Acoustic evidence
19Expert evidence on acoustic issues was provided on behalf of Amazonia by Mr Nick Koikas and on behalf of the Council by Mr Glenn Leembruggen. Mr Koikas and Mr Leembruggen provided individual Statements of Evidence (exhibits C and 2), and a joint report (exhibit 4), and gave oral evidence. The acoustic experts addressed the likely noise impacts for residential receivers in the vicinity of the hotel. Those residential receivers are identified on Figure 1 of Mr Leembruggen's Statement of Evidence (exhibit 2):
20The approach adopted by each expert differed. Mr Koikas provided an acoustic assessment dated 29 April 2014, and in his Statement of Evidence (exhibit C) provided responses to matters raised in relation to that assessment by the Council. In summary, Mr Koikas' approach was to determine the octave band background noise levels at two unattended sites, one logger 2.5m from the hotel and 1.5m above the solid metal awning fronting Darlinghurst Road and the other 1m from the façade of the hotel and 1.5m above the solid metal awning fronting Roslyn Street. Attended measurements were taken at one site at the rear of the premises, and within the hotel. Mr Koikas converted patron noise levels as sound power levels per person, and used CadnaA computer software modelling to provide noise contours. Noise modelling was undertaken in octave band noise levels. The modelling was based on a maximum of 30 persons occupying the smoking terrace, and considered 15 patrons talking. Mr Koikas stated that his assessment considered the worst case scenario being that the premises operates at maximum capacity after midnight whereby inaudibility is to be achieved and that noise levels do not exceed the background noise levels at the boundary of a residential premise (exhibit C, p 7). In his Statement of Evidence Mr Koikas identified (pp11-12) that there would be a marginal exceedance at two frequency bands (2000 and 4000 Hertz frequency bands) which would be barely audible and unlikely to be perceived as intrusive because there would be other persons talking in the immediate neighbourhood and vehicles passing. In his opinion the Plan of Management prepared by Mr Martin would mean that instances where noise may become audible would be isolated instances.
21In his Statement of Evidence (exhibit 2) Mr Leembruggen stated that he had undertaken attended short-term monitoring at six locations, identified in Figure 4 of his Statement of Evidence:
22The relevant residential receivers were referred to in the experts' evidence by reference to their numbering in Figure 4. R1 is the units at 1A Roslyn Street; R2 is backpackers' accommodation at 34 Darlinghurst Road, on the corner of Roslyn Street; R3 is hotel accommodation ("Lido") adjoining the hotel at 2 Roslyn Street; R4 is residential apartments at 24-30 Llankelly Place; R5 is residential apartments at 11 Springfield Avenue; and R6 is backpackers' accommodation at 34-50 Llankelly Place.
23Mr Leembruggen noted that his assessment of background noise levels differed from that of Mr Koikas, those of Mr Koikas being higher (exhibit 2, p 13). Mr Leembruggen applied the noise criteria used by the Council, including that noise from the use must not be audible within any habitable room in any residential use between the hours of 12.00 midnight and 7.00am regardless of transmission path (exhibit 2, p 15). Mr Leembruggen explained his methodology at pp 17-21 of his Statement of Evidence. Mr Leembruggen applied the "noise begets noise" mechanism identified in research by Rindel (JH Rindel "Acoustical Capacity as a means of noise control in eating establishments", exhibit H), which shows that sound pressure levels increase by 6dB for every doubling of the number of people talking simultaneously in contrast to the usual rule of 3dB increase per doubling of talkers. Mr Leembruggen predicted exceedances of the noise criteria by up to 15dB at location R1 (1A Roslyn Street) and location R2 (34 Darlinghurst Road), and of up to 5 dB at locations R4 (24-30 Llankelly Place) and R6 (34-50 Llankelly Place) (exhibit 2, pp22-23). Mr Leembruggen outlined (at pp 13-14) results of measurements of patron noise taken by himself and his staff after preparation of exhibit 2 at outdoor smoking terraces at two hotels, the Aurora Hotel in Surry Hills (8 patrons) and the Newtown Hotel in King Street Newtown (30 patrons), concluding that his predicted levels were similar to the levels measured at Newtown Hotel and that even though there were fewer people at the Aurora Hotel the intoxication of the talkers resulted in their levels slightly exceeding the level in the Newtown terrace.
24In their joint report (exhibit 4) Mr Koikas and Mr Leembruggen agreed that in applying the inaudibility criteria, a sound is generally inaudible if it is 10dB below the level of background noise, and that if the level of the interfering sound is below the threshold of hearing it will not be heard (p 8). They differed as to whether backpackers' accommodation should be assessed to the inaudibility criterion after midnight. They agreed that sound levels produced by persons smoking and drinking in dedicated smoking terraces are not necessarily similar to those of patrons in restaurants eating and drinking, due to the lower number of patrons and the open façade of such terraces, which reduces the reverberation time in the terrace (p 9).
25The experts were generally agreed as to the background noise levels. The Table at p 8 of the joint report (exhibit 4) records that for a full week, background noise at R1 was 53dB, R2 52; and for weekdays R3 was 50, R4 47, R5 49 and R6 51. Applying the Council inaudibility criteria that a sound is generally inaudible if it is 10dB below the level of background noise sound in the same frequency range and if the level of the interfering sound is below the threshold of hearing it will not be heard, the noise criteria for the six sites is respectively 42, 42, 40, 37, 39, and 41 (exhibit 4, Table, p 9). The experts agreed that noise levels at R1, R2 and R6 is crucial, on the basis that if there is compliance at those locations there would be compliance at all residential receivers.
26In his contribution to the joint report Mr Koikas noted that in the course of providing data to Mr Leembruggen he had noted that the sound power level module used was incorrect and when the correct sound power levels were entered exceeding levels at some sites became apparent. With noise mitigation measures, the calculated resultant noise levels showed that marginal exceeding levels would occur for some frequency bands to the nearest premises. With 30 patrons the exceeding noise levels were significant, and Mr Koikas considered it necessary to reduce the number of patrons to 20 to produce exceeding levels that he considered to be insignificant (exhibit 4, p 22). Mr Koikas undertook further modelling based on 20 patrons, which demonstrated exceedances at R1, R2 and R6. To address those exceeding levels Mr Koikas considered further noise attenuation measures to the underside of the street awning adjacent to the hotel (exhibit 4, p 17, p 22).
27The experts agreed that the installation of sound absorbing ceiling and walls would reduce the general level of patron noise emitted from the hotel (exhibit 4, p 21), and proposed measures are now contained in a draft condition (exhibit G).
28The issues as to differing methodologies, and the significance of a reduction from 30 to 20 patrons after midnight as recommended by Mr Koikas, were pursued in the oral evidence.
29Amazonia was critical of Mr Leembruggen's reliance on the research undertaken by Rindel. The Rindel article (exhibit H) addresses the problem with predicting noise from people speaking in eating establishments and other social gatherings, noting (at 2.3) that there are unknown parameters relating to individual differences and how much people actually want to talk, which may depend on the type of social gathering, how well people know each other, their age, consumption of alcohol, and other social circumstances. The article refers to the Lombard effect, where persons with normal hearing raise their voice when subjected to noise, and notes that applying the prediction model developed in the article, the ambient noise level increases by 6dB for each doubling of a number of individuals present (2.2). Rindel notes that the prediction model was verified by comparison with measured data for a varying number of persons between 50 - 540 in large food courts and in a canteen, and being based on statistical conditions, "should not be applied to small rooms with a capacity less than, say 50 persons" (3.1). In oral evidence Mr Leembruggen was of the opinion that notwithstanding that qualification, the Rindel model that noise begets noise is still relevant; he agreed that while the circumstances are dissimilar to those on which the model was verified, the acoustic modelling is supported by his subsequent monitoring for the Newtown Hotel, and he chose a lower Lombard ratio in his modelling to be conservative.
30Mr Leembruggen addressed the significance of the change in maximum patron numbers from 30 to 20. His initial modelling for his Statement of Evidence and his contribution to the joint report was undertaken on the basis of 30 patrons, and overnight during the hearing he modelled for 20 patrons. Based on 20 patrons together with the addition of absorption measures, and based on groups of 3 with 7 people talking at the same time and a Lombard slope of 0.3, Mr Leembruggen's evidence was that there would be compliance other than for locations R1 and R2, and at those locations the exceedances would be at the 250 and 500 Hertz bands, at 4.4dB (which could be 3dB out either way). Mr Koikas agreed that R1 and R2 are critical; that his figures were in the same ballpark; and that there would be a need to ensure management of the terrace by staff on a regular basis. Mr Koikas agreed that if people were speaking in a loud voice the residents would be disturbed, and considered that the best way to address that would be with the proposed management approach.
31It is not necessary in my view to determine whether Mr Leembruggen's adoption of the Rindel prediction model, or Mr Koikas' modelling based on the CadnaA model, is to be preferred. Mr Leembruggen accepted that the differences between them of 3 dB in the initial modelling for the joint report could be accepted due to different assumptions between the CadnaA model and his model (exhibit 4, p 19). Whichever approach to modelling is adopted, there was ultimately common ground as to the background noise levels, and also that with 20 patrons and the additional absorption measures, there would be an exceedance of the noise criteria for R1 and R2. On Mr Koikas' evidence that exceedance would be between 1 and 2 dB at certain octave bands, and on Mr Leembruggen's evidence it would be 4.4dB (subject to a margin of 3 dB either way). On Mr Koikas' evidence there would also be an exceedance at R6 at two octave bands (exhibit 4, Table, p 17).
32Mr Koikas considered that the noise exceedance levels were marginal and would not be perceived by a person with normal hearing particularly when considered against the average maximum noise levels in the vicinity of Roslyn Street and Darlinghurst Road intersection (exhibit 4 p 17). He considered that the noise exceedances would not constitute offensive noise for the purposes of the Protection of the Environment Operations Act, when considered against the guidance issued by the EPA for Council officers which requires responses to six questions: is the noise loud in an absolute sense, and relative to other noise in the area?; does it include characteristics that make it particularly irritating?; does it occur at times when people expect to enjoy peace and quiet?; is the noise atypical for the area?; does the noise occur often?; and are a number of people affected by the noise? (exhibit 4, pp 23-25). Given the agreement between the experts that the applicable noise criterion is the Council's criterion for inaudibility after midnight (subject to Mr Koikas differing as to whether that should apply to backpackers' accommodation), I prefer the approach of Mr Leembruggen (exhibit 4, p 25) which was that if the noise does not comply with that criterion especially after midnight there are significant risks that residents will find the noise disturbing. On that basis, I agree with Mr Leembruggen that the most important step in mitigating the risk of noise disturbance after midnight is to ensure that the hotel complies with the post-midnight noise criterion of inaudibility, and if it does not comply, then arguments as to whether the noise is offensive can be pursued.
33As noted above, one of the deferred commencement conditions imposed by the Council required preparation of a Plan of Management to align with the operation of the smoking terrace. Amazonia's planning expert, Mr Andrew Martin, provided a revised draft Plan of Management (September 2014) with his individual Statement of Evidence (exhibit D). During the course of the acoustic evidence, further amendments were proposed to that Plan of Management (exhibits F and 11). The proposed management regime for the smoking terrace between 12 midnight and 4.00am as provided in exhibit F includes having a staff member with a current responsible service of alcohol (RSA) certificate patrol the terrace between 12.00 midnight and 4.00am and fill out a compliance sheet every 15 minutes noting patron numbers and whether there is too much noise, and ensure that there are no more than 20 patrons and that they keep the volume to a minimum. The staff member is to rove between the gaming terrace and the south western terrace and be principally stationed in the front/main bar area, and carry out minor duties such as glass collection, table wipe down, bar service and ashtray collection. Exhibit 11 proposes that a staff member conduct a listening test outside the apartments at 1A Roslyn Street, listening for patrons' voices during the lulls in noise from traffic and pedestrians and if voices are continually audible, instruct patrons to reduce their speech level. The proposed management would also require independent testing by an acoustic consultant based on the log recordings to determine the likely times of high patron numbers.
34Mr Koikas and Mr Leembruggen discussed the proposed management regime in their oral evidence. They disagreed as to the desirable frequency of independent acoustic monitoring and the extent to which the acoustic expert would be in contact with neighbouring residents or rely on the log and records of complaints. Those differences were not resolved, with Mr Leembruggen expressing reservations about the proposed frequency of testing and the drafting of the proposed conditions, stating that there would need to be strong management in the interests of residential amenity.
35Mr Leembruggen did not consider that reliance on a Plan of Management would be feasible. If there were no alcohol, 20 people talking quietly would not be a problem, however with alcohol there would be a risk of non-compliance with a Plan of Management, and it would depend on the nature of the patrons and how well they were managed. Mr Koikas agreed that if the Plan of Management failed there would be a problem, however in his opinion the present operators would do their best to comply.