Note: With this installment, we kick off a series of technical articles dealing with various aspects of acoustics technology, penned by experienced designer/acoustical consultant David Kennedy of David Kennedy Associates. In future installments, we will delve into topics such as creating subwoofer arrays, the mechanics of perceived loudness, the science of transducer design and more, so stay tuned! —ed.
With recent talk about democracy for listeners, this is a good time to discuss why it is important for the sound coverage in a venue installation to be as even as possible, how to evaluate sound coverage, and how speaker modeling programs can be used to design better sound coverage, along with some modeling issues to watch out for. Consistent sound coverage has actually been important for sound-reinforcement design engineers, trained by SynAudCon, for decades. For non-engineer readers, we will present some basics to help you better understand array coverage.
Since they were first introduced in the mid-1980’s, I have been using sound modeling programs for array coverage simulation (and later on, room/acoustic simulation) on almost all of my sound system designs. But my training on these modeling programs was decades ago. So, we have begun polling the loudspeaker manufacturers for updates to follow in later articles, to include: Array Modeling and Control updates along with Modeling Engineering details, with quotes from Array Modeling experts and vendors. Follow-up articles (in Part 2 and Part 3 of this article) will delve into the history and current trends in array modeling/control, with another on the modeling of cardioid subwoofer arrays. Another follow-up article will also discuss the universal programs to compare array models, and detailed acoustical analysis vs. array-specific modeling and control programs.
Music sound coverage need not be as consistent as speech, especially if the overall program has a high sound level. Getting great coverage to every seat in your venue requires a great deal of time, energy and money to get right. But in reality, there simply is no way that every seat in the venue will experience the exact same sound. That said, sound-reinforcement systems quality, and the proper design tools for coverage (dispersion) have come a long way over the past few decades.
The first step in getting great coverage is evaluating your current system. This is commonly done with listening tests that consist of pink noise followed by some music. Pink noise is very useful in this situation because it’s easy to hear when you walk out of the coverage of your horns. If your system has Left & Right speakers, do the Pink noise test through only one output channel/array. Most rooms that suffer from poor coverage have problems with HF (high-frequency) coverage more than lows and subs. In a venue where speech intelligibility is of paramount importance, the mid- and high-frequency coverage should be the highest priority. More on what coverage models (AKA heat maps) should look like and a few types of fill speakers will follow below.
In the November 2018 issue of FRONT of HOUSE (“Sound Sanctuary,” page 58), author Vince Lepore, the director of event technical operations at Full Sail University, noted that “if you hear high-frequencies fall off at any point within the coverage area, you’ve got a problem that needs to be addressed. Before jumping to conclusions about needing a new speaker system, think critically about how your system is aimed. It’s hard to get into specifics here because every system is different, but poor speaker aiming has been the culprit in a number of projects I’ve been involved with. It’s amazing what some aiming changes can do for a system that isn’t covering correctly. A few degrees of adjustment can go a long way to improving your system’s coverage.” (See fohonline.com/articles/sound-sanctuary/dispersion for the article, and its tips on using fill speakers.)
Fig. 1 shows two views of an EASE sound model in a large, box-style proscenium venue with a balcony, showing the Broadband Total SPL in the room, which includes reflections from the surfaces in the room and does not represent the direct sound field from the loudspeakers themselves. This is typical of many sound modeling examples and proposed/free designs from loudspeaker factories and suppliers.
Fig. 2 shows a modeling image from another array modeling program called Soundvision, from L-Acoustics (similar to EASE and others). This coverage map is a more realistic representation of how “comb-filtering” — caused by interaction between multiple loudspeakers — should be displayed (setup in the program). But both of these coverage map examples (in Fig. 1 and Fig. 2) show a 20 dB SPL (sound pressure level) variation. The problem with the marketing approach of showing almost perfectly consistent sound (Broadband) coverage, is that it is not helpful, as the sound engineering industry standard for displaying SPL variation is 6 dB to 10 dB. Array coverage models should have the dB scale set to look more like the image in Fig. 4 (below).
Showing two times the standard scale over an average of a wide range of frequencies and/or multiple arrays at the same time, is not realistic or true. And of course, someone must pay for these free, but basic services.
A recent trend is the offering by pro-loudspeaker factories to do “free” array modeling of coverage for their design/build dealers. We applaud the live-sound speaker manufacturers for taking the initiative and making the investment to create and support these and other programs, and for their willingness to provide a free service to their dealers and end-users, to help ensure that their products get implemented well, to the good of all. However, most of the modeling programs from a brand-name source only include their own loudspeaker models in their database.
(We will touch more on loudspeaker array modeling trends in Part 2 of this article series.)
Free array modeling has many limitations. The results can be very hard to compare, and the modeling results can sometimes be of questionable value. The limitations with “free” array modeling services from loudspeaker vendors include:
- Vendors may display multiple arrays at the same time (hiding poor coverage)
- Comb-filtering from interaction between loudspeakers may not be displayed
- The interaction between loudspeaker vendors and sound designer is limited
- There is no universal program (various programs use different calculations)
- Lack of consideration of echoes from loudspeakers off of walls
- Lack of objective recommendations (limited to brand offerings)
- Lack of standard scale for how sound coverage is displayed
- Lack of sound system clean power & conduit plans
- Lack of venue site survey by loudspeaker vendor
An expert and independent sound consultant/engineer, should not only provide all of the above services, but should do so in an objective and integrated manner.
What is Array Coverage?
While estimating total sound coverage is fairly intuitive, the understanding of speaker array interaction and psychoacoustics does not come easy. Extensive training and years of array modeling will train a sound designer about loudspeaker interaction and sound design methods. Author Bob McCarthy has written extensively on the subject. The graphics in Fig. 3 are from his book, Sound Systems: Design and Optimization — Modern Techniques and Tools for Sound System Design and Alignment. They show how coverage patterns are measured and shown.
dB SPL Basics
The decibel (dB) is a unit of measurement used to express the ratio of one level/value to another on a logarithmic scale. Suffixes like SPL (Sound Pressure Level) are commonly attached to the basic dB unit in order to indicate the reference value. One decibel, the smallest sound level change detectable to the average listener, is one-tenth of one bel, named (a century ago) in honor of Alexander Graham Bell. A quadrupling of sound power is commonly described as “6 dB” And that 6 dB figure is the pro-sound industry standard for defining the usable limit for a given speaker coverage area.
What Should Array Coverage Models Look Like?
Loudspeaker array coverage models should look more like the lower image in Fig. 4 (from d&b audiotechnik; dbaudio.com). The upper image shows a realistic representation of how “comb-filtering” — caused by interaction between multiple loudspeakers — should be displayed (with more realistic scale not shown). But the lower image clearly shows how a single “point-source” loudspeaker — with dispersion almost wide enough to cover the venue — has much smoother sound level coverage across the seating area.
The remaining problem below is the lack of coverage in the front corners (circled in red). Small side- or front-fill speakers, digitally delayed to synch with the main speakers, will not only provide the needed supplemental sound level coverage in the front corners, but also lower the sound image, making the sound seem to come from the platform (more natural sounding). While low-level interaction between the loudspeakers will occur, it will not be very noticeable. Extremely long/tall rooms benefit from small digitally-delayed loudspeakers in the rear of the venue.
Modeling of Venue Acoustics & Directional Realism
Reasonably accurate directional realism and intelligibility of the talent/worship-leaders is another related issue for me; it is quite hard to do in most venues with just Left and Right loudspeakers. Sound imaging with poor directional realism causes cognitive dissonance —essentially stress due to contradictory information. A central array is important for directional realism and optimized speech intelligibility. For more on this subject, and quotes from a few engineers on this directional realism topic, see comments online at: www.line-arrays.com/stereo-or-mono-arrays.html
I feel the need to debunk the trend of thinking that modern loudspeaker arrays can eliminate the need for proper acoustics — that is rarely the case! The right venue-specific loudspeaker arrays, with proper aiming, can minimize the need for extensive acoustic treatment. But selection of the proper loudspeaker array for a given venue can rarely be separated from the acoustic space that it operates in. So, allowing vendors to recommend their loudspeakers without the proper study of how they will cause echoes from room acoustics is not a complete solution! Reverberation Time should be optimized, and long echoes controlled in meeting, performance and worship spaces. More on the proper design of room acoustics, will be covered in the next installment in this article series.
The proper use of a 3D coverage modeling program to design acoustics and sound by a credible expert and, ideally, an objective sound engineer, can help ensure that everybody in your venue can hear the sound loud and clear.
Author David Kennedy operates David Kennedy Associates, a company specializing in the design of architectural acoustics, AV Systems and custom loudspeaker arrays. Since 1991, DKA has worked on more than 300 projects, including churches, performing arts centers, theaters, commercial buildings and auditoriums. Visit him at d-k-a.com.
The Other Side: Independent Array Experts on the Misuse of Modeling
It cannot be overemphasized that modeling/analysis systems — like any other form of technology — are useful items in the designer’s toolkit. When correctly applied in the right circumstances, the results can be impressive and essential. Yet, the widespread availability of modeling tools, which can be improperly utilized, may have its dark side as well.
Commenting on this topic on SynAudCon’s prosoundtraining.com website was noted consultant/designer/electroacoustics expert Dr. Richard Honeycutt. “By the mid-1960s, independent consultants and contractors had taken over most sound-system design and installation, with manufacturers providing some occasional support as requested,” he explains. “Lately, however, things have changed a bit. Along with some colleagues, I have noticed a trend toward sound contractors getting at least the speaker system design done by a manufacturer. In at least three cases, this procedure has caused problems. And there are at least four reasons why:
- Factory engineers are not in a position to perform a proper needs assessment, which is the first step in any successful system design.
- Factory engineers are not always experts in managing the acoustical/electro-acoustical interface at which the speakers operate.
- Factory engineers just may — hold your breath here — be tempted to specify more expensive speakers, even when they may not be the best choice.
- The contractor’s engineer (or salesperson) may choose the speaker brand without a dispassionate evaluation of multiple manufacturers’ offerings, potentially leading to sub-optimal system performance for the venue.
The posting also included the example shown in Fig. 5 (above). The image to the left shows direct coverage from the manufacturer’s design; the right image shows the results of the improved coverage from using a single center speaker.
Pat Brown, a sound engineer trainer at SynAudCon added, “I’d never trust a vendor to model my coverage. They can’t be expected to take the time to do it right.”
Jerrold Stevens, associate principal at Marsh/PMK International replied, “So true how important it is for knowledgeable and honest persons to do the modeling and design loudspeaker systems. It is quite easy for someone to ‘fudge the settings’ to make their design appear to work better than reality. It would be quite easy for someone to drop a single speaker in a model, aim it at the ceiling, and then run a coverage map that looks like +/- 0 dB everywhere!”
Jeremy Johnston at Wide Open Audio wrote, “Modeling results are often misinterpreted or incomplete, and yet well-meaning contractors and DIY’ers still try to build arrays based off the model results. They don’t have all the information. They’re trying to make good decisions (or someone is selling them something based on poorly interpreted or incomplete modeling results), and the end results aren’t always good (or useful).”
Rocky Giannetta, CTS-D, president of Layer 8 Inc., said “the most unfortunate trend is the misuse of software programs as marketing tools. The bottom line is that both modeling and measurement software programs are useful tools in the hands of pros who have a good understanding of sound energy, and electro-acoustics physics.”
So, it is very important to qualify who is doing the new loudspeaker array modeling and sound system engineering for your venue (and who pays them).
—David Kennedy
Anechoic Listening
Among the most useful tools in analyzing any acoustical space are your own ears. As mentioned, a simple playback of pink noise can be very revealing. Engineers often also depend on certain disks as a playback reference — Donald Fagen’s Nightfly is one particular favorite. Unfortunately, no matter how well recorded or mixed, essentially any commercial recording also includes a fair amount of ambience (or sometimes a large amount of ambience) that can limit their use in evaluating system performance in a performance space. When the CD plays back, you hear both the ambience of the space combined with whatever reverbs were used (natural or artificial) during the record/mix process.
One notable example is Denon’s 1988 Anechoic Orchestral Music Recording, a CD of classical music — which was close-miked with the players performing inside a makeshift anechoic chamber set up on a recording stage. Acoustically, “dry” is the word here, so when played back in a performance space, you get an actual feel for the reverberation characteristics of that environment. Unfortunately, the CD is long-since out of print and no one has followed it up with a sequel, but although pricey, it can be found on the secondary market. Also noteworthy is a 1995 Syn Aud Con “Test CD for Sound Reinforcement Systems,” which contained a dry speech track featuring narrator John Royer reading an excerpt from Don and Carolyn Davis’ Sound System Engineering text. Although also long out of print, a downloadable WAV file of that track can be found via a simple web search and it’s a useful tool for quick intelligibility checks. —George Petersen
