Published in November 2005

Indoor Stadium Design
By David E. Marsh, FASA


Figure 1. Rendering of the new stadium currently under construction for the NFL Arizona Cardinals.

     A stadium can be open-air or enclosed and can seat a few thousand (e.g., high school football stadiums, minor league ballparks, etc.) all the way up to 100,000 or more (e.g., NFL and large university football stadiums). The term arena, on the other hand, normally refers to an enclosed facility primarily used for court sports and/or hockey. The largest arenas, such as for NBA and NHL games, only seat in the range of 18,000 to 22,000 on the high end—about 20% to 25% the size of larger stadiums. We’ll talk here about acoustical treatment of enclosed stadiums, which presents unique problems in terms of sound system design and acoustical treatments, making them more difficult than arena projects.

Reverberation Time (RT60)
     Within the design discipline of architectural acoustics, there is a sub-discipline called room acoustics that generally deals with how sound behaves in an enclosed space. There are many objective and subjective attributes of room acoustics used in the design and measurement of auditorium spaces, but for stadiums, the one that stands out is reverberation time.
    "Reverberation” refers to sound that “hangs on,” or reverberates, after cessation of the sound source. The objective measure of reverberation is called reverberation time. This is the time, in seconds, that it takes for the sound level to drop off to a point 60 decibels (dB) below the starting sound level. Therefore, it is abbreviated RT60 or T60 (sometimes just RT or T).
     Many of us have experienced the pleasant effect of sound in a well-designed concert hall taking two seconds or so to die away when the orchestra stops abruptly. However, even a classical music hall can have too much reverberation. This causes the notes to run together in a way that reduces or even eliminates the essential clarity. Likewise, too much reverberation in an opera hall or theater makes lyrics and speech unintelligible. It might surprise you then to learn that:
•   Large stadiums typically have RTs averaging about seven seconds for the midrange frequency octave band of 500Hz.
•   Low frequency RTs in these facilities generally are in excess of 10 seconds and 63Hz values may be on the order of 15 to 20 seconds!
     This “large room” phenomenon occurs not only in enclosed stadiums, but also in those that are partially open, such as Texas Stadium (current home of the NFL Dallas Cowboys) and the new stadium being designed for the NFL Arizona Cardinals with its retractable roof in the open position (Figure 1).

Figure 2. Table of reverberation time (RT60) data for pro sports stadiums.

Figure 3. Chart of reverberation time (RT60) data for pro sports stadiums (ref. Figure 2).

     RT60 is an important acoustical design parameter for stadiums because it affects speech intelligibility and the clarity of both prerecorded and live music. RT60 typically is calculated and/or measured for each of six different octaves with center frequencies of 125Hz, 250Hz, 500Hz, 1kHz, 2kHz and 4kHz, ranging from bass to treble tones. When a single reverberation time is given without reference to frequency, it usually is the 500Hz value or an average of the 500Hz and 1kHz values referred to as the midrange average RT60.
     Measured RT60 values become increasingly less reliable at 250Hz and lower frequencies because of signal-to-noise ratio limitations in the measurement process. Predicted RT60 values (e.g., calculated during the design stages of a new project) normally are limited on the low end to the 125Hz octave band because sound-absorption test data is not commonly available below that point for typical building materials. Yet, as we shall see, low-frequency reverberation cannot be ignored in these “super-sized” sports venues.

	Figure 4. BankOne Ballpark, home of the MLB Arizona Diamondbacks, with retractable roof open.
	Figure 5. BankOne Ballpark with retractable roof closed.

Stadium RT60 Measurements
     Figures 2 and 3 show data from a total of 10 RT60 measurements made in a variety of stadiums. Some are permanently enclosed, some have retractable roofs and a couple can be characterized as having a “hole” in the roof. This data shows that even partially enclosed stadiums are very reverberant. Figures 4 through 8 show the size of stadiums exhibiting the types of RT characteristics described here. Captions point out acoustical treatments that were employed (or lack of treatments).
     Pelton Marsh Kinsella (PMK) has performed RT60 measurements in some of these large facilities using a “yachting cannon,” with 10-gauge shotgun shell blanks, as an impulse source, collecting the data with multiple spectrum analyzers that have fast time-store capabilities, and post-processing the data with computer software. The yachting cannon idea was suggested by Jack Randorff (Randorff & Associates, Ransom Canyon TX) as a way to generate enough low-frequency energy for measurements all the way down to 63Hz. PMK’s cannon-based 63Hz measurements are not presented here because of ongoing research to establish a consistent method for presentation of this less-reliable low frequency data.
     Randorff & Associates used the yachting cannon at the Astrodome in 1988 (see Figure 9) to obtain audio recordings of the impulses, using an instrument grade recorder, and then processed the data with a strip chart recorder to arrive at RT60 values. At the same time, PMK employed time-delay spectrometry using a TEF Analyzer for parallel RT60 measurements because of the inherent signal-to-noise superiority of this measurement process. The TEF-based RT measurements tracked the cannon-based measurements very well, as documented in a paper presented to a Fall 1988 Audio Engineering Society meeting.
     However, interpretation of TEF data generated from multiple loudspeaker clusters in the house sound system proved difficult compared with data provided from the cannon’s point-source impulse. Modern portable spectrum analyzers now allow quick simultaneous collection of impulse responses at multiple locations, and the cannon serves as an ideal source for huge venues such as sports facilities.

	Figure 6. BankOne Ballpark used a 3" fluted “acoustical deck” for both the retractable roof and the permanent roof deck areas. Rock wool roofing insulation was used above the deck, instead of the usual polyisocianurate, to increase sound absorption. Many of the wall surfaces employed perforated metal backed with fiberglass for additional sound absorption and various types of treatment were used on other vertical surfaces as well. The combined effect was the shortest reverberation for the group of retractable roof stadiums measured with the roof closed.
	Figure 7. For Miller Park, home of the MLB Milwaukee Brewers, an “acoustical roof deck” was used, but vertical surfaces were not treated and the internal volume is very high. This resulted in reverberation times that are among the longest of the retractable roof group measured with the roof closed.

Evaluating Low Frequency RT60
     One parameter often examined in room acoustics work is the Bass Ratio (BR), defined as the 125Hz and 250Hz Early Decay Time (EDT) average divided by the 500Hz and 1kHz average EDT. By definition, EDT includes only the first 10dB of sound-energy decay. In stadiums, EDTs are very short; the decay rate is similar to a measurement made outdoors. This is because there are few reflections in the early part of the impulse response measurement. Therefore, EDT is useless for determining BR in a stadium.
     However, BR is still a useful parameter for stadium design if RT60 values are used instead of EDT for the low-frequency and midrange averages. By adding 63Hz RT60 values into the low-frequency average to arrive at an “extended” low-frequency average, the BR becomes an Extended Bass Ratio (EBR). This is shown in Figure 2 where appropriate.
     As a point of reference, consider that halls for music performance should have BR values close to unity in small venues (less than 2000 seats) and for halls with RT60 values greater than 2.0 seconds. BR should be in the range of 1.1 to 1.2 for larger halls with shorter RT60 values.
     Now, let’s go back to sports arenas. Remember that we began with a discussion of the huge size difference between stadiums and arenas. PMK’s collection of more than 20 arena measurements allows categorization into three groups with:
•   midrange average RT60 = 2.08 seconds, average BR = 1.45
•   midrange average RT60 = 2.83 seconds, average BR = 1.11
•   midrange average RT60 = 3.91 seconds, average BR = 1.18.
      BR values for the less reverberant arenas are higher on average than acceptable for a music hall, but the more reverberant arenas have BR values in the acceptable range. This is just an observation, not a suggestion that more reverberation is a good thing. Now look at Figure 2 and note that the average BR is 1.55 compared to the average EBR of 1.82 for the four stadiums with 63Hz data. This shows the extreme nature of low-frequency reverberation problems that can occur in stadiums compared to arena-size venues. To be fair, the four stadiums with EBR values have a BR average that is higher than the BR average of the other six measurements (1.8 and 1.4, respectively). The bottom line here is that sports venues in general are difficult acoustic environments for music, but stadiums present exceptional challenges, especially in the low frequencies.

Arizona Cardinals Stadium Acoustics
     Figures 10 and 11 show interior views of the Cardinals Stadium within a computer-based acoustical model created with the EASE software package. RT60 values were calculated using this model, plus a “fudge factor” derived from numerous predicted vs. measured stadium reverberation times. It was determined that inadequate sound absorption would be offered by the already-purchased metal roof deck with “acoustical” features consisting of 1½-inch perforated corrugations (flutes) with fiberglass infill.
     PMK recommended using horizontally suspended acoustic banners under all treatable areas of the roof deck. The retractable portion of the roof uses a fabric-like material and there is an additional “halo” of the same material around the “hole” created when the roof is open. This limits the treatable area of the roof. Because of cost concerns, the owner requested to hear an audible demonstration (auraliza-tion) of the stadium with three different scenarios:
•   with only the minimal acoustical treatments specified for available wall surfaces and 50% of the seats occupied (i.e., no roof treatment)
•   with the addition of the recommended acoustical banners covering only 50% of the treatable roof area
•   with 100% of the treatable roof area covered by banners.
     Having worked jointly with Miami-based Pro Sound on the stadium audio and video systems design, PMK was able to produce the required auralizations with all loudspeaker levels balanced and equalized within the model. Three additional scenarios were modeled where the loudspeaker line arrays covering the lower seating bowl were set to a reduced elevation to get them closer to the seats and re-aimed appropriately. The owner picked the 50% banner treatment with the lowered loudspeakers as the best balance between cost and overall clarity of words and music.
     We were concerned that the own-er’s decision did not adequately account for the inevitable long reverberation times in the 63Hz octave band that could not be modeled or auralized in EASE. Therefore, an alternative treatment was recommended using 4-inch-thick fiberglass duct liner attached to the underside of the fluted roof deck, thereby creating a 1½-inch airspace behind much of the fiberglass for additional low-frequency sound absorption. A recommendation was also made to replace the polyi-socianurate roofing insulation on top of the deck with rock wool to further increase low-frequency sound absorption. The lower installed cost of fiberglass ductliner as compared to the banners would allow a greater percentage of the roof to be treated, and the owner accepted this option.

	Figure 10. View from the Club seating level within a computer-based acoustical model for Cardinals stadium showing main line-array loudspeaker clusters.
	Figure 11. View from the Upper Deck seating level within a computer-based acoustical model for Cardinals stadium highlighting the upper deck satellite loudspeaker clusters.
Figure 12. Cardinals Stadium roof deck acoustical treatment final layout (one of three options considered).

     Figure 12 shows the final layout of acoustical treatment on the roof deck as approved by the owner. This was one of three options considered. “Option 2” was selected based on improved control of echoes from the end-zone walls, scoreboards and other vertical surfaces. Figures 15 and 16 are construction progress photographs showing some of the treated areas.
     Comparison of modeled vs. measured RT60 data in stadiums along with analysis of available 63Hz information has given PMK a proprietary method for estimating RT60 values during the design stage of new stadiums, including a range of values for the 63Hz octave band. This method was used while comparing the acoustical treatment options for the Cardinals stadium. The 50% banner solution is directly compared to the “Option 2” ductliner solution on Figures 13 and 14. Notice that the midrange average reverberation time is about one second longer than the average of the other stadiums discussed earlier, but the BR value is lower. Also, the ductliner treatment gives an EBR that is equal to or lower than three of the four stadiums for which EBR is presented (depending on the 63Hz value used), and below the overall average of those four.

Figure 13. Predicted RT60 values and low-frequency analysis for Cardinals Stadium.

Figure 14. Predicted RT60 values for Cardinals Stadium with estimated range for 63Hz octave band.

Summary and Conclusions
     Stadiums have categorically longer reverberation times than other types of public assembly spaces and low frequency reverberation is a particularly troublesome issue. Simultaneous averaged reverberation time measurements in these types of venues are best accomplished with a single loud impulse source such as a yachting cannon that can provide enough sound energy to obtain data at very low frequencies. However, further research is required with regard to post-processing low frequency impulse responses for repeatable RT60 results because of the increasingly unreliable nature of measurements at progressively lower frequencies.
     Prediction of low-frequency reverberation during the design stage of a new project is even more difficult than measuring it reliably in a completed facility. Producing audible simulations from commonly available computer modeling tools adds value, but such tools do not address the very low end of the frequency spectrum where the biggest problem lies. Yet, it is possible to at least establish a range of 63Hz RT60 values and this is useful in stadium acoustical design.
     The acoustical design of a new stadium for the Arizona Cardinals employed novel methods to address the acoustical challenges detailed here, including extensive modeling, aural-ization and low-frequency RT60 calculation techniques. The prescribed acoustical treatments, in combination with loudspeaker clusters located as close as possible to the seating areas, will provide improved speech intelligibility and musical clarity, compared to the original “acoustical” roof deck solution with the main loudspeaker clusters at their original elevations.

	Figure 15. Cardinals Stadium acoustical treatment on the underside of the roof deck.
	Figure 16. Black faced ductliner attached to the underside of the Cardinals Stadium roof deck is covered with a white fabric resulting in a light color that matches the interior design.

Managing Principal of Pelton Marsh Kinsella (PMK), Dallas TX, David Marsh is president-elect of the National Council of Acoustical Consultants (NCAC), a Fellow of the Acoustical Society of America (FASA) and a member of Sound & Communications’ Technical Council. He writes Sound & Communications’ monthly “Sound Advice” column. Send comments to him at dmarsh@testa.com.