This report summarizes acoustical analyses performed by David Coate Consulting (DCC) to address issues related to the mezzanine community area in REI's Conceptual Store. The community area is a space designed for training and educating staff about REI products. Normal conversation, lectures, and video presentations are examples of the types of communication that will be required in this space. The REI design team wants to retain a degree of openness of the community level to the rest of the store. The current design calls for transparent walls which would be open at the top.
DCC analyzed the potential for speech interference caused by customer activity and various design recommendations to protect the acoustics of this space.
Noise Measurements in Framingham, MA REI Store
The most important parameter in determining the amount of potential sound intrusion into the mezzanine level is the general ambient noise level in the store. Ambient noise sources include customer activity, music and paging from loudspeakers, and HVAC noise. REI staff suggested the Framingham, Massachusetts store as a candidate location for measuring in store noise. Obviously, store noise in yet to be constructed REI stores could vary from this, but it is unlikely that customer noise would vary by a large amount. Other noise sources are controllable, and customer noise should only increase by 3 decibels if customer activity were double that of the Framingham store. A three decibel increase is barely noticeable to most people.
DCC collected noise measurement data at the Framingham location on March 17, 2007 and March 30, 2007 for a total of eight hours at eight locations within the store. Four locations were established on a grid system on each floor and then averaged to generate an overall store spatial average. The overall frequency spectrum of these measurements is shown in Figure 1.
Figure 1. REI Framingham Overall Average Ambient Noise Levels
This ambient noise spectrum is at NC 50 which falls in the range typical for shops.
Community Area Sound Levels
DCC employed two different computer modeling techniques to estimate store-generated noise levels within the community area. Room acoustics modeling requires sophisticated software and it turns out that analysis of the noise barrier effects of the partial height mezzanine walls is a very complex room acoustics problem. DCC used CADNA, an environmental noise prediction program, to analyze the noise barrier effects, and EASE, a room acoustics program to analyze the absorption effects of architectural finishes. The effects of reverberation (sound scattering off of various surfaces) could be an important issue because sound could reflect off of the 2nd floor ceiling, and then over the top of the mezzanine wall. That is, the extent of absorption treatment on the 2nd floor ceiling could affect how much sound actually passes through those openings into the community area. DCC modeled noise levels on the 2nd floor of the prototype store using the measured data shown in Figure 1, represented by twenty point sources randomly distributed on the second floor. The number of point sources is somewhat arbitrary, but they represent a moderate to high number of customers (i.e., customers are the point sources). Figure 2 shows the results of this modeling. The overall spatial average (i.e., the average of all the colors in the figure) matches the data in Figure 1.
Next, DCC modeled the mezzanine level, specifically the barrier effects of the partial height mezzanine walls, using the 2nd floor noise model as the noise source. This model assumed a mezzanine floor elevation of 3.5 meters, top of glass wall elevation of 7 meters, second floor elevation of 6.09 meters, and 2nd floor ceiling elevation of 10.6 meters. Figure 3 shows the results of this analysis- the estimated store generated noise levels penetrating into the community area. This modeling exercise shows that mezzanine levels would be at about 44 dBA. The combination of the barrier effect of the glass walls, and physical distance from the noise sources would provide approximately 10 dBA noise reduction.
Figure 2. CADNA Model of Second Floor Noise Levels
Figure 3. CADNA Modeled Mezzanine Noise Levels
Based on the estimated ambient noise spectrum in the mezzanine, the level of speech intelligibility can then be estimated. That is, in order for speech to be intelligible, the sound pressure level of the speech has to be higher than the ambient noise level by a certain amount. The Speech Transmission Index (STI) is a quantitative measure of speech intelligibility, where 0 is completely non-intelligible, and 1 is completely intelligible. There are also subjective ratings for STI values between 0 and 1. Table 1 shows the results of the STI analysis for the community area for various speaker to listener configurations.
Table 1. Speech Intelligibility in the Community Area
The data in Table 1 shows that there will be no problems associated with normal conversation in the community area. However, there will be problems understanding unamplified speech for someone presenting at a lectern to a group of people seated within the community area. Consequently, in this configuration, speech will need to be amplified via a lavaliere or lectern microphone through an appropriate sound system. Loudspeaker to listener distances may need to be adjusted, along with overall sound system levels, so as to not create excessive noise within the 2nd floor store area.
Reverberation in the Second Floor Retail Area
The previous analysis is based on the assumption that the acoustics in the 2nd floor retail area will have no effect on sound entering the mezzanine area. However, it is likely that reverberation and reflections from the 2nd floor ceiling will cause higher noise levels in the mezzanine than assumed in the previous section. In order to address this issue, DCC modeled reflections using EASE room acoustics software. Room dimensions and architectural finishes were modeled, including 1.5" of Tectum on the ceiling above the mezzanine, and steel deck for the rest of the ceiling. Next, DCC assumed that the entire 2nd floor ceiling was comprised of 1.5" Tectum. EASE predicts that there would be a 5 dB reduction in noise level at the mezzanine area by treating the entire ceiling. However, there are a number of complicating factors including the amount of absorption in the room caused by store displays, clothing, and other products which were not included in the analysis. Furthermore, reverberation is likely more localized than accounted for in this level of model detail.
It is likely that both the magnitude of the reduction in noise level and the extent of the ceiling treatment needed will be less than predicted in this modeling exercise. Because of the uncertainty in the modeling detail and accuracy, DCC recommends that the entire 2nd floor ceiling be treated with 1.5" Tectum.
In addition, DCC recommends that care be taken in locating 2nd floor sound sources such as ceiling mounted loudspeakers as well as HVAC ducts. Those sources should be located as far as possible from the openings above the glass walls. HVAC servicing the community area should be quieted to minimize the sound masking effects described in the previous discussion.