Podium Acoustics for the Symphony Orchestra

Transcription

1 Podium Acoustics for the Symphony Orchestra An investigation of correlations between subjective and objective acoustic parameters M A L I N C E D E R L Ö F Master of Science Thesis Stockholm, Sweden

2 Podium Acoustics for the Symphony Orchestra An investigation of correlations between subjective and objective acoustic parameters M A L I N C E D E R L Ö F Master s Thesis in Music Acoustics ( credits) at the School of Media Technology Royal Institute of Technology year Supervisor at CSC was Erik Jansson Examiner was Sten Ternström TRITA-CSC-E : ISRN-KTH/CSC/E--/--SE ISSN Royal Institute of Technology School of Computer Science and Communication KTH CSC SE- Stockholm, Sweden URL:

3 Sammanfattning En människas upplevelse av en ljudvåg består av en källsignal faltad med reflektioner av vågen. Reflektionerna kallas för ljudets impulssvar. Den akustiska upplevelsen är, om vi bortser från vissa psykologiska faktorer, orsakad av objektiva egenskaper från ljudfältet. De objektiva parametrarna är beskrivna och definierade direkt från ljudets impulssvar. Det akustiska intrycket är inte enbart orsakat av det objektiva ljudfältet. Människans hörselsystem spelar en avgörande roll när det gäller upplevelsen av ett ljud. Det subjektiva intrycket en person har av att ljud skiljer sig från en annan persons upplevelse beroende på en mängd olika faktorer. Detta examensarbete behandlar korrelationer mellan subjektiva och objektiva rumsakustiska parametrar vid spel på podium i konserthus. Fem konserthus i Sverige har studerats, dessa är Berwaldhallen i Stockholm, Gävle Konserthus, Västerås Konserthus, Malmö Konserthus och Louis de Geer i Norrköping. De objektiva rumsakustiska parametrarna uppmättes på konsertsalarnas podium i var och ett av konserthusen. De subjektiva parametrarna undersöktes genom en enkätstudie vilken delades ut bland professionella musiker som arbetar i konserthusen, sammanlagt enkätsvar samlades in. Enkätsvaren och de objektiva mätningarna analyserades därefter. Till en början gjordes en statistisk analys med Spearmans rangkorrelation för att se korrelationer mellan de subjektiva och objektiva parametrarna. Analysen gav inga starka korrelationer varav arbetet gick vidare genom att studera de diagram som tagits fram för medelvärdena av resultaten. Genom denna studie utkristalliserades vissa tendenser. Analysen visade att Västerås Konserthus fick de högsta betygen av orkestermusikerna där varje orkester svarade på enkäten för sin hemarena. På den andra sidan av skalan befann sig Berwaldhallen och Malmö Konserthus. Västerås Konserthus jämfördes därför med Berwaldhallen och Malmö Konserthus gällande de objektiva parametrarna. Dessa parametrar jämfördes genom en studie av medelvärdena för varje parameter som uppmättes på samtliga podium. Detta visade att Västerås Konserthus hade höga Support (ST), Clarity (C) och Definition (D5) värden i jämförelse med de andra konserthusen. I Västerås Konserthus var Reverberation Time (T), Early Decay Time (EDT) och Center Time (Tc) lägre än i Berwaldhallen och Malmö Konserthus. Gävle som fick de näst högsta betygen av musikerna visade objektiva värden på liknande sätt som Västerås Konserthus i jämförelse med Berwaldhallen och Malmö Konserthus. Den parameter som skiljer är ST vilken inte visade på lika höga värden i Gävle Konserthus som i Västerås Konserthus. ST har diskuterats bland många akustiker huruvida parametern har en betydelse eller inte vid de subjektiva intrycken vid spel på podiet. I denna studie kan resultaten visa på en tendens att om STvärdet inte är högt så behöver det inte vara avgörande för den akustiska upplevelsen. Alla parametrar spelar en roll för hur musikerna upplever podiumakustiken. Det faktum att konserthusen har olika volymer skall även tas i beaktning vid jämförelsen. En statistisk analys av de subjektiva parametrarna visade att den subjektiva parametern som fick den högsta korrelationen med Totalupplevelse var Support. I Berwaldhallen och i Louis de Geer visar diagrammen för T att efterklangstiden är lägre i låga frekvenser än vid medelhöga frekvenser vilket inte följer den rekommenderade efterklangskurvan framtagen av akustikern Michael Barron []. I Berwaldhallen är detta tydligt och skulle kunna vara orsaken till varför musikernas betyg av Totalupplevelse är så lågt. I Louis de Geer är tendensen inte lika tydlig som i Berwaldhallen vilket gör det svårt att säga hur detta påverkar musikernas subjektiva bedömning. Uppdragsgivaren för detta examensarbete var Akustikon AB i Göteborg.

4 Summary A persons perception of a sound consists of the source signal convolved with the reflections of the sound. The reflections are called the impulse response. The acoustic impression is, if we ignore certain psychological factors, caused by objective properties of the sound field. The objective parameters are described and defined directly from the impulse response and the sound field. The acoustic impression is not only caused by the objective sound field. The human auditory system plays a conclusive role when it comes to perception of sound. The subjective impression a person have of a sound differ from another due to many things. This master thesis deals with the relationship between subjective and objective room acoustic parameters on concert hall podiums. Five concert halls in Sweden have been under investigation, these are Berwaldhallen in Stockholm, Gävle Concert Hall, Västerås Concert Hall, Malmö Concert Hall and Louis de Geer in Norrköping. The objective acoustic parameters were measured on the podiums in each concert hall and the subjective parameters were investigated through a questionnaire which was distributed among proffessional musicians working in the concert halls, in all responses were received. The questionnaire responses and the measurements of the soundfield were analysed. As a starting point the results, the subjective parameters and the objective parameters, were analysed statistically with the Spearmans rank order coefficient. This method did not give many strong correlations and the analysis method became instead to compair the results by studying the diagrams of the objective parameters and in that way see differences and similarities between the sound field and the questionnaire responses from the concert halls. Through this analysis some tendencies were found. The analysis showed that Västerås Concert Hall got the highest marks from the musicians and on the other side of the subjective ranking are Berwaldhallen and Malmö Concert Hall. Västerås Concert Hall was therefore compaired to Berwaldhallen and Malmö Concert Hall. The objective parameters were compaired by studying the mean values of each parameters measured on the podiums. This showed that Västerås Concert Hall has high ST, C and D5 values in comparison with the other halls while T, EDT and Tc are lower. Gävle who got the second best rank from the musicians showed similar objective results as Västerås Concert Hall in comparison to the other concert halls. The parameter that differ is ST who is not high in comparison to the other halls. Many acoustician have discussed whether ST is important or not for the acoustic perception while playing on the podium. This study could show that if the ST is not high, this may not be conclusive for the acoustic perception. All parameters play a part in the way of how the musicians perceive podium acoustics. The fact that the concert halls not have the same volume must be considered in the comparison. The analysis showed also that the subjective parameter which got the highest correlation with Overall Impression is Support. In Berwaldhallen and Louis de Geer the reverberation time diagrams show that the reverberation time is shorter for low frequencies than for the middle frequencies which is not like the recommend reverberation time curve made by the acoustician Michael Barron []. In Berwaldhallen this is prominent and it may be one of the reasons why the musicians ratings of Overall Impression is so low. In Louis de Geer the tendency is not as clear as in Berwaldhallen which make it difficult to say how this affect the musicians subjective judgements. The assigner for this project was Akustikon AB in Gothenburg.

5 Acknowledgements I was able to conduct this master thesis with help from a number of inspiring, professional and supporting persons. I would therefore like to give my thanks to: Jan-Inge Gustafsson Akustikon AB Mats Olsson Akustikon AB Gunilla Sundin Akusikon AB Liselott Tillander Akustikon AB Erik Jansson TMH All musicians in: The Swedish Radio Symphony Orchestra Berwaldhallen in Stockholm The Malmö Symphony Orchestra Malmö Concert Hall The Västerås Symphonietta Västerås Concert Hall The Norrköping Symphony Orchestra - Norrköping Concert Hall The Gävle Symphony Orchestra Gävle Concert Hall All employees at each concert hall for all your help and kindness during my visits.

6 List of Abbreviations Tc C D5 EDT ST SPL T R/D Center Time Clarity Definition Early Decay Time Support Sound Pressure Level Reverberation Time Reverberant/direct ratio

7 Table of contents. Introduction.... Presentation of concepts..... Objective parameters..... Subjective parameters.... Earlier studies.... Description of the halls..... Berwaldhallen..... Louis de Geer, Concert- and Congress Hall The Gävle Concert Hall..... Västerås Concert Hall Malmö Concert Hall.... Project Aims...9. Method...9. Questionnaire Survey Choice of subjective parameters..... Analysis of the subjective parameters..... Presentation of results Correlations between subjective parameters...7. Measurement of objective parameters..... Choice of objective parameters..... Equipment and Measurement Set-Up WinMLS..... Measurement implementation..... Analysis of the objective parameters Presentation of results.... Analysis method Discussion.... Conclusion Future work.... References...7

8 APPENDICES A Mean value diagrams of the objective parameters A. Mean values for Reverberation T A. Mean values for Clarity C A. Mean values for Definition D5 A.5 Mean values for Early Time Delay EDT A.5 Mean values for Center Time Tc A. Mean values for Support ST B The Questionnaire C Diagrams of the subjective parameter results from the questionnaire responses C. Berwaldhallen C. The Gävle Concert Hall C. Malmö Concert Hall C. Västerås Concert Hall C.5 Louis de Geer, Concert- and Congress Hall D Results from the questionnaire responses - questions about instrument, hearing impairment and ear defenders D. Berwaldhallen D. The Gävle Concert Hall D. Malmö Concert Hall D. Västerås Concert Hall D.5 Louis de Geer, Concert- and Congress Hall

9 . Introduction A musician works during his lifetime on reaching a particular tone and timbre on his instrument. Nevertheless, if the acoustics of the concert hall does not support the musician while playing, the optimal tone will never reach the listener. The concert hall, which can be regarded as an extension of the musician s instrument, plays a conclusive part when it comes to how a musician perceives his instrument and the tones from the other musicians on the podium. How does the musician perceive the acoustics on the podium? How does the objective sound field correspond to the subjective impression? The purpose of this master thesis is to investigate correlations between subjective and objective acoustic parameters on the podium by studying five concert halls in Sweden. The target group of this thesis is students in the field of sound and acoustics, professionals in the field and musicians. The assigner for this project is Akustikon AB in Gothenburg. The objects studied in this project are five concert halls in Sweden Berwaldhallen in Stockholm, Louis de Geer Concert- and Congress Hall in Norrköping, Gävle Concert Hall, Västerås Concert Hall and Malmö Concert Hall. The study was carried out partly by measurements of the sound field on the podiums and partly through a questionnaire survey among the members of the professional symphony orchestra at each concert hall. This chapter will present the background to this master thesis. Firstly, concepts of importance will be described. Secondly, comes an overview of earlier studies within the field. Finally, the five Swedish concert halls studied in this thesis will be described in detail.. Presentation of concepts The acoustics of a room is of complex nature and must be investigated from several points of view. In this master thesis the objective sound field and the subjective impression will be investigated... Objective Parameters A persons perception of a sound consists of the source signal convolved with the reflections of the sound (the impulse response). A sound field is therefore often described by the impulse response (see Figure ) which gives almost all information about the sound for a particular source and receiver position. A sound is generated from a vibration of an object. The vibrations produce a pattern of changes in pressure in the surrounding medium (usually air) which make that the atmospheric particles are squeezed closer together than normal, called condensation, and then pulled father apart than normal, called rarefaction. The air particles move backwards and forwards parallel to the direction of the motion of the wave, therefore are sound waves longitudinal.[] The acoustic impression is, if we ignore certain psychological factors, caused by objective properties of the sound filed. A listener receives about reflections of a short sound within one second of the direct sound.[] Each reflection has a delay, a sound level and a direction. The objective parameters are described and defined directly from the impulse response and the sound field. See chapter.. for the objective parameter definitions of interest in this thesis.

10 Figure : Impulse response on stage. [].. Subjective parameters The acoustic impression is not only caused by the objective sound field. The human auditory system plays a conclusive role when it comes to perception of sound. The auditory system, composed of the outer, middle and inner ear, works in a particular way with the basilar membrane (BM) as its bandpass filter. Each point at the BM, which runs along the length of the cochlea, corresponds to a certain frequency which gives the maximum response at the corresponding point. The inner hair cells supply mechanical movements into neural activity. [] The ear localize the direction of a sound through the direct sound, the first wave front received. The localization ability is such that it can function even though the direct sound is weaker than the combined energy of later reflections. Sound perception is complex and especially the neural activity is far from completely understood by the researchers in the field.[] Persons perceive sounds in different ways and the subjective hearing is very complex. The impression a person have of a sound differ from another due to many things. Among others different persons have different experiences in perception of sound which gives them different references in the field. Good acoustics for one person may not be perceived as good acoustics for another individual. The Tacit Knowing Theory which was proposed by M. Polany tries to explain the human cognitive process. This theory was applied to music players and deals with the relation between the musician and his instrument where actions and reactions are exchanged between the musician and the instrument. In the same way, if the musician is playing in a concert hall, he adjust his playing technique according to the support he receives from the hall.[] In this master thesis the subjects who answered the questionnaire are all employees in a professional symphony orchestra which have an own home arena with which they are familiar, and where they are used to perform and rehearse. See chapter.. for subjective parameter definitions of interest in this thesis.

11 . Earlier studies Concert hall acoustics deals with conditions for the audience and for the musicians on the podium. Today the preferred conditions for the audience are quite well understood while the optimal conditions for the musician are less clear. Studies on how musicians perceive the acoustic in concert halls have been through the years investigated both with experimental setups in anechoic rooms, experimental set-ups on the podiums and also thorough interviews and questionnaires. This chapter is a short review of earlier studies in the area. In 9 M. Barron used a questionnaire with eight continuous bipolar semantic differential scales. The parameters were Clarity, Reverberance, Envelopment, Intimacy, Loudness, Balance, Background noise and Overall Impression. It was found that the subjects could be divided into two groups: those that prefer Intimacy, which describes how the listener experiences that he is acoustically involved or detatched by the music, and those that prefer Reverberance. The scale Envelopment, which describes how the listener experiences that he is sourrounded by the music, was found to be interpreted differently by the two groups.[] In 99, the Danish researcher A.C Gade investigated the relationship between subjective room acoustic needs of performers of classical music and objective properties of the sound field on orchestra podiums in his licentiate project. A.C Gade suggested two new objective parameters: Support (ST) and Early Ensemble Level (EEL). Gade define ST as the ratio between the energy of the early reflections and the energy of the direct sound. This is measured m from the source. Gade intend that a high value of ST correspond to a strong feeling of the subjective parameter Support. To compensate the uncertainty concerning the contribution of later reflections, Gade suggest upper integration limits within the ST parameter. The limits are ms and ms which are the X in the equation below. ST = log ( E(-X ms)/ E (dir) ) The parameter EEL is defined as the ratio between the received early energy and the energy emitted. This is also measured at a distance of m as the ST parameter. Gade meant that a higher value of EEL correspond to a better level of the subjective parameter Hearing each other. EEL = log ( E(- ms)/ E (dir) ) [9] In A.Farina studied eight opera houses employed for symphonic music in Italy. A questionnaire was distributed among well-known musicians and objective parameters were measured. Some correlations were found between the physical acoustic parameters and the subjective parameters. For example the subjective pair pleasant-unpleasant which was included in the questionnaire seemed to be affected by many of the objective parameters, for example C, SPL and R/D. Finally a short questionnaire suitable for further psychoacoustic analysis was obtained. [] In J. Sanders, Marshall and Day Acoustics studied the hypothesis that, for musicians in small ensembles rather than large orchestras, the ability to hear each other is of higher importance in their assessment of a hall. They did a survey in halls in New Zealand to find the halls which were most suitable for chamber music. Using the average of Overall Impression the halls were ranked from most preferred to least preferred. The correlation were calculated between subjective parameters and it was observed that Support was the most highly correlated with Overall Impression. []

12 . Description of the halls The five concert halls which constitute the investigation subjects in this master thesis are all located in Sweden and they all plays the role of being the home arena for a professional symphony orchestra. Table Main characteristics for the five concert halls No. Name Number of seats Height (m) Volume (m ) Stage area (m ) Construction year Malmö Concert Hall - 95 Berwaldhallen 979 Västerås Concert Hall Gävle Concert Hall, Louis de Geer Berwaldhallen After many years of planning, Berwaldhallen was inaugurated in 979. The concert hall which also is a music studio for the Swedish Radio is the home arena for The Swedish Radio Symphony Orchestra which has musicians and was established in 9.[] The concert hall is named after the Swedish composer Franz Berwald (79-). At the time for the building process the acoustic consultant was Vilhelm Lassen Jordan who also was the acoustic consultant for the Opera house in Sydney and the Oslo Concert Hall.[] The concert hall has a six-sided shape and not the traditional shoe-box shape. There are balconies around the podium. The building is designed with two thirds of the volume down in the ground.[] At the time of the inauguration the acoustics got a lot of criticism. The violins, who where famous for their timbre, did not sound as good as they used to.[] Changes where made at that point, and there have been some further adjustments after. Berwaldhallen view towards the podium

13 .. Louis de Geer, Concert- and Congress Hall The concert hall in Norrköping -Louis de Geer, a former industry building, was inaugurated in 99. It has a number of seats and a volume of m³ (9.*.5*5. m²). The stage area is -55 m² including an orchestra pit. The architect was Lund & Valentin architects in Gothenburg and the acoustic consultants were Akustikon AB in Gothenburg. The hall has a reverberation time of.. seconds, it is adjustable by different acoustic tools. The shape of the hall is not the traditional shoe-box but an eight-sided shape with three balcony sections and an opening for the podium.[] Louis de Geer is the home arena for the Norrköping Symphony Orchestra which has musicians.[5] The ceiling has a cone shape and are constructed with insets of ripples for sound diffusion. The sound in the hall can be adjusted by textiles on rollers placed partly in the middle area of the hall and partly in the upper sections of the wall. From the beginning the hall was not flexible enough and was reconstructed after a few years. One change that improved the acoustic on the podium was when the reflectors in the ceiling were divided into smaller pieces. Above the stage there are five movable reflectors that can be used to adjust the balance of the orchestra. [] Louis de Geer, Concert- and Congress Hall view towards the podium 5

14 .. Gävle Concert Hall Gävle Concert Hall is a former swimming pool complex and was in 99 inaugurated as the new home for the Gävle symphony orchestra with its 55 musicians. The former pool area became the concert hall with a volume of 5 m³ (.5*.*. m²) and seats- 7 on the lower stall, 5 on the upper stall, 7 on the first balcony and on the second balcony.[] The reverberation time of the hall is..9 seconds in the hall and the stage area is m². The architects were the same as in the Louis de Geer project, Lund & Valentin architects and the acoustic consultants were Akustikon AB in Gothenburg.[] The hall has an oval shape. A grid with 7 acoustic boards are located above (7. m. m) the stage and first stall.[7] The hall is, unlike Louis de Geer, integrated with the podium in one piece. The hall and the podium is separated only by the reflectors which are located over the podium, which are there to give early reflections to the hall and to the orchestra.[] The Gävle Concert Hall consists of two halls for music performance. The larger which is used by the symphony orchestra is called The Gevalia Hall. The smaller is called The Bo Linde Hall and is used for chamber music and smaller concerts.[] The Gävle Concert Hall view towards the podium

15 .. Västerås Concert Hall Västerås Concert Hall has a different solution to adjust the acoustics then the other concert halls described in this chapter. There are diffusion chambers around the hall with doors which can adjust the acoustics by be opened gradually. The reverberation time can be reduced from. sec to.5 sec by using the diffusion chambers. Behind the doors it is possible to roll down textile curtains to change the reverberation time. It is also possible to vary the degree of how much the doors are open. By only open the doors a little you can get a low frequency absorption.[] The concert hall which was inaugurated in has a volume of 95 m³ (7.* 9.7*. m²) and a stage area of 9 m². The reverberation time is adjustable between..5 seconds and the clarity is.. db. The architects of Västerås Concert Hall is Archus Arosia Västerås and the acoustic consultants were Akustikon AB in Gothenburg.[] The concert hall is the home arena for Västerås Symphonietta, which is one of the oldest orchestras in Sweden with roots from a music society in. The orchestra has members who are half-time employees.[] Västerås Concert Hall view towards the podium 7

16 ..5 Malmö Concert Hall Malmö Concert Hall is the home arena for the Malmö Symphony Orchestra and its 9 professional musicians. The concert hall has seats and its podium has an area of (*m) m². The hall which was inaugurated in 95 is situated in the city centre of Malmö.[] The stage height is in the front metre and at the back metre. The hall has an shoe-box shape with an opening for the podium. In the ceiling there are reflectors to give early reflections to the podium and the orchestra. The arcitect of Malmö Concert Hall is Sten Samuelson and the acoustic consultants were Ingemansson AB.[] Malmö Concert Hall view towards the podium

17 . Project aims The aim of this project was to investigate correlations between the physical sound field on the podium and the musicians perception of the sound. The main questions were: - How are the concert halls ranked on the basis of the questionnaire responses? Which hall got the highest rank? - What similarities and differences of the sound field can be seen between the concert hall with the highest rank in comparison to the hall with the lowest rank? - How do Overall acoustic impression correlate with the other subjective parameters? The investigation included the following steps: - Performing measurements of the impulse response in 7 positions on the podium in 5 concert halls. In all 7 measurements on each podium (see chapter..). - Setting up a questionnaire which included a number of subjective parameters, each with a -step numerical scale. - Performing a statistical analysis of the questionnaire responses. In all responses were received. - Analysis of the objective measurements with WinMLS.. Method The study was carried out partly by measurements of the sound field on the podium and partly through a questionnaire survey among the members of the professional symphony orchestra at each concert hall.. Questionnaire Survey The aim of the questionnaire survey was to map out the performers perception of the acoustics in their concert hall. The subjects for the questionnaire were professional musicians in The Swedish Radio Orchestra, Malmö Symphony Orchestra, Gävle Symphony Orchestra, Norrköping Symphonietta and Västerås Symphonietta. The questionnaire was sent by mail to the orchestra directors at each concert hall. The orchestra director distributed the questionnaire to the musicians. Depending on the orchestras schemes, the questionnaires were completed and returned after -5 weeks. Assuming that all musicians in the orchestras received a questionnaire, the response rate was: - The Swedish Radio Symphony Orchestra % - Malmö Symphony Orchestra % - Norrköping Symphony Orchestra 55% - Gävle Symphony Orchestra % - Västerås Symphonietta 7%. A total number of musicians completed the questionnaire. 9

18 .. Choice of subjective parameters A set of subjective parameters has been carried out as the basis for the questionnaire. The starting point for the questionnaire of this master thesis was a questionnaire that Jan-Inge Gustafsson at Akustikon AB designed in collaboration with the Danish researcher A.C Gade. The questionnaire survey they did was at Rotundan in The Opera House in Stockholm in 999 []. By studying their questionnaire and by studying an investigation made by A.C Gade (Investigations of musicians Room Acoustic Conditions in Concert Halls) where he interviewed musicians of classical music and asked which acoustic qualities they judge as the most important at the podium [9], the parameters for this questionnaire was carried out. The parameters in the questionnaire are: - Overall acoustic impression - Support - Dynamics - Hearing each other - Loudness - Hearing oneself - Reverberation - Clarity of the conductor speech Overall Acoustic Impression is the average impression the musician have of the acoustics [9]. Support is the parameter which describes how the musician feel that it is necessary to force the instrument to be able to hear himself. It is effective even while short tones are played and therefore a property different from reverberation [9]. Dynamics describes the dynamic range obtainable in the room and the degree to which the room obeys the dynamic intention of the player [9]. Hearing each other is the property necessary in order for a number of musicians to play in ensemble, i.e. with rhythmic precision, in tune and in balanced levels [9]. Loudness stands for how loud the musician perceive the music on the podium [9]. Hearing oneself is the ease of hearing your own instrument while playing [9]. Reverberation is heard for example in a break and consists of the tone that just have been played [9]. Clarity of the conductor speech is defined as the ability to hear speech on the podium. The questionnaire has two different parts. One part has questions about the perception of the acoustics when playing with full orchestra while the other part asks the same questions but while playing in a smaller ensemble. It is important to include both parts in the investigation for two reasons. First it is interesting to know if there are some of the concert halls where the musicians experience that there is better to play in a small ensemble than in a bigger, or vice versa. The other aspect is to compare for example Västerås Concert Hall (where the Symphonietta and the house is smaller than the other halls and orchestras) with the other

19 larger concert halls, i.e. it is relevant to compare Chamber music in Berwaldhallen with Full orchestra in Västerås Concert Hall. In the questionnaire questions about hearing impairment and ear defenders were included to see how the musicians work environments are and also to see if their perception is affected by any hearing impairments. To assure that the questionnaire was easy to understand for the musicians, three musicians all active players in amateur symphony orchestras revised the questionnaire. After their review changes were made before it took its final form (see Appendix B)... Analysis of the subjective parameters The questionnaire responses were, for each concert hall, statistically analysed with the computer programme MATLAB[]. Mean value, standard deviation, max value, min value and the median were calculated (see chapter..). During the analysis, the scale for Hearing oneself and Reverberance was translated to a - scale to make the statistical analysis easier. The correlations between Overall impression and other subjective parameters were analysed statistically with the Spearmans rank order coefficient. This method is good to use when you have data where the specific value is not important, it is the order of the data that you want to use. This means that, if you have a number of data values, for example.,.9 and., you translate. by,. by and.9 by etc. [5] One example of Spearmans rank order coefficient, when you have two data series, x and y, and the orders x-rank and y-rank, d is the difference between the two ranks: [] Spearmans rank order coefficient, rs, is then calculated by the fomula: [] where n is the number of data pairs In this project Spearmans rank order coefficient was calculated using a free web application from VassarStats. [7]

20 .. Presentation of results Berwaldhallen Berwaldhallen was one of two (Malmö Concert Hall is the other) who got the lowest marks from the musicians. Comments from the musicians are (translated from Swedish by the author) - The hall sounds blaring with a big orchestra but unpersonal with a smaller orchestra., Hard unsympathetic acoustics. Fortissimo in brass and timpani is a huge problem for the audience, We must actively adjust us to the hall through a softer way of playing, no matter what the conductor says, The hall is great for chamber music but terrible for symphonic music For chamber music Berwaldhallen got a higher rank than for full orchestra. In Berwaldhallen (%) musicians answered the questionnaire. Table Mean value for subjective parameters in Berwaldhallen Full orchestra Mean value Std Max Min Overall acoustic impression,, Support,5 Dynamic,5 Hearing others,,9 Sound level,7, 9 *Hearing oneself,,7 5 *Reverberation,5, Conductor speech,9, Chamber music Overall acoustic impression 5,9, 9 Support 5,,5 Dynamic,, Hearing others 5, Sound level 5,7, *Hearing oneself,5, *Reverberation,9, *The optimal value are 5 (see chapter..) Std - Standard Deviation

21 Malmö Concert Hall Malmö Concert Hall got a low rank in comparison to the other concert halls. A commentary in the questionnaire showed a certain discontent - The only way to improve the podium acoustic is unfortunately to build a new concert hall. Malmö Concert Hall was the hall where the number of responses were lowest, (%) fully compiled answers. Even though the responses in Malmö were few, the answers came from different instrument groups which will give a good view of the opinions in the orchestra. Table Mean value for subjective parameters in Malmö Concert Hall Full orchestra Mean value Std Max Min Overall acoustic impression,, 7 Support,5, Dynamic,, 9 Hearing others, 7 Sound level,, 9 *Hearing oneself,9,9 *Reverberation,, Conductor speech,, Chamber music Overall acoustic impression 5,9 Support,,9 Dynamic,,5 Hearing others,9, 7 Sound level,,5 *Hearing oneself,, 7 *Reverberation, *The optimal value are 5 (see chapter..) Std - Standard Deviation

22 Louis de Geer Louis de Geer is the concert hall where the highest number of musicians answered the questionnaire, (55%) responses. The commentaries in the questionnaire were (translated from Swedish by the author) - The acoustic is blaring and has a lack of bass frequencies, The acoustic perceives as having to much of high frequency content and being sharp, Good spacious sound but it could have been softer and more bass, Hard timbre, lack of over tones. Judging from the commentaries it is clear that the musicians experience an acoustic with lack of bass frequencies. Table 5 Mean value for subjective parameters in Louis de Geer in Norrköping Full orchestra Mean value Std Max Min Overall acoustic impression,5,9 Support 5, Dynamic,,9 Hearing others,5, 9 Sound level,,9 9 *Hearing oneself,7,9 7 *Reverberation 5,, 9 Conductor speech 5,,9 Chamber music Overall acoustic impression 5,5, Support 5,, Dynamic,5, Hearing others,7, Sound level,, *Hearing oneself,,9 *Reverberation 5,,5 9 *The optimal value are 5 (see chapter..) Std - Standard Deviation

23 Gävle Concert Hall Gävle Concert Hall is one of the two (the other is Västerås Concert Hall) with highest rank from the musicians. The commentaries are (translated from Swedish by the author)- The double-basses sound a bit muddy, The strings sound weak and the brass has a tendency to be too loud. In all 7 (%) responses from The Gävle Symphony Orchestra. Table Mean value for subjective parameters in Gävle Concert Hall Full orchestra Mean value Std max Min Overall acoustic impression 5,5,7 9 Support,,7 Dynamic 7, Hearing others 5,,9 Sound level 5,, *Hearing oneself,,7 7 *Reverberation 5,9 9 Conductor speech,, Chamber music Overall acoustic impression,,5 Support,, Dynamic,9, Hearing others 5,, Sound level 7,5, *Hearing oneself,5,5 7 *Reverberation 5,5,9 9 *The optimal value are 5 (see chapter..) Std - Standard Deviation 5

24 Västerås Concert Hall Västerås Concert Hall is very popular among the musicians in Västerås Symphonietta. Overall acoustic impression was 9, out of. Comments from the musicians are (translated from swedish by the author) - A little bit to much treble, It would be good if it was possible to, through moderate the hall, could get the same conditions on rehearsals as at concerts. It is not the podium that is the problem, the problems disappear at concerts. In all (7%) responses from The Västerås Symphonietta. Table 7 Mean value for subjective parameters in Västerås Concert Hall. Full orchestra Mean value Std max Min Overall acoustic impression 9,,9 9 Support 9,7,99 7 Dynamic,9,7 Hearing others 7,,7 Sound level,59, *Hearing oneself 5,, 7 *Reverberation 5,7,7 Conductor speech 7,9,5 Chamber music Overall acoustic impression 9,5, 9 Support 9,, 7 Dynamic,9, 7 Hearing others,9,95 7 Sound level 9,9, *Hearing oneself 5,9,9 *Reverberation 5,7, *The optimal value are 5 (see chapter..) Std - Standard Deviation

25 The results imply that the musicians in Västerås Symphonietta are pleased with their acoustic conditions. Their answers do not differ much. In The Swedish Radio Orchestra and Malmö Symphony Orchestra the answers differ more and there are the musicians not as pleased as in Västerås. Uneven opinions can be a sign of discontent. Hearing oneself and Reverberation had a very low standard deviation when compared to the other subjective parameters. By averaging Overall Impression, the Concert Halls are ranked, on the basis of the questionnaire responses, in the following order (highest rank first): Full Orchestra:.Västerås Concert Hall.Gävle Concert Hall.Louis de Geer. 5. Berwaldhallen and Malmö Concert Hall Chamber music:.västerås Concert Hall.Gävle Concert Hall.Louis de Geer.Berwaldhallen 5.Malmö Concert Hall A question was included in the questionnaire where the musicians are asked to point out their position on the podium. This question did not give any clear results, it was not possible to conclude where the musicians were seated. Therefore are the results of that question not included in the results. The results of the questions on hearing impairment and ear defenders as well as the number of responses from different instrumental groups can be seen in Appendix D.... Correlation between Overall impression and other subjective parameters To see what parameters the musicians perceive as Overall Impression, correlations were calculated with Spearmans rank order coefficient (see..) between Overall Impression and the other subjective parameters in the questionnaire. The calculations showed that Support has the highest correlation with Overall Impression and the second most correlating parameter was Hearing each other. These results were expected from earlier studies, among others in []. Table Correlation coefficients, rs, between Overall Impression and the other subjective parameters. Level of significance at,% are written in bold numbers. Overall Impression rs Support, Hearing each other,7 Sound Level,5 Dynamic, Hearing oneself, Reverberation -, 7

26 . Measurement of objective parameters The measurements of the objective parameters were done when the podium was empty. It was necessary to have an empty podium to be able to reconstruct the same situation in all halls... Choice of objective parameters The objective parameters of interest in this study are calculated using the ISO standard [] by the WinMLS software. The parameters analysed are Clarity, ST, Definition, T, Early Decay Time and Center Time. These parameters are important and commonly used when studying ensemble perception.[] Clarity is the early to late index and is defined as: where t e is the early time. h(t) is the impulse response. C is Clarity with an early time of ms. C5 is Clarity with an early time of 5 ms. C is typical for music application and C5 is used for speech evaluation.[] Support is defined as: where the impulse response is assumed to be measured with an omnidirectional microphone at a distance of meter from the source. t d is the direct time, t e is the early time and t l is the late time. ST is defined with the direct time ms, early time ms and late time ms. There is also ST, which is defined with the late time of ms. But for measurements of ensemble conditions, ST is more relevant than ST which is because in ensemble playing, it is more important with a fast transmission of the energy. [][]

27 Definition is the early to total index and is defined as: Definition is often used when analysing speech signals and the early time is usually taken as 5 ms. [] Reverberation time EDT (s) is defined as the db decay time calculated by a line fit to the portion of the decay curve between and - db. T (s) is defined as the db decay time calculated by a line fit to the portion of the decay curve between -5 and -5 db. If the decay range is less than 5 db, the T is in WinMLScalculated using the range available and is displayed with red colour.[] Center Time: is the first moment of the squared pressure impulse response, expressed in milliseconds: [] 9

28 .. Equipment and Measurement Set-Up At each podium 7 positions were fixed after the same template (see Fig ). Position - is located meters in from the middle point at the front of the podium and then meters to the left from that point. Positions, and are all located at a straight line, metres from the centre. Position and 7 - are located meters from the front of the podium and meters from the side wall. Position 5 is located meters in from the side wall and at the same height as position. Position are located meters from the back side of the podium and meters from the side wall. These positions were carried out according to this pattern to cover the podium in a way that represent the different instrument groups and the main part of the podium. Fig measurements positions with position distances from Västerås Concert Hall

29 Fig Measurement equipment set-up All the measurements were performed with a microphone and a multidirectional dodekaeder loudspeaker. The loudspeaker is designed by Akustikon AB and built by Jan Stridbeck. A sine sweep was used to measure the impulse response which was done with the software WinMLS. The microphone used under the measurements was an omnidirectional microphone AUDIX TR- serial number BOOE-5. The loudspeaker and microphone were both. meters from the floor. That distance positioned represents the height of the musician while sitting on a chair.... WinMLS WinMLS is a Windows based software which for performing and exaluating audio, acoustical and vibrational measurements. The first part in the name WinMLS stands for Windows and the second part, MLS (maximum length sequence), is a measuring method which was the most powerful in 99. Today often a sine sweep is used instead, but MLS is still available as an option in WinMLS and is recommended for speech measurements. FFT and octave band analysis is possible with WinMLS and impulse responses can be measured with the software. [] The sound card used was Digigram VXpocket V..

30 Fig WinMls- screen dump.. Measurement implementation Measurements were performed between all the different positions. The loudspeaker was placed at position and the microphone at position, then the microphone was moved to position,, 5,, 7 and new measurements were performed. The loudspeaker was then moved to position and the microphone was moved to position,, 5,, 7 and new measurements were performed. With this method measurements between all positions were performed. The parameter Support was measured at the source-receiver distance meter. That was done at positions,,,, 5 and. The microphone was located meter in the direction to the middle line of the podium. In all, 7 measurements on each podium were performed... Analysis of the objective parameters The impulse responses that were measured in the concert halls were analysed with WinMLS. Values for the objective parameters were calculated for each podium position and a mean value for the whole podium of every hall was also calculated by the programme (see appendices A). During the analysis, the frequencies of interest were.5- khz, these are regarded as the most important when it comes to music perception. []..5 Presentation of results In this project it was not possible to analyse each position and each instrument group. Further investigations have to be carried out to see any correlation results from each position. In this project the mean values for each concert hall (see appendices A) were compared and analysed.

31 Table : Mean values of objective parameters Table 9: Mean values of objective parameters at the podium in Berwaldhallen. at the podium in Malmö Concert Hall. Measurements performed -5-7 Measurements performed -- Berwaldhallen mean values 5 Hz Hz Hz Malmö mean values 5 Hz Hz Hz T (s). C (db)... D5 (%) Tc (ms) 9 9 EDT (s).7 ST (db) T (s)...9 C (db)... D5 (%) 5 Tc (ms) 9 EDT (s)... ST (db) Table : Mean values of objective parameters Table : Mean values of objective parameters at the podium in Louis de Geer. at the podium in Gävle Concert Hall. Measurements performed -- Measurements performed -- Louis de Geer mean values 5 Hz Hz Hz T (s)...9 C (db).. 5. D5 (%) Tc (ms) 9 EDT (s)...5 ST (db) Table : Mean values of objective parameters at the podium in Västerås Concert Hall. Measurements performed --7 and -5-9 Gävle mean values 5 Hz Hz Hz T (s)...7 C (db) D5 (%) Tc (ms) 7 EDT (s)...5 ST (db) Västerås mean values 5 Hz Hz Hz T (s).9..7 C (db)..5. D5 (%) Tc (ms) EDT (s).7..7 ST (db) Västerås Concert Hall has, in comparison to the other concert halls (above all Berwaldhallen) high ST values and low T, EDT and Tc. In Västerås Concert Hall, C and D5 are high in comparison to for example Berwaldhallen which got low rank from the subjective investigation. Berwaldhallen has a T curve (see appendix A) that looks different from the other halls. In the lower frequencies the reverberation time is lower than in for example Västerås Concert Hall. Gävle Concert Hall showed the same tendencies as Västerås Concert Hall. In comparison to Berwaldhallen and Malmö Concert Hall, Gävle Concert Hall has high C and D5 values at the same time as T, EDT and TC are low. On the other hand, the ST values is not as high as in Västerås, and in comparison with Berwaldhallen and Malmö Concert Hall the ST values in Gävle Concert Hall is not high.

32 . Analysis method The subjective parameters and the objective parameters were analysed statistically with the Spearmans rank order coefficient (see chapter..) to see how the subjective impression and the objective sound field correlated. For each subjective parameter, a correlation coefficient was calculated with each objective mean value at 5 Hz, Hz and Hz. The method did not give many significant correlations due to the low number of concert halls under investigation. For this reason, the analysis method became instead to compare the results by studying the diagrams of the objective parameters. Västerås concert hall got the highest rank on the basis of the questionnaire responses from the musicians. At the other side of the subjective ranking are Berwaldhallen and Malmö Concert Hall. These two halls were therefore compared to Västerås Concert Hall to see if there are any similarities or differences between their objective parameter values. 5. Discussion The results of the analysis of the questionnaires showed that it is, in some of the concert halls, a large spread in how the musicians perceive the acoustics, a larger spread then was expected. On the other hand, the musicians in Västerås Symphonietta are generally very pleased with the acoustics on the podium and the responses did not differ much (see appendices C). The strong deviation of opinion on some of the concert halls could depend on discontent with the acoustic conditions. It can also depend on the musician s position on the podium, and what instrument they are playing. One drawback with the survey is that it was a low number of questionnaire responses from some of the concert halls. This could depend on the orchestras rehearsal scheme which made it difficult to reach all the musicians. Nevertheless, the responses cover most of the instrument groups and gives a reliable picture of the opinions in the orchestras. All musicians in the study are used to reherase and perform in their home arena which gives them similar experience in how they perceive the acoustics of the concert hall. This was important for the study. If the musicians had different experiences of the concert hall, their answers would not have been comparable. The subjective parameters that had a very low standard deviation in comparison to the other subjective parameters where Hearing oneself and Reverberation. This could imply that musicians have a more similar experience in how they perceive these parameters in comparison to other subjective parameters. One explanation could be that it is easier for the acoustic consultants to satisfy the musicians needs for Hearing oneself and Reverberation[].

33 . Conclusion Subjective rank Västerås concert hall got the highest rank on the basis of the questionnaire responses from the musicians. The hall seems to have the ultimate conditions playing symphonic music according to the musicians. At the other side of the subjective ranking are Berwaldhallen and Malmö Concert Hall. It can be seen from the analysis of the subjective parameters that the parameter which got the highest correlation with Overall impression is Support. The result was the same as can be found in other investigations, among others in []. Differences and similarities The analysis showed that Västerås Concert Hall has high ST values in comparison to the other halls. Västerås got a value for ST of -db (5Hz) while Berwaldhallen got -db (5 Hz). In A.C Gade s study in 99, he found that the highest ST value was measured in the Tivoli Concert Hall, Copenhagen, where ensemble playing was regarded as very easy. It can be seen, with Gade s results and the results in this thesis, that a high ST may be desirable for a good acoustic experience. The ST value can on another hand though not be too high which can be experienced as the sound becoming too loud, which was found by Gade in the Royal Theatre in Copenhagen []. Gävle Concert Hall got the second best marks from the musicians, and also shows similar objective values as Västerås Concert Hall. The C and D5 values are higher than in Berwaldhallen and Malmö Concert Hall. T, EDT and Tc are lower. The parameter value that differ in comparison with the other halls are ST. Gävle Concert Hall do not show as high ST values in comparison to Berwaldhallen and Malmö Concert Hall as from Västerås Concert Hall. This imply that even if the ST is low, this may not be conclusive for the acoustic perception. All parameters play a part in how the musicians perceive podium acoustics. In comparison between Västerås Concert Hall and Gävle Concert Hall the C and D5 is higher in Gävle Concert Hall. Clarity and Definition have been found to be important when it comes to the perception of music []. For auditorium acoustics, the recommended reverberation time is.-. s. The recommended characteristics for reverberation time show that for low frequencies the reverberation time should not be lower than for middle frequencies []. In Berwaldhallen and Louis de Geer, the reverberation time diagrams show that the reverberation time is lower in low frequencies than in middle frequencies (see Apendix B). In Berwaldhallen this is prominent, and it may be the reason why the musicians marks of Overall Impression is so low. In Louis de Geer, the tendency is not as clear as in Berwaldhallen. This make it difficult to say how it affect the musicians subjective judgements. By comparing the concert halls with lowest rank with the highest ranked halls, some tendencies can be seen. The highest ranked halls have high clarity and definition and short reverberation time. It is not possible to isolate one parameter, all parameters play a role when it comes to acoustic perception. When it comes to ST Gävle Concert Hall has low values while Västerås Concert Hall has high values. This imply that if the ST is not high, this may not be conclusive for the acoustic perception. 5