Global noise score indicator for classroom evaluation of acoustic performances in LIFE GIOCONDA project

Abstract The LIFE GIOCONDA is an ongoing project that aims to provide an innovative methodology to the authorities for supporting the environment & health policies by involving the young people in the decision-making processes. The project suggests a web platform able to relate air and noise pollution data in the schools with the students’ pollution awareness. GIOCONDA aims to enhance the awareness of students, teachers and local administrations on the noise issues in schools, presenting suitable tools to improve the public participative processes. This paper presents a new method that has been developed within the Project. It aims to evaluate the acoustic performances of a classroom and to suggest the use of a global indicator based on a group of acoustic parameters compared with their limit values. Whit the new method the comparison between different classrooms or different schools becomes possible, together with a homogeneous evaluation of the priority for planning noise mitigation actions. Several noise measurement campaigns have been performed to characterize the students’ exposure in eight Italian schools. The results are useful to describe the acoustic performances of classrooms.


Introduction
The exposure to high noise levels is a well know issue in modern society, so much that the scientific community 30 and the consultant bodies gave a lot of attention to it in the past decades. The problem is much more complicated inside the school environment, where the children and teachers need quietness for their activities [1].
Impairs cognitive performance in schoolchildren [2- 35 7], as well as stress or voice problems for teachers [8,9] can be caused by both the noise coming from outside or inside the classrooms. Indeed, several studies shown that high reverberation and high levels of background noise, due to external or internal noise, can seriously affect speech per- 40 ception, short-term memory, task assignments and general understanding, resulting in a significant negative impact upon school and working performance and an induced hearing threshold shifts in the worst cases.
Therefore, the acoustical conditions in classrooms do 45 not often fit the specific needs of young listeners.
Minimizing the noise and reverberation in classrooms thus become of relevant importance [7].
In this background, the LIFE GIOCONDA Project [10]  The project has the purpose of filling the gap between young people and public administrations about environ-10 ment and health issues, considering that young people will lead the environment and the health of tomorrow by means of their perception and behaviour. This is a chance to understand the gap between subjective perception and real pollution. 15 Indeed, providing accessible information for public decisions on environment and health is a priority of the current LIFE+ Environment Policy and Governance, supported also by the World Health Organization, which studied the effects of noise pollution on vulnerable cate-20 gories [11,12].
The Aarhus Convention [13] establishes several rights concerning the environment, among which the right to the "access to environmental information" and the right to the "public participation in environmental decision-making". 25 The first one deals with the information about both the state of the environment, human health and safety, and the policies or measures taken to improve them. The second one drives the public administrations to involve population planning and programmes related to the environ- 30

ment.
After the Aarhus convention, several projects are born to enhance the information exchange between people and public administration. The GIOCONDA Project provides an innovative procedure to effectively support the young peo- 35 ple involvement in the decision-making processes on environment and health. The procedure is based on the combination of air and noise pollution measurement results with the risk perception and willingness-to-pay (WTP) [14] related to environmental health issues. Since the project 40 involves cities with different kind of pollution sources, a holistic approach [15] has been applied. Alongside the measurement campaigns, the students reported their risk perception on noise, air, waste and water pollution and also provide their WTP related to each issue. The combina-45 tion of these data could allow the understanding of the gap between the perceived and the objective pollution. Some previous studies already investigated this gap [16][17][18][19][20] but this project aims to offer a further step in communicating the results to the surveyed students and to increase their 50 awareness about the negative effect of a noisy environment on the learning process. Finally, an online platform will be developed in order to facilitate the application of environmental and health risk governance and policies.. The platform will include tools for the decision makers that may 55 be useful to estimate the costs and benefits of policies regarding the air pollution and/or the noise exposure, while other tools will enable schools to measure the students' perception of their surrounding environment. As a further result, in the second year of implementation of the project, 60 it will be possible for students to discuss and use the platform in order to suggest solutions to the public administrations.
In the international scientific literature many studies involved the effects of noise on teachers and students 65 or the characterization of classrooms' acoustical performance [21,21,22]. Unfortunately, for the authors' knowledge, none suggested a method to properly evaluate a classroom and to allow their comparison.
In the present paper, an innovative method for the 70 noise characterization of classrooms is presented. It is based on the noise data acquired during measurements campaigns performed in 2015 along the Italy and it develops a new global indicator that would allow a simpler evaluation of the state of the acoustic environment inside the 75 schools. This innovative method has been applied in some pilot schools in order to test the reliability and the suitability for its application in Italy and Europe.

Schools involved in the project
The GIOCONDA Project involves eight schools belonging 80 to four different project areas: Napoli, Taranto, Ravenna and Valdarno. In each project area, two schools have been chosen: -a first grade secondary school, age from 11 to 13 (middle school MS);

85
-a second grade secondary school, age from 14 to 18 (High School HS).
The four areas are spread across all Italy: -Valdarno is a portion of a rural area in Tuscany. The schools are in two small villages, San Miniato and 90 Ponte a Egola. The San Miniato HS is in a very silent context, with few roads and a low traffic flow. The Ponte a Egola MS is along a regional urban link road, presenting a high traffic flow at low speed. -Ravenna is a medium city, with the HS placed in the 95 city centre, in a traffic-restricted area. The MS is in a quiet suburb placed near the sea.  -Napoli is one of the biggest cities of south of Italy and both the schools are placed in the city centre, even if they are placed in different districts: the HS is in the historical centre and the MS is on a main road in the district of the railway station.

5
-Taranto is a big industrial city. The HS is located in the city centre, whilst the MS is located in the industrial area. High traffic levels are present around both the schools.
In Figure 1 the schools and theirs surrounding are 10 shown.
A summary of context characteristics, main noise sources and qualitative judgements about both status of school buildings maintenance and noise pollution of the area, according to a first subjective judgment of the opera-15 tor during the inspection, is reported in Table 1.
In each school, three classrooms have been characterized, evaluating both the noise exposure and the building acoustic characteristics. iPOOL, a spin-off company of National Research Council of Italy, has been put in charge 20 for developing the whole procedure to evaluate the acoustic performances of the classrooms and for carrying out all measurement sessions. The investigated classrooms are described in Table 2.

Procedure to evaluate the acoustic performances
The following main steps are proposed in order to set-up a procedure to acoustically evaluate any classroom with a single indicator representing the judgment of the overall 30 noise situation: 1. setting a list of significant acoustic parameters to investigate; 2. establishing a score range for each parameter; 3. establishing a "Global Noise Score" to be assigned 35 to the classroom, with a related score range; 4. carrying out the measurement campaigns; 5. analysing the data and providing the results; The quality and intelligibility of speech in a classroom mainly depends on both the noise level and the amount 40 of reflected sound, which increase the noise level and masks the speech itself. Thus, the noise and the reverber-   ation outline the acoustical environment of a classroom. Concerning the noise, outside the school it is mainly due to transport infrastructures and industrial areas, whilst inside the classroom it is also related to other sources, such as building services (heating, lighting, ventilation 5 systems), teaching aids (overhead projector, computers) or the ongoing lesson. Reverberation describes the amount of reflected sound and it depends on the room volume and the acoustic characteristics of all the surfaces inside the room, as walls, ceiling, floor, desks and whiteboards. 10 Bearing in mind these considerations, a common set of six parameters, defined in accordance with international standards, is proposed: -the L DAY for investigating the exposure to external sources, calculated from: 15 1. external noise monitoring (L DAY−Ext ); 2. internal short-term measurements (L DAY−Int ); -the following four parameters for investigating the building acoustics characteristics: Each parameter has been categorized in five classes: the higher one (score 5) has been chosen to fulfil the Italian 25 limit values [27][28][29][30], which are very hard to be observed, the others have been set according to the scientific literature or to the international optimal values [31][32][33][34][35]. The score ranges proposed are reported in Table 3 and better explained in the following. 30 In Italy, the national regulation for environmental acoustic [19] requires the school areas to be in the lower class of acoustic zooning (L DAY−Ext < 50 dB(A)), considering the exposed façade as representative for the whole school. Nonetheless, the schools are often placed near 35 congested streets, as documented in several cases analysed in this project (Table 2)  The limits for façade insulation parameter D 2m,nT,w and for wall insulation parameter R' of school buildings are defined in a specific decree [28]: D 2m,nT,w must be higher than 48 dB(A) and R' higher than 50 dB(A). The na-45 tional technical standard [32] sets a very low requirement of 38 dB(A) for D 2m,nT,w, that the authors used as base for the very poor score level.
In proper decrees, the Italian regulation defines the optimal values as a function of both frequency and volume 50 of the room [28] and the limits of 1.2 s [27, 30] for reverberation time in classrooms. However 1.2 s is a value considered too high by the scientific community [31,33,37], thus the higher class has been set TR=0.8 s and the sufficient class has been considered equal to the law limit.

55
The STI index does not have a proper limit in the national legislation, however a scale of values is proposed according to the international standard [26]. The speech intelligibility is related to the signal to noise ratio (S/N), which is the difference between the speech and back-60 ground noise in a room.
A qualitative judgment is assigned to each class in addition to the numerical score in order to enhance the students' understanding.
All the scores can be summed up to obtain the Global 65 Noise Score of each classroom, reported in Table 6.
The total values range from 6, corresponding to the simultaneous minimum score for all the parameters, to 30 when each parameter has the maximum score.
A single indicator could allow the students to compare their classroom or school with other ones. Moreover, 5 the simple metrics is easily understandable even for young people: the highest is the "Global Noise Score", the best is the environment of the classroom.
For each classroom, the "Global Noise Score" is obtained summing up the score of all parameters. So, the best 10 is the acoustic situation in the classroom, the highest is the Global Noise Score, ranging from 6 to 30.

Measurements protocol
Both the external and internal noise levels influence the acoustic comfort in classrooms: with high noise the stu-15 dent pay less attention and the teachers have to raise their voice above the background noise. A detailed analysis of both noise levels is then necessary.
No railways or airports are located in the proximity of the studied schools and the main noise source is al- 20 ways the road traffic. However, in Napoli also the anthropic noise from markets and public spaces strongly affects the acoustic comfort of the classrooms, whilst in Taranto also some industrial sources are close to the schools.
The external noise level is acquired through a weekly 25 lasting measurement, with the microphone placed at 1 m from the main façade and at 4 m height [36]. The internal noise level is obtained with short-term measurements lasting 30 minute with open windows, taking the average values from two different positions: centre of the room (stu-30 dent average listening position) and 1 m from the windows (student worst listening position). All short measurements were performed in the afternoon, without the students.
In Italy the open windows is a common condition due to the warm average climate and to a general classroom over- 35 crowding. Then, according to the guidelines of the Tuscany Regional Environmental Agency [38], a correction factor can be applied to the internal short term noise level to evaluate the average daytime noise level inside the classroom (L DAY−Int ). The procedure assumes that the internal and ex- 40 ternal noise levels have the same time-evolution. Furthermore, from a single long term external measurement for each school and an internal long term measurement calculated in each class, some calculations are needed to obtain the daily noise level outside (L DAY−Ext ) the 45 classrooms that are not facing the external measurement. An attenuation term "A" is calculated between the outside measurement and the internal one close to it. This atten-uation is considered to be the same for the outside wall, thus the external noise levels where a proper measurement 50 were not performed have been calculated by adding A to the corresponding internal noise levels After the characterization of the environmental noise, the façade insulation have to be tested. Indeed, the interior sound quality is due to the external sources and by the 55 building characteristics [39,40], therefore the standardized level difference D 2m,nT parameter is monitored using a pink noise generated by a dodecahedron emitter. This measurement allows the evaluation of the room insulation from the external noises.

60
Similar measurements are carried out to evaluate the insulation between the adjacent rooms and the corridors through the R parameter.
Finally, the acoustic characteristics of the rooms are evaluated: reverberation time and STI index were mea-65 sured through the MLS signal.

Exposure to external sources:
The external noise exposure levels daytime are reported in Table 5 together with the external levels in front of each 70 classroom.
Due to various security constraints the external noise acquisition in S5 and S6 is not performed with a longterm external monitoring station, as in the other sites. Anyway, it was possible to evaluate the time evolution of noise 75 and the equivalent noise level during the daytime through some short-term measurements in the morning, according to [38].
In Table 5 also internal levels with opened windows are reported, together with the classrooms position.

80
The environment surrounding the main façade is critical for all schools such as most cases, the measured levels exceed the local regulation limits. However, internal and external values show wide variations, depending on the location of the classroom respect to the main road. 85 It is relevant that lessons result seriously affected by noise in the classrooms of S1, S2, S5, S6, S7 and S8 schools, when windows are open. This clearly affects the school activities and the student's wellbeing, especially during the springtime.

90
In Figure 2 the judgement of classrooms in terms of both external and internal levels is reported, highlighting how even when a school is in a noisy context, the class- rooms placed in a backside position can be exposed to low noise from external main sources.

Acoustic insulation of the vertical partitions
The measured D 2m,nT,w and R' values in all the schools are 5 reported in Table 6. The walls insulation of façade presents D 2m,nT,w values below the law limits, whereas insulation between the classrooms present R'w values slightly better. R'w is higher between the rooms than between rooms and corridors. 10 In some school (e.g. S1, S4) the façade insulation values are quite different among the three classrooms. The variability in S1 is mainly due to the conditions of doors and windows fixtures that show generally very bad conditions and in some case they were even broken. 15 In S1, S2, S3, S4, S5, S6 partition insulation values are quite different in the three classrooms and this variability is mainly due to different construction technology: in some  case partition is a load-bearing, otherwise is a false partition walls. 20 The rooms with no adjacent classrooms are less disturbed by adjacent lessons but at the same time they suffer more the noise coming from the corridor and stairs.
In Figure 3 judgement of classrooms in terms of external and internal insulations are reported. 25

Acoustical characteristics of classrooms
The parameters used in the study to describe the acoustical quality of classroom are the reverberation time (RT) and the intelligibility index STI [41,42]. Table 7 reports the measured values. 30 Most of the classrooms have very high reverberation time values (>1.50 s) due to the big volume, with height greater than 4 meters. Some rooms (J, L, P in Table 5) present better values thanks to a false porous ceiling installed.   In Figure 4 judgements in terms of architectural acoustical parameters based on values in Table 5 without taking into account students' presence, are reported. However, people inside the room significantly affect values and further analyses are presented in the following.

Global noise score
In Figure 5, all the "Global Noise Scores" of the classrooms involved in the project are reported. A chromatic scale is used in order to quickly evaluate the acoustical condition of the classroom. 10 As foreseen, the most polluted classrooms are placed on the main road ( Table 2). As an example, the rating summary for S2, where classroom F is the only one exposed on the main street, is reported in Table 7.
The use of these scores could help the students to 15 identify the problems in their school and to encourage the local administrations to solve them. The project also intends to stimulate students to take actions whenever their own behaviour could improve their environment. Therefore, specific advices could be given after the project to im-20 prove the specific noise problems.

Simulation of classrooms' RT with "full-room" condition
The acoustical performance of a room, described by means of the RT and STI index, depends also on the absorption 25 coefficient of all surfaces. Presence of people, operating like a diffusive and absorptive media, influences the room absorption coefficient. Thus, even though the Italian regulation takes into account only the acoustic parameters   computed in empty-room condition, the full-room condition has been also analysed in this project, as a pilot study using a data modelling, based on the acoustic absorption areas and coefficients reported in Table 9.
The presence of students does not exercise any in-5 fluence on the measured noise level outside and inside the room, whilst it strongly influences the Reverberation Time [43] and consequently the STI index and to a lesser extent the D 2m,nT,w and R'w. Thus, the full-room condition has been simulated for the S2. 10 First of all, the acoustic absorption area in case of empty-room condition has been estimated from Sabine formula using the measured values of volume and Reverberation Time. Then, the full-room acoustic absorption area is estimated summing people absorption, using val-15 ues reported in Table 9 and considering one student per each desk in the classroom. The new reverberation time and all other indexes have been obtained applying the Sabine formula, as reported in Table 10.
The acoustical comfort is better in the simulated full-20 room condition, thanks to the absorption due to the presence of students. The RT decreases of about 35% and STI increases of nearly the 15%, without considering noise due from students inside classroom. This analysis shows that before planning actions for improving the acoustical com-25 fort in a school, it could be worthy to evaluate also the fullroom condition, in order to obtain a realistic priority order among classrooms.

Comments and future developments
The main source of noise around the schools under study 30 is the road traffic. According to the perceptual evaluation,   the perception of road traffic noise by the students is confirmed.
The insulation and absorption are critical in all the schools: the main observed problems are the low façade insulation and the high reverberation time.

5
As a consequence the D 2m,nT,w resulted the worst parameter by far, being very poor in 24 classrooms out of the 25. The window fixtures, which always present a bad maintenance or are lacking of insulated glazing, are the main cause of low façade insulation in most of the cases. Very 10 high reverberation is the second main problem observed.
A reverberation time higher than 1.5 s is spotted in 80% of classrooms, whilst the suitable values for correct speech intelligibility are close to 1.0, caused by the big volumes and the lack of false ceiling in several classrooms. 15 Only 3 out of 24 classrooms have a false ceiling installed.
Unfortunately, most of the examined schools show bad results for several parameters, as reported in Figure 6. This is consistent with many other studies around the world, showing that school characteristics are often below 20 requirements.
GIOCONDA is carried out combining two monitoring systems: one based on the collection of data on air and noise pollution in GIOCONDA's sites, the other based on risk perception of teenagers and their willingness-to-pay 25 (WTP), collected with questionnaires. The environmental monitoring of both air pollution and noise, indoor and outdoor, provided a noise characterization of each school. The students produced "conceptual maps", "trees of problems and solutions", completed a questionnaire on risk perception, produced recommendations, they did inter-5 views and peer-to-peer discussions.
The questionnaire on risk perception, which was part of a larger questionnaire, was modelled on the psychometric paradigm [44], assessing several important qualitative characteristics of the risks associated with environmental 10 pollutants.
The specific results of risk perception questionnaires regarding noise will be object of a further work when all questionnaire results will be available. Anyway, preliminary results show some correspondence between noise 15 measured and noise perceived by students.
Then, the results of acoustic measurements will be presented to the students in order to improve the awareness of pollution and the perception of health risks. These outcomes could encourage the use of the GIOCONDA plat-20 form to share environmental data, opinions and ideas, in order to improve the future environment.

Solid noise method
The general poor conditions found for classrooms' interior with windows and doors badly maintained, or false walls 25 with holes, lead to the need of identifying which are the most critical elements. A new device developed by Noiselab [45] has been tested in order to evaluate the propagation of vibrations in a classroom's surface due to a noise source in an adjacent room. Noise is transmitted to the re-30 ceiver room by various architectonic elements, thus a measure of their vibration allows an assessment of their contribution to the overall noise transmission [46]. This method consists in placing an omnidirectional noise source in an adjacent classroom and measuring the 35 vibrations of each surface in the receiver room with an accelerometer. A measuring points' grid has been established to evaluate every possible room elements and 6 seconds measurements were taken on each chosen point. Particular care has been given to characterize every possi- 40 ble architectural elements of the room, whilst keeping the number of points manageable in a reasonable time span.
A frequency analysis has been performed on each measurement point, obtaining 1/3 octave band levels for the acceleration signals. For each frequency the band 45 maximum, minimum and average values have been computed and used to obtain a colour mapping. This mapping method aims to obtain a graphical overview of the measurement campaign that may allow the identification of the weak points in the noise propagation in a visual way 50 understandable also by the inexpert students. Each measurement point has been then associated to a colour value and the overall result have been superimposed to a room picture. Figure 7 shows results for 50 Hz, 500 Hz, 1000 Hz and 2000 Hz centre band value. Red areas are associated 55 to higher signal and green to lower values.
The noise source has been located in the adjacent room, separated by the wall on the right of Figure 7.
The preliminary results of this innovative method highlight that different frequencies show different be-60 haviour of the room elements. The windows, as an example, contribute significantly to the noise transmission only at low frequencies, while at the higher ones the ceiling seems to be the strongest source. The partition wall between source and receiver room shows an inhomogeneous 65 pattern of vibration above a threshold noise frequency. The central part of the wall is less vibrating due to a large corkboard on its surface, effectively acting as a sound absorber.
This kind of analysis appears to be very useful to vi-70 sualize the critical path of sound propagation, allowing to directly act on the specific issues in the noise mitigation phase.

Conclusions
Noise in schools is a serious issue for the young vulnerable 75 children and teachers, thus a good environment for learning and working is mandatory for improving the education quality. However, the learning environment in a school is the complex result of relationships between acoustical and psycho-acoustical, such as the internal and external 80 noise as well as the attitudes of the schools' users. Therefore, in order to improve the schools' environment and facilitate a better learning in a wider perspective, the identification of problems by means of the only physical traditional approach could be not enough. At this purpose, the 85 GIOCONDA project is a pilot study born to create a useful tool for schools and local authorities to highlight, discuss and solve the noise issues, and above all to raise in the citizens the awareness of improving the school environment quality. The set-up of a web platform for sharing data and 90 ideas is a practical goal of the project, to connect data of air and noise pollution in schools with the students' pollution awareness, contributing to understand the gap between perception and objective pollution. This paper showed the results of the noise measurements campaigns performed to acoustically evaluate several classrooms in the schools involved. A new index, named "Global Noise Score", has been proposed in order to better understand the acoustic performance of a 5 classroom and to allow their comparison. The index is obtained by summing up each single score correspondent to a set of parameters: internal and external daily noise levels, façade and wall insulations, reverberation time and speech intelligibility (STI). 10 The results showed that most of the parameters do not comply with the law requirements for schools. The computed Global Noise Score confirmed how noise in schools is a serious issue that should not be neglected by local administrations. Indeed, only 4 classrooms over the 24 anal- 15 ysed report a Global Noise Score at least sufficient.
The most serious issue related to the learning process is confirmed to be the high reverberation time, thus further analyses on it are necessary to find a solution that really comply with the structural problems of each classrooms. 20 The solid noise method has been found to absolve this purpose. Indeed, when the problem are localised, it came out that easy and low-cost actions could be often more efficient than costly mitigation on the external noise sources. Moreover, in some cases the greatest part of noise pollu-25 tion in classrooms is due to internal sources (machineries, students). Thus, a simple good maintenance of the windows fixture and doors could improve the insulation, while the reverberation time could be reduced increasing the amount of acoustic absorption in the rooms by means 30 of false ceilings and wall coverings.
The data provided by this study will be the first contribute to the GIOCANDA web platform and the Global Noise Score presented can be extended to other schools contributing to spread the awareness of environmental 35 health issues.