Interlaboratory comparison of ISO 23351-1: speech level reduction of furniture ensembles Valtteri Hongisto 1 , Jarkko Hakala, and Jukka Keränen Turku University of Applied Sciences, Acoustics Laboratory Joukahaisenkatu 3, FI-20520 Turku, Finland ABSTRACT Mobile enclosures (booths, pods) or partially enclosed furniture ensembles are increasingly used in offices, hospitals, education buildings, traffic centers and other public places to provide spaces for conversations, meetings, or private work. International standard ISO 23351-1:2020 describes a la- boratory method for determining the speech level reduction D S,A of such mobile products. D S,A de- scribes how much the product reduces the A-weighted sound power level of standard speech to the exterior space. The method is based on repeated application of ISO 3741 standard. The purpose of our study was to determine the precision of the ISO 23351-1 for two extreme kinds of mobile products that the method covers: a workstation and a phone booth. Eight laboratories from six countries par- ticipated in the accuracy experiment. Two specimens were circulated in these laboratories and tested according to ISO 23351-1. The measured D S,A values varied from 3.7 to 5.5 for the workstation and from 27.2 to 30.3 dB for the booth. The corresponding reproducibility standard deviations were 0.6 and 1.1 dB, respectively. The results show that the precision of ISO 23351-1 is high and in good agreement with ISO 3741 standard. The results have been published in Annex C of ISO 23351-1 standard. 1. INTRODUCTION

Mobile enclosures (known as booths, pods, or boxes) or partially enclosed furniture ensembles (such as workstations, sofa groups, chairs) are increasingly used in offices, hospitals, education buildings, traffic centers and other public places to provide both ad hoc and planned spaces for private conver- sations, small meetings, or concentration demanding work tasks. The increased popularity of above- mentioned mobile products is based on the fact, that both building owners and tenants prefer to rent open spaces and the number and location of isolated rooms should be customized according to ten- ants’ needs. Small rooms are easier to erect using mobile products than by building fixed rooms. The sound level reduction (or insertion loss) provided by a mobile product is usually tested by ISO 23351-1:2020 standard in laboratory conditions [1]. It is largely based on the method developed by Hongisto et al. [2,3]. The outcome of the test is speech level reduction D S,A . It describes how much the A-weighted overall sound power level of standard effort speech (within 125  8000 Hz) is reduced when the speaker moves inside the mobile product. The acoustic test consists of two sound power level tests ( L W , dB) according to ISO 3741 [4] in reverberation room: speaker alone ( L W,1 ) and speaker inside the tested mobile solution ( L W,2 ). Sound reduction D [dB] provided by the test object in octave bands 125  8000 Hz is determined by

1 valtteri.hongisto@turkuamk.fi

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW

,1 ,2 W W D L L   (1)

The speech level reduction D S,A [dB] is obtained by

, , ,1 , ,2 S A S A S A D L L  

where L S,A,1 [dB] is the A-weighted sound power level of standard effort speech determined by ISO 3382-3 standard [5] and L S,A,2 [dB] is the A-weighted sound power level of standard effort speech reduced by the sound reduction D . The sound reduction of a mobile product depends on many factors, such as sound reduction index of wall, roof, glazing, and door, the insertion loss of ventilation routes, the absorption properties of all interior surfaces, and internal diffusion / geometry. Sound leaks around the element seams are often determining the outcome to a large extent. D S,A takes all these factors into account unlike, e.g., sound reduction index, since no measurements are done inside the mobile product. The only thing that mat- ters is how much sound is radiated outside. Most mobile products are nowadays tested by ISO 23351-1. It is no longer justified to apply sound insulation test method ISO 16283-1 for mobile products because of many factual reasons. The high practical relevance of the single-number quantity D S,A is the reason why the standard has gained large popularity among furniture manufacturers, workplace designers, acousticians, and end users. The measurement uncertainty is important to know for this method because the D S,A values can be quite small for partially enclosed furniture ensembles, even 4 dB or lower [3, 5]. Furthermore, the potential buyer wants to compare the D S,A values of different products. It is important to know, how small differences between product values are meaningful. Decision making should not be confounded by measurement uncertainties. The purpose of this study is to summarize the main results of an accuracy experiment (a.k.a. Round Robin test, Interlaboratory comparison) testing the precision of the ISO 23351-1 standard. The full version of the study is available in Ref. [6]. 2. MATERIALS AND METHODS

The accuracy experiment was conducted according to the general instructions of ISO 5725-1 [7] and ISO 5725-2 [8]. They suggest that the following issues are considered in accuracy experiments:  The experiment is managed by a panel experienced on the test method.  At least two levels of single-number values are tested  At least eight participants join the experiment  All participants are given the same information  Data is treated anonymously  Data is shared to relevant standardization bodies These suggestions were carefully accounted for. The panel consisted of the authors and standardiza- tion working group ISO TC 43 SC 2 WG 35 developing the ISO 23351-1 standard. The test was conducted using a draft standard ISO/DIS 23351-1 (2019). However, the differences between the draft and the final published standard ISO 23351-1 are negligible and the outcomes of this accuracy experiment are valid for the published standard.

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW

The experiment involved the measurement of two products representing the two extreme types of products that can be tested using the standard: A. a workstation (partial enclosure), and B. a phone booth (enclosure). The products are shown in Fig. 1. Eight test laboratories from six countries partic- ipated in the experiment (see Acknowledgements). All laboratories tested both products once. In ad- dition, one laboratory tested both products five times to obtain the repeatability standard deviation (SD). The main outcome of the work is the reproducibility standard deviation, s R [dB], defined by

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW

2 2 R r L s s s   (2)

where s L [dB] is the between-laboratory SD, and s r [dB] is the repeatability SD. The former is based on the values reported by the eight participants. The latter is based on the five repeated tests conducted by a single participant. It is expected that s L > s r , because s r depends only on the variations of repeated measurements in the same laboratory (same room, apparatus, source and measurement positions, and specimen installation) while s L depends on many other things that differ between laboratories (differ- ent operator, apparatus, source and measurement positions, and specimen installation).

Fig. 1. Workstation and phone booth tested in eight laboratories. The external dimensions of the workstation and the phone booth and were 122 x 190 x 140 cm and 100x100x221 cm, respectively. The corresponding weights were 110 and 350 kg. The same products were circulated in the eight laboratories. 3. RESULTS AND DISCUSSION The single-number values obtained in different laboratories are shown in Fig. 2. The essential single- number values describing the outcomes of the interlaboratory comparison are collected in Table 1.

6

32

A

B

5

31

4

30

D S,A [dB]

D S,A [dB]

3

29

2

28

1

27

0

26

1 2 3 4 5 6 7 8

1 2 3 4 5 6 7 8

Laboratory ID

Laboratory ID

Fig. 2. The test results of the eight laboratories for the workstation (A) and for the phone booth (B). Table 1. The main outcomes of the accuracy experiment for the workstation (A) and the phone booth (B): mean (M), repeatability (within-laboratory) standard deviation ( s R ), between-laboratory standard deviation ( s L ) and reproducibility standard deviation ( s R ).

Specimen

Quantity A B M 4.2 28.7 s r 0.09 0.17 s L 0.64 1.12 s R 0.6 1.1

4. DISCUSSION

The main outcome of the study is the reproducibility standard deviation, s R , since it considers both between-laboratory and within-laboratory differences. The s R value was ±0.6 dB for the workstation (A) and within ±1.1 dB for the booth (B). The between-laboratory SD determined the main outcome since the within-laboratory SD was so small. The numbers mean that the between-laboratory values are within 2  s R from each other with a probability of 68.3%. The expanded uncertainty, U, can be calculated by U = k  s R . The constant k is chosen according to desired confidence level. The usual val- ues of k are 1.00, 1.28, 1.65, 1.96, 2.58, or 3.29 corresponding to two-sided confidence intervals of 68.3, 80, 90, 95, 99, and 99.9%. ISO 3741 [4] reports a typical upper bound values of the reproducibility standard deviation for A- weighted sound power level to be 0.5 dB. The results are in good agreement with this. However, the larger value for the phone booth is expected to be caused by slight variations in workmanship which can lead to variation in, e.g., the airtightness of door seams. Since the two tested specimen represented well the two extremes available in the market [2,3,9,10], it is justified to expect that the outcome represents well the interlaboratory variations in general. The

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW

study was successful since the number of participants agreed with ISO 5725 standard, all data was successful, and the two samples represented well the whole range of mobile products in the market. The values of Table 1 were also adopted to the informative Annex C of ISO 23351-1 standard. It would be useful to conduct more similar studies since this is only one study and new kinds of mobile products (furniture ensembles) emerge in the future. 5. ACKNOWLEDGEMENTS

The authors want to thank Framery Ltd. who funded the research work, planned the transportations, funded the transportation of the test specimens, and the assembly and disassembly of the test speci- mens in each test laboratory. Special thanks belong to the seven external participants who put a sev- eral days of voluntary work to this project: Lars Sommer Søndergaard and Rasmus Stahlfest Holck Skov (FORCE Technology, Denmark), Moritz Späh (Fraunhofer Institute for Building Physics, Ger- many), Dag Glebe (RISE Research Institute of Sweden, Sweden), Mikko Tamminen (Framery Ltd., Finland), Remy Wenmaekers (Level Acoustics, The Netherlands), Johan Jernstedt (Akustikverkstan, Sweden), and Elzbieta Nowicka (Building Research Institute, Poland). 6. REFERENCES

1. ISO 23351-1:2020 Acoustics — Measurement of speech level reduction of furniture ensem- bles and enclosures — Part 1: Laboratory method. 2. Hongisto, V., Keränen, J., Virjonen P. (2016). Acoustic testing of office workstations and booths, Proceedings of Internoise 2016, 6226-6231, 21  24 August, Hamburg, Germany. 3. Hongisto, V., Keränen, J., Virjonen, P., Hakala, J. (2016). New method for determining sound reduction of furniture ensembles in laboratory, Acta Acust. Acust. 102 67–79. 4. ISO 3741:2010 Acoustics—Determination of sound power levels and sound energy levels of noise sources using sound pressure – Precision methods for reverberation test rooms. 5. ISO 3382-3:2012 Acoustics — Measurement of room acoustic parameters — Part 3: Open plan offices. 6. Hongisto, V., Keränen, J., Hakala, J. (2020). Accuracy experiment of ISO DIS 23351-1 – speech level reduction of furniture ensembles and enclosures. Appl. Acoust. 164 107249. 7. ISO 5725-1:1994 Accuracy (trueness and precision) of measurement methods and results — Part 1: General principles and definitions. 8. ISO 5725-2:1994 Accuracy (trueness and precision) of measurement methods and results — Part 2: Basic method for the determination of repeatability and reproducibility of a standard measurement method. 9. Hongisto, V., Keränen, J. (2020). Acoustic performance of eleven commercial phone booths according to ISO 23351-1. Research Reports from Turku University of Applied Sciences 51, 20 pp., Turku University of Applied Sciences, Turku, Finland. At: http://julkaisut.tur- kuamk.fi/isbn9789522167743.pdf . 10. Keränen, J., Hongisto, V. (2021). Measurement and Prediction of Speech Level Reduction of a Phone Booth in Three Different Open-plan Offices. Proc. Euronoise 2021, pp. 1194  1202, 25  27 Oct, Madeira, Portugal.

21-24 AUGUST SCOTTISH EVENT CAMPUS. ? O? ? GLASGOW