Effect of presence of furnishings on designed and measured reverberation time in school spaces

Elena Prokofieva 1

Edinburgh Napier University

Unit 1, 7Hills Business Park, 37 Bankhead Crossway South, Edinburgh, EH11 4EP, Scotland, UK

ABSTRACT According to Building Bulletin 93 (BB93), reverberation times for teaching and study spaces are expected to be assessed to comply with the pre-set requirements, when the room is finished, furnished and empty. Presence of furniture and people is considered as variable. Typically, the teaching spaces are modelled by acousticians at the design stage, considering only geometry and any surfaces’ treatments proposed. During site visits, when sample testing is conducted, the spaces are not furnished, however at the compliance testing the classrooms typically have the furniture present. In this paper the author proposes ways of taking into account the furniture into reverberation time calculations, and a single “furniture coupled coefficient” (FCC) is suggested. Recommendations concerning the stage at which FCC should be introduced into the design of new or refurbishment projects are also discussed. Examples of predictions with FCC coefficient and on-site tests are provided. 1. INTRODUCTION

According to information officially released by the Department of Education (published in 2021) there are currently over 30,000 schools functioning in the UK. Of these, over 3,000 are nurseries or early-learning centres, around 21,000 are primary schools, and over 4,000 are secondary schools. The remainder are special schools and pupil referral units [1]. In Scotland there are over 5,000 schools, including 2,630 early learning centres, over 2,000 primary schools, 357 secondary schools and 111 special schools.

The proposals from the government also approve new schools. More than 400 new free schools and technical schools have been approved since 2010 across England, and around 100 schools are proposed to be built or refurbished in Scotland in 2019 [2]. These new schools are a vital part of the government’s plan for education as they increase choice for parents and help to drive up standards across the board. The variety of school types introduce new ideas and approaches to teaching and educating pupils, as well as provide working placements to teaching and supporting staff.

The school buildings’ design includes, among others, input from the architect, the landscape designer, the energy specialist, the mechanical ventilation and electrical engineer, and the acoustic consultant. The cost of the acoustic consultancy typically varies between 0.5% and 1% of the total project design cost.

2. PROJECT DESIGN STAGES AND MAIN STANDARDS FOR SCHOOL

The Plan of Work for design of buildings is published by the Royal Institute of British Architects (RIBA) [3] and, where applicable, by BREEAM [4]. The design stages for a project as set in RIBA document are described in [3] as follows: Stage 0 - Strategic definition; Stage 1 - Preparation and briefing; Stage 2 - Concept design; Stage 3 - Spatial coordination; Stage 4 - Technical design; Stage 5 - Manufacturing and construction; Stage 6 – Handover; Stage 7 - Use.

Input from the acoustic specialist is required in Stages 2 to 4, then some involvement is possible at Stage 5, and additional input at Stage 6, when commissioning testing is required. There is no

1 E.prokofieva@napier.ac.uk

requirement to have the same acoustic specialist involved in design and in commissioning testing, however for the acoustician it is a great opportunity to verify their own design, when testing the completed building.

The main design document for the acoustic design of educational buildings is Building Bulletin: BB93, updated in February 2015 [5]. Additional guidance is also provided by Association of Noise consultants in “Acoustics of Schools - a design guide November 2015” [6]. These two documents cover all main aspects of the acoustic design required for any type of educational buildings: from the nursery or early years centre to university campuses.

The standard acoustical design approach for a new or refurbished building is to review and provide advice on the key aspects related to the sound and noise within and around the building: control of external sound break-in, sound insulation between spaces, control of reverberation time, and control of building services and plant sound and vibrations, both airborne and duct-borne.

3. REVERBERATION TIME REQUIREMENTS SET IN BB93

Reverberation time (RT) is the duration it takes a sound to decay by 60 dB within a room. The speech conditions require controlled reverberant conditions, whilst musical performances sound better in more reverberant conditions, depending on the musical style. A good level of speech intelligibility is also desirable in all teaching and learning areas such that a single un-amplified speaker can clearly communicate to a teaching group. The speech intelligibility is affected mainly by two factors: ambient noise and reverberation time. BB93:2015 [4] outlines the reverberation time design criteria for each considered space (as given in Table 2.1 in the document and reproduced in Table 1 below). The values vary for new build and refurbishment school projects.

The reverberation time in above Table 1 is quoted in terms of T mf which is the arithmetic average of the reverberation times in the 500 Hz, 1 kHz and 2 kHz octave bands, or the arithmetic average of the reverberation times in the one-third octave bands from 400 Hz to 2.5 kHz (although these are not mathematically equivalent, in practice the difference will be small and in the interests of simplicity and ease of measurement, either is acceptable) [4].

Table 1: BB93: Reverberation time requirements

Recommended Reverb. Times, T mf (s)

Type of room

New build Refurb

Nursery school rooms (Early years) Primary school: teaching and learning spaces Secondary music practice rooms (small volume)

≤0.6 ≤0.8

Secondary school: teaching and learning spaces, Art rooms ≤0.8 ≤1.0

ASN classrooms ≤0.4 across 125 Hz to 4 kHz, no more

than 0.6 at each frequency

4. THE IMPORTANCE OF REVERBERATION TIME IN ACOUSTIC DESIGN

The reverberation time in all main spaces within the proposed school building is part of the design, but typically it does not include any thorough study and/or modelling using acoustic modelling software, due to shortage of time for the design process and lack of budget to cover the modelling. The modelling only occurs in large-scale projects with complex-shaped rooms in which the acoustics are expected to play a significant role (i.e. recording studios, performance theatres, broadcast rooms and similar). However, all other rooms designed for teaching (classrooms), for learning (breakout spaces, library) and for ancillary use (offices, meeting rooms, etc) also require input from the

acoustician on the sound insulation and on the reverberation time. This input, therefore, is usually limited and provided only for an exemplar room type, which is defined by the floor, walls and ceiling material coverage and other unique features within the rooms.

In the guidance provided by Association of Noise Consultants it is noted that a classroom with a long reverberation time shows degradation of the speech intelligibility. Long reverberation times can occur in large rooms, or in rooms with reflective wall and ceiling surfaces [5].

5. SABINE AND EYRING FORMULAS FOR PREDICTION

Sabine's formula for the reverberation time prediction is straightforward, and therefore very widely applied, however it is only recommended to be used for the fairly 'live' rooms, where there is only a small amount of absorption (with α AVGE lower than 0.2). The Sabine formula is applied with the main assumption that the sound within the room is diffuse and is distributed evenly in all directions. A secondary school classroom falls into this category, whilst a primary classroom may or may not be considered ‘live’.

The Sabine formula is shown as Equation 1:

RT = 0.16V/[(S*  ave )] (1),

where V – volume, S – surface area,  ave – average absorption coefficient, which depends on frequency and size.

The Eyring formula is recommended to use for 'dead' rooms, i.e. rooms containing a lot of absorption. The Eyring formula (and its further iterations) is applied in rooms where the sound field may not be diffuse, i.e. distributes in one direction more than in the others.

Classrooms for primary schools, nurseries and teaching spaces for children with special needs are rooms with additional absorption and here it is recommended to use Eyring formula instead of Sabine’s.

The generic Eyring formula is shown in Equation 2:

RT = 0.16 V / [k*V – S*ln (1 - α ave )] (2),

where V – volume, S – surface area,  ave – average absorption coefficient, which depends on frequency and size, k – air coefficient, which depends on humidity and frequency.

Absorption of air is generally accounted for in large spaces, as the addition from this coefficient in smaller rooms is considered to be insignificant, and adds the value only for larger rooms.

Both these formulas, Sabine and Eyring, provide very close prediction results for rooms with the average absorption coefficient below 0.2. Both formulas are considering all surfaces within the room on the grounds of their absorption value and their dimensions. Neither of these formulas can recognise the presence of the prominent room modes (standing waves), nor the unevenness of spread of absorptive/reflective surfaces throughout the room.

The presence of people and furniture in classrooms will also influence the absorption, and therefore, on the reverberation in the room.

The absorption coefficients for the various materials across the frequencies of interest are provided in scientific literature (i.e. in [6] and [7]), or supplied by the manufacturer and confirmed by laboratory test certification. Some of the absorption coefficients are shown in Table 2 below.

Table 2: Examples of absorption coefficients for various materials

Frequency

Material

125 250 500 1000 2000 4000

Vinyl on concrete 0.02 0.02 0.03 0.05 0.05 0.05 Carpet on concrete 0.12 0.15 0.18 0.28 0.28 0.4 Plasterboard on dabs 0.15 0.10 0.06 0.05 0.05 0.05 Exposed concrete 0.01 0.01 0.02 0.02 0.02 0.05 Class A panel (generic) 0.25 0.75 0.90 1.00 1.00 0.80 Soft furniture 0.18 0.35 0.54 0.54 0.54 0.64 Hard furniture 0.02 0.02 0.17 0.14 0.32 0.32 6. EFFECT OF FURNITURE ON REVERBERATION TIME

The reverberation time values given in BB93 in [4] are for rooms that are finished, furnished for normal use, but unoccupied. Where rooms are to be used without furnishings, the performance standards normally apply in the empty condition. Normal furnishing is not anticipated to have any significant effect on indoor ambient noise levels or sound insulation, but may reduce measured reverberation times by providing diffusion and absorption.

The standard classroom in the secondary school has vinyl flooring (rarely with thin pre-bonded carpet on concrete), suspended ceiling with Class A acoustic tiles, plasterboard-covered walls, at least one large window, and one door. The classroom also has furniture in it, such as chairs, school desks and benches, lined in rows, providing a typical layout in any secondary school teaching space.

The primary classroom is usually designed with more variability in choices for wall and floor materials, as well as furniture types and arrangement. The floors are mainly vinyl for easier cleaning purposes, suspended ceiling with Class A acoustic tiles (in some cases – exposed concrete or cross- laminated timber), plasterboard-covered walls, and minimum one large window and a door. The primary classrooms and nursery rooms may have additional rugs on the floor, soft furniture, sheltering booths for playing, reading and quiet times, as well as chairs, tables, randomly arranged across the floor area, rather than lined in rows.

The effect of the furniture present in classrooms is typically not taken into account during the calculations for RT prediction. It is noted that hard furniture in rows is not expected to provide significant effect on absorption, but can be beneficial for scattering sound and creating the diffuse field ([5], [7]). The soft furniture, arranged randomly within the space, is expected to provide some absorption in addition to scattering, and also affect the diffuse fields’ re-distribution of sound within the room.

During site visits and tests it had been noted that the desks and benches may occupy part of the soft flooring and reduce the overall absorptivity of the room, and therefore increase the measured reverberation time in this room.

Room modelling using software (as shown, i.e. in [7], [8]) allows to include the proposed furniture layout, however still provides an approximate answer, with more than 20% deviation from the measured reverberation time results at the key frequencies (500 Hz, 1 kHz and 2 kHz)([7]).

7. COMPLIANCE TESTING AND COMPARISON WITH PREDICTIONS

Compliance testing for school projects is conducted at the handover stage (RIBA Stage 6), and often it is the only chance for acoustic consultants to verify the proposed design in-situ. In rare occasions during the acoustic site visit during the construction stage (RIBA Stage 5), preliminary sound tests are conducted, but they are always conducted without furniture and even in unfinished spaces, so therefore could produce higher figures than the requirements set in BB93 for completed classrooms.

7.1. Secondary school testing example

A secondary school classroom for music practice (BB93 requirement for RT = 0.6sec) was tested during the construction and at the completion testing stage. The size of the room in this example is small, with overall volume around 30m 3 . The ceiling was concrete with Class A absorption rafts hanging from the ceiling. The floor covering was vinyl, and the walls were finished with plasterboard. There was no window in the tested room. At the first test, there was no furniture in the room, at the second test there were hard desks and chairs.

The summary of the tests is shown in Figure 1.

0.80

0.60

0.40

0.20

Measured RT classroom with furniture, RT = 0.57 sec Pass

Measured RT classroom without furniture, RT = 0.62 sec Fail

0.00

125 250 500 1k 2k 4k

Figure 1: Measurements of RT in classroom with and without furniture

The results show that the reverberation time in the classroom tested without furniture has failed to comply with requirements set in BB93, however with furniture, tested at the later date, the results complied with BB93. The variation between the measurements at key frequencies (500 Hz, 1kHz, and 2kHz) was varying between 5 and 11%.

Comparison of predicted reverberation times with interim measurements without furniture, and are shown in Figure 2.

0.80

0.70

0.60

0.50

0.40

0.30

0.20

0.10

0.00

125 250 500 1k 2k 4k

RT classroom, no furniture predicted, no furniture

Figure 2: Measurements of RT in class room and comparison with predicted values (no furniture).

It can be noted that the predicted reverberation time at key frequencies (500 Hz, 1kHz, and 2kHz) is exceeding by 6% at 500 Hz, 7% at 1kHz and 5% at 2kHz over the measured during the construction stage. However, as it had been mentioned above, this is a very rare occasion of comparison of the predicted and the measured value, and quite impractical for the completion testing, as it does not have any presence of furniture.

If the prediction is conducted with the addition of a single coefficient for the furniture presence (hard furniture for a secondary classroom), then the predicted reverberation time is coming closer to

the measured reverberation with the furniture present. The furniture is covering part of the carpet, and therefore the reduction of the carpet area was also introduced.

The comparison of measured and predicted reverberation times with the furniture present in the room is shown in Figure 3.

0.70

0.60

0.50

0.40

0.30

RT classroom with furniture predicted, with furniture

0.20

125 250 500 1k 2k 4k

Figure 3. Measurements of RT in classroom and comparison with predicted values (with furniture)

It is shown in Figure 2 that the predicted reverberation time at key frequencies (500 Hz, 1 kHz, and 2 kHz) is now much closer with deviation of no more than 5% at all key frequencies, comparing with measured at the compliance stage. It is also noted that the prediction underestimates the room performance at some frequencies.

7.2. Primary school example

A primary school room was tested at the completion stage and had the furniture present. There was no opportunity to measure without the furniture at the day of the testing. The room was large, with a volume over 100m 3 . The ceiling had Class A ceiling tiles installed, vinyl on the floor, and walls finished with plasterboard. There were large windows on one of the walls. Soft furniture, rugs and hard desks were present in the room.

The results of the on-site reverberation time measurements and the comparison with the predicted reverberation are shown in Figure 4.

1.50

1.00

0.50

Predicted, with furniture

measured with furniture

0.00

125 250 500 1k 2k 4k

Figure 4. Measurements of RT in large (more than 100m 3 ) room and comparison with predicted

values (with furniture)

It is shown in Figure 4 that the predicted reverberation time at key frequencies (500 Hz, 1kHz, and 2kHz) is close to measured: by 2% at 500 Hz, -1% at 1kHz and 6% at 2kHz. It is also noted that the prediction underestimated the room performance at one of the key frequencies.

8. DISCUSSION AND FURTHER WORK

The furniture effect coefficients used in the predictions are estimations, based on the figures provided in literature [6, 7] for hard (secondary school example) and soft (primary school example) types of chairs and tables, and are not confirmed with laboratory testing, or detailed research.

As in the simple formulas, such as Sabine and Eyring RT prediction formulas, it is not possible to connect the effect of the furniture with relation to its location (i.e. in the middle of the room, or in the corner, on carpet or on vinyl, etc). However, the locations of the elements of furniture may affect the measured value of the reverberation time. It is proposed to investigate the generic single “furniture coupled coefficient” (FCC), which will provide the absorption spectrum of the coupled “furniture” and “surface”, on which the furniture is placed, to be measured and used in Sabine and Eyring formulas for predictions.

The FCC, once determined for the different types of furniture and floor coverings, can be used at the design stage, allowing the prediction to be more precise in relation to the expected real measurements at the compliance stage and reducing the level of uncertainty through the design project. It can also be useful for refurbishment projects, especially when the use of room is expected to change, the floor coverings and the furniture are replaced or upgraded, or the existing room is not compliant with the current school standard requirements. With use of FCC in such predictions, the change/ addition of furniture could also be offered additionally to the wall panelling to improve the sound absorption.

Establishing this coefficient across various key frequencies could be taken forward as an independent research project.

ACKNOWLEDGEMENTS

I would like to thank my colleagues at Robin Mackenzie Partnership for providing the site testing data and invaluable advice for this work.

REFERENCES

1. Department for Education website (2021) 2. RIBA Plan of Work, (2020) 3. BREEAM Education. BRE Global (revision 2010) 4. Acoustic design of schools: performance standards. Building Bulletin 93. Department for Education (updated February, 2015) 5. Acoustics of Schools - a design guide. Published jointly by the Institute of Acoustics (IOA) and the Association of Noise Consultants (ANC). November 2015, ANC 6. J. Borwick. Loudspeaker and Headphone handbook. Pages: 339-341, 2001 7. F. Fantozzi, M. Rocca, N. Spinelli. Assessment of Reverberation Times in University Classroom: Comparison between Analytic Formulae, Software Simulations and Measurements. Proceedings of the 16th IBPSA Conference, Rome, Italy, Sept. 2-4, 2019 8. A. P. O. Carvalho. The Use of The Sabine and Eyring Reverberation Time Equations to Churches. 129th meeting of the Acoustical Society of America, Washington DC, May 30-June 03, 1995.