaluation

Perc

enta

ge (%

)

Peace Gardens, Sheffield

Sound Sound PreferencesPreferencesEssential preferencesEssential preferences

Positive attitudes towards natural and culturePositive attitudes towards natural and culture--related sounds.related sounds.MicroMicro-- preferencepreference

0

20

40

60

80

100

10~17 18~24 25~34 35~44 45~54 55~64 >65Age groups

Perc

enta

ge (%

)

Favourable

Annoying0

20

40

60

80

100

10~17 18~24 25~34 35~44 45~54 55~64 >65Age groups

Perc

enta

ge (%

)

Annoying

Favourable

Difference in sound preference between age groups

bird songs car music

MacroMacro--preferencepreferencecultural background and longcultural background and long--term environmental experience play an important role term environmental experience play an important role in peoplein people’’s judgment of sound preference.s judgment of sound preference.

A lim os Thessaloniki Sesto San G iovanni Sheffield K assel

K ara. Square Seashore M aked.

Square K ritis

Square P etazzi Square

IV N ov. Square

P eace G ardens

B arkers Pool

F lorent. Square

B ahns-. p latz

F 27.7 84 .0 74 .7 80 .3 74 .5 N 66.6 14 .8 20 .5 17 .9 22 .4 W ater

A 5 .7 1 .2 4 .8 1 .8 3 .1 F 37 .7 33 .2 34 .0 23 .5 N 43 .1 46 .1 59 .1 75 .3 Insects

A 19 .2 20 .7 6 .9 1 .2 F 31 .1 56 .8 47 .9 N 68 .9 35 .4 37 .6

B ells of church

A 0 .0 7 .8 14 .5 F 44 .2 48 .8 57 .3 88 .0 N 38 .3 28 .8 27 .2 12 .0 M usic played

on street A 17 .5 22 .4 15 .5 0 F 2 .3 7 .0 32 .2 23 .5 44.6 17 .9 18 .0 18 .5 15 .3 N 77 .8 77 .6 17 .0 69 .8 47.2 68 .3 69 .3 80 .5 84 .7 Surrounding

speech A 19 .9 15 .4 50 .8 6 .7 8 .2 13 .8 12 .7 1 .0 0 F 20 .3 25 .5 54 .1 29 .5 27 .4 11 .7 6 .9 1 .7 1 .0 N 54 .3 50 .8 19 .9 53 .0 53 .4 48 .4 40 .3 69 .0 54 .8 C hildren 's

shouting A 25 .4 23 .7 26 .0 17 .5 19 .2 39 .9 52 .8 29 .3 44 .2 F 5 .5 8 .0 8 .6 12 .9 7 .1 N 89 .9 84 .7 62 .0 58 .4 17 .9 Pedestrian

crossing A 4 .6 7 .3 29 .4 28 .7 75 .0 F 0 .6 3 .5 31 .3 2 .7 1 .6 2 .4 1 .0 N 26 .0 53 .0 16 .6 59 .8 35.4 38 .7 43 .6 Passenger

cars A 73 .4 43 .5 52 .1 37 .5 63.0 58 .9 55 .4 F 3 .4 1 .3 1 .6 3 .7 2 .1 N 52 .3 84 .3 39.2 38 .9 37 .9 Passenger

buses A 44 .3 14 .4 59.2 57 .4 60 .0 F 2 .9 1 .0 1 .4 2 .0 N 32 .2 35 .3 57 .9 54 .7 V ehicle

parking A 64 .9 63 .7 40 .7 43 .3 C onstruction F 2 .1 32 .5 2 .2 2 .1

N 52 .9 11 .5 18 .0 19 .2 A 45 .0 56 .0 79 .8 78 .7

Significant difference exists for some sounds among the cities, likely caused by cultural factors.

Classifications for various sounds in urban open public spaces (%).

People’s noise sensitivity may differ considerably in different citiesKassel has the quietest home environment, whereas Alimos has the noisiest. With similar Leq, the evaluation score in Kassel is much higher than in Alimos. • People from a noisy home environment adapt more to noisy urban spaces? • Cultural difference - people in Germany are more aware of urban noises?

in residential buildingsTaipei/Beijing people prefer

music from outside but Sheffield people not

A related study

Semantic differential analysisSemantic differential analysisThe technique, developed by Osgood et al in

order to identify emotional meaning of words, has been extended to a large variety of concepts

•• A soundscape walk in A soundscape walk in four typical open urban four typical open urban public spaces in public spaces in SheffieldSheffield

•• Comparison between Comparison between general public and general public and architectural studentsarchitectural students

•• Comparison between Comparison between Beijing and SheffieldBeijing and Sheffield

Very Fairly Little Neutral Little Fairly Very *Discomfort -3 -2 -1 0 1 2 3 Comfort *Noisy -3 -2 -1 0 1 2 3 Quiet *Unpleasant -3 -2 -1 0 1 2 3 Pleasant Ugly -3 -2 -1 0 1 2 3 Beautiful *Boring -3 -2 -1 0 1 2 3 Interesting *Artificial -3 -2 -1 0 1 2 3 Natural Unfriendly -3 -2 -1 0 1 2 3 Friendly *Dislike -3 -2 -1 0 1 2 3 Like *Harsh -3 -2 -1 0 1 2 3 Gentle Pure -3 -2 -1 0 1 2 3 Impure *Soft -3 -2 -1 0 1 2 3 Hard Sad -3 -2 -1 0 1 2 3 Happy *Slow -3 -2 -1 0 1 2 3 Fast Dark -3 -2 -1 0 1 2 3 Bright Low -3 -2 -1 0 1 2 3 High *Flat -3 -2 -1 0 1 2 3 Sharp Heavy -3 -2 -1 0 1 2 3 Light *Everywhere -3 -2 -1 0 1 2 3 Directional *Simple -3 -2 -1 0 1 2 3 Varied *Deadly -3 -2 -1 0 1 2 3 Echoed *Close -3 -2 -1 0 1 2 3 Far Unsafe -3 -2 -1 0 1 2 3 Safe *Unsocial -3 -2 -1 0 1 2 3 Social *Meaningless -3 -2 -1 0 1 2 3 Meaningful Weak -3 -2 -1 0 1 2 3 Strong Unsteady -3 -2 -1 0 1 2 3 Steady *Calming -3 -2 -1 0 1 2 3 Agitating *Smooth -3 -2 -1 0 1 2 3 Rough

Indices used in the evaluationconnotative and denotative meaningsconnotative and denotative meanings

Identify the main factors that characterize the soundscapeIdentify the main factors that characterize the soundscape

FACTORS INDICES 1 2 3 4 Comfort-discomfort .701 .164 .138 Quiet-noisy .774 Pleasant-unpleasant .784 .258 .157 Interesting-boring .435 .272 .274 .103 Natural-artificial .532 .102 .240 Like-dislike .519 .575 .247 .151 Gentle-harsh .502 .531 .123 Hard-soft .812 Fast-slow .827 Sharp-flat .220 .345 .488 Directional-everywhere .234 .441 .267 Varied-simple .115 .674 .167 Echoed-deadly .204 .531 Far-close .550 Social-unsocial .672 .462 Meaningful-meaningless .126 .585 .469 Calming-agitating -.143 .708 .286 Smooth-rough .683 .396

Factor analysis of the soundscape evaluation in Barkers Pool and the Peace Gardens in winter and spring. Kaiser-Meyer-Olkin measure of sampling adequacy, 0.798; cumulative%, 52.77; extraction method, principal component analysis; rotation method, varimax with Kaiser normalization; N=491.

Factor 1 (26%) Factor 1 (26%) –– relaxationrelaxationcomfortcomfort--discomfort, quietdiscomfort, quiet--noisy, pleasantnoisy, pleasant--unpleasant, naturalunpleasant, natural--artificial, likeartificial, like--dislike and gentledislike and gentle--harshharsh

Factor 2 (12%) Factor 2 (12%) –– communicationcommunicationsocialsocial--unsocial, meaningfulunsocial, meaningful--meaningless, calmingmeaningless, calming--agitating and roughagitating and rough--smoothsmoothFactor 3 (8%) Factor 3 (8%) –– spatialityspatialityvariedvaried--simple, echoedsimple, echoed--deadly and fardeadly and far--closecloseFactor 4 (7%) Factor 4 (7%) –– dynamicsdynamicsfastfast--slow and hardslow and hard--softsoft

The situation is more complicated than that for product soThe situation is more complicated than that for product sound quality, but still possible to identify major factors.und quality, but still possible to identify major factors.

Part IIPart II

SOUNDSCAPE DESIGNSOUNDSCAPE DESIGN

System

SoundsSounds

SpaceSpace

PeoplePeople

EnvironmentEnvironment

ToolsTools

Measurement/monitoring

Evaluation index

A systematic approach towards intentionally planning and designing

Soundscape description: Soundscape description: factors to be consideredfactors to be considered

system

Each source (type)

Sound pressure level

Spectrum

Psychological/social characteristics

Source movement

Location

Variation (hour, day, season)

Duration

Impulsive characteristics

Meaning

Natural or artificial sound

Relation to activities

Soundmark Effect of the space

Reverberation

Reflection pattern and/or echogram

General background sound

Sounds around the space

Users Acoustic condition at users’ home and work, experience, etc

Social/demographic factors, activities and behaviours

Other environments Temperature, humidity, lighting, etc

Visual, landscape and architectural characteristics

Descriptive or holistic

Temporal conditions

Soundscape design: activities and sound elementsSoundscape design: activities and sound elements

In addition to dividing sounds as keynotes, foreground sounds anIn addition to dividing sounds as keynotes, foreground sounds and d soundmarkssoundmarks, sound sources in an urban open space can be divided into: , sound sources in an urban open space can be divided into:

active sounds: active sounds: relate to sounds from the activities in the space, relate to sounds from the activities in the space, e.g. group dancinge.g. group dancing

passive sounds:passive sounds: relate to the sounds from the landscape elements, e.g. relate to the sounds from the landscape elements, e.g. fountainsfountains

Spectrum Spectrum –– also important when using also important when using psychoacoustic magnitudespsychoacoustic magnitudesdynamic processdynamic processacoustic zones and scaleacoustic zones and scale -- suitable aural space suitable aural space or sourceor source--listener distance for each zonelistener distance for each zone… …

An urban open space can be designed to An urban open space can be designed to encourage activities generating active sounds.encourage activities generating active sounds.

sounds

Designable factors/potentials

ActiveActive sound: musicsound: music

People are not only interested in the music itself, but are alsoPeople are not only interested in the music itself, but are also attracted by the attracted by the activitiesactivities of the players. In this case, the type of music (e.g. classicalof the players. In this case, the type of music (e.g. classical music or music or pop music) is not a very important issue. pop music) is not a very important issue.

However, when music is played using loudspeakers, the type of muHowever, when music is played using loudspeakers, the type of music as well as sic as well as the sound level needs to be considered carefully. the sound level needs to be considered carefully.

Most people donMost people don’’t like loud music played from loudspeakers, whatever the music tt like loud music played from loudspeakers, whatever the music typeype. .

Spectrum analysisSpectrum analysis dynamic processdynamic process

Acoustic zones

PassivePassive soundsound: water: waterEndless effects in colouring the soundscape – ‘primary soundscape quality’

Landscape theory: ‘primary landscape qualities’ - water and foliage

Spectrum: most water sounds have significant high frequency components around 2k to 8kHz and some of them also have notable low frequency components.

making water sound distinctive from the backgroundDynamicDynamic: The flow rate of a water feature should not be constant. : The flow rate of a water feature should not be constant.

Cascade with Temple Pavilion in the Chatsworth Garden

50

55

60

65

70

75

80

16 31.5 63 125 250 500 1k 2k 4k 8k 16kFrequency (hz)

SPL

(dB

)

1 step 2 steps5 steps 7 steps

Examples showing design potentials:high frequency components come from the water splash itself, whereas when a large flow of water is raised to a very high level and then dropped to a water body or hard surface, notable low frequency components can be generated

Sound field simulation and parameter studiesSound field simulation and parameter studies

-50

-45

-40

-35

-30

-25

-20

-15

-10

0 20 40 60 80 100 120

Source-receiver dis tance (m)

SPL

(dB)

25x25, diffuse25x25, geometrical50x50, diffuse50x50, geometrical100x100, diffuse100x100, geometrical

-50

-45

-40

-35

-30

-25

-20

-15

-10

0 20 40 60 80 100 120

Source-receiver distance (m)

SPL

(dB

) 50206

-50

-45

-40

-35

-30

-25

-20

-15

-10

0 20 40 60 80 100 120

Source-receiver distance (m)

SPL

(dB

)

50x50, diffuse100x25, diffuse50x50, geometrical100x25, geometrical

-50

-45

-40

-35

-30

-25

-20

-15

-10

0 20 40 60 80 100 120

Source-receiver distance (m)

SPL

(dB

)

0.1, diffuse0.3, diffuse0.5, diffuse0.7, diffuse0.9, diffuse0.1, geometrical0.3, geometrical0.5, geometrical0.7, geometrical0.9, geometrical

SizeSize

ShapeShape AbsorptionAbsorption

HeightHeight

BoundaryBoundary

space

Based on several models [Future Computational Acoustics Workshop]

Case studiesCase studiesFurther showing design potentialsFurther showing design potentials

Preconceptions:Preconceptions: Respondents predicted the concrete barrier to be the most effecRespondents predicted the concrete barrier to be the most effective, tive, although the transparent barriers and deciduous vegetation barrialthough the transparent barriers and deciduous vegetation barriers judged most ers judged most aesthetically pleasing aesthetically pleasing These preconceptions were reflected in the results of the furtheThese preconceptions were reflected in the results of the further perception exercise, r perception exercise,

virtual reality environment RAVEvirtual reality environment RAVEfive barrier: concrete, metal, timber, transparent acrylic, and aive barrier: concrete, metal, timber, transparent acrylic, and a hedgerow of deciduous vegetationhedgerow of deciduous vegetation

peoplesocial and demographic factors, including age, gender, education, profession, residential status (i.e. local and non-local), cultural background, and acoustic environment at home and at working place; activities including moving type such as walking, playing with children, and sport; and non-moving type such as sitting, standing, reading, and watching;behaviours such as wearing earphone and sunglasses. The results have clearly demonstrated the importance and potential in considering the characteristics of the users.

Preconceptions, aesthetics: Preconceptions, aesthetics: barrier as examplebarrier as example

Landscape and acoustic satisfaction

Visual and acoustic satisfaction

Beijing

Taipei

SheffieldCorrelation between quiet and view when

choosing a living environment(Beijing data)

EU data

Consider auralConsider aural--visual interactionsvisual interactions

Environment if a place is very hot or very cold, perhaps none cares about soundscape…

Process of ANN learningProcess of ANN learning

• Important to develop a model to predict the subjective evaluation of soundscape quality at the design stage, using known design conditions.

• Given the complicated relationships between various factors, ANN is suitable• Initial modelling results suggest that there is a good convergence

Neural Network Analysis of SoundscapeNeural Network Analysis of Soundscape

Biological neural netBiological neural net

Soundscapeevaluation

Other physical comfort

Social/demographic

Foregroundsounds

ThermalOther physical comfort

Visual

Lighting

Physical parameters relating to thermal

Social/demographic

Thermal

Foreground sound

Background sound

Activities

behaviours

Social/demographic

Acoustic space

Source characteristics

tools

To aid urban soundscape design and for public To aid urban soundscape design and for public participation, it would be useful to present the 3D participation, it would be useful to present the 3D visual environment with an acoustic animation toolvisual environment with an acoustic animation tool

Multiple sourcesMultiple sourcesSource and receiver all movingSource and receiver all movingCalculation speed fastCalculation speed fastCalculation accuracy less criticalCalculation accuracy less critical

ProgressProgressConsiderable simplifications can be made in terms of algorithms Considerable simplifications can be made in terms of algorithms for urban spacesfor urban spacesFurther simplification possible in terms of subjective evaluatioFurther simplification possible in terms of subjective evaluation, depending on source typesn, depending on source typesA fast algorithms CRR has been developed, especially for SPLA fast algorithms CRR has been developed, especially for SPLAuralization techniques are explored with various levels of accuAuralization techniques are explored with various levels of accuracyracy((HRTF for spatial sound with headphone or using RAVE)HRTF for spatial sound with headphone or using RAVE)

Acoustic animation and auralisationAcoustic animation and auralisation

Design guidelinesDesign guidelines

Concluding remarksConcluding remarks

Importance of interdisciplinary approachImportance of interdisciplinary approachImportance of introducing Importance of introducing ‘‘scientificscientific’’ methods into soundscape research/designmethods into soundscape research/design

Great potentials of planning/designing soundscape has been demonGreat potentials of planning/designing soundscape has been demonstrated.strated.A systematic approach towards intentionally planning and designiA systematic approach towards intentionally planning and designing ng soundscape will greatly benefit practice and policessoundscape will greatly benefit practice and polices ……

Policy and regulations needed for urban open public spacesPolicy and regulations needed for urban open public spaces

Looking forward to other talks today Looking forward to other talks today …….. ..

Space

Presentation design tools Evaluation Prediction

Practice

Policies

PeopleSounds EnvironmentDescription Framework

Research

Practice and potentials

Many other issues and other kind of spaces/soundscapes