Nominal and Advanced GLL operation modes
The loudspeaker development team at Biamp is introducing a new way to use GLLs for simulating loudspeaker performance in a venue project. We believe this will be helpful and informative, giving you both the most accurate data available and allowing for comparisons with the loudspeaker offerings from other manufacturers. This also facilitates comparison to some older Biamp, Apart, and Community loudspeakers.
Our new GLLs will have two modes of operation, Nominal and Advanced. For GLLs constructed on the simple-loudspeaker framework, like the Desono EX-S series, these two modes of operation are built into the GLL itself. For GLLs constructed on a line array-loudspeaker framework, like the LVH-900 series, there will be two different GLLs, one for each mode of operation. Why we decided to break things down this way is beyond the scope of this brief
article. Suffice it to say that we think this will result in the best workflow for those using these GLLs.
Why are these different GLLs needed?
It’s no secret that different manufacturers rate their loudspeakers using different methods. This can make it quite difficult to accurately compare the performance of different loudspeakers. Some manufacturers follow published standards and/or recommended practices. Some don’t.
At Biamp we believe it is critically important to not only follow published standards (when applicable) and best practices for measurement and specification, but to also do our best to put forth the most accurate data that correlates well with how our loudspeakers perform in real-world applications, both objectively and subjectively. We want our simulations and reality to match as closely as possible.
Providing accurate loudspeaker modeling data accomplishes that task. However, if these types of specifications and modeling data are compared to data that were derived in a different manner it might seem that one loudspeaker is inferior to another on paper or simulations. As an example, an accurately specified loudspeaker might have a maximum SPL rating of 118 dB and a very similar (almost identical) loudspeaker might be specified with a purely calculated
maximum SPL of 123 dB, or higher. Does the second loudspeaker really have that much more maximum output capability? In most instances a real-world, side-by-side comparison could show this to demonstrably not be the case. The first loudspeaker might have just as much, or more, output capability than the one with the larger maximum SPL specification or mapping in a software simulation.
So how do we provide the most accurate data we can and still show that our loudspeakers have comparable, or superior, performance to others that might be rated differently? We use two different sets of data to accomplish this.
What are the Nominal and Advanced GLLs for Biamp loudspeakers?
The Advanced GLLs (and the Advanced mode within GLLs) use the most up-to-date methods and best practices known today to measure and derive data that accurately characterizes loudspeaker performance. These are standards-based procedures and specifications that correlate well with the real-world performance of the loudspeaker.
The Nominal GLLs (and the Nominal mode within GLLs) rate the maximum continuous input to the loudspeaker, and subsequently the maximum continuous SPL output capability, with the methods that have been used over the last several decades by loudspeaker manufacturers. Years ago, these methods were the most up-to-date and best practices. So, we shouldn’t dismiss them as being wrong. It was just a different way to do things back then. However, today there are better approaches that can be used.
The Nominal GLL (mode) allows our newer loudspeakers to be compared to our older loudspeakers. It also allows our loudspeakers to be compared to loudspeakers from manufacturers that might not have adopted these more modern methods for measuring, characterizing, and rating their loudspeaker systems.
Which GLL should I use?
The answer to that question is, it depends. What are you attempting to accomplish?
If you want to know what you can actually expect the maximum SPL of the loudspeaker to be in a particular application, you should probably use the Advanced GLL (mode). These simulation results should correlate well with in-situ measurements, provided the measurement conditions are reasonably similar to the parameters and conditions of the model/simulation.
If you want to compare the maximum SPL of a newer Biamp loudspeaker to an older model, or to a loudspeaker from another manufacturer, you should probably use the Nominal GLL (mode).
Both types of GLLs can be used in an EASE® model or Focus project to facilitate both types of investigations/comparisons.
So how does one know if a GLL for another manufacturer’s loudspeaker uses maximum input/output data more like the Nominal or the Advanced GLL (mode)? Loudspeakers that use similar designs with similar components should perform in a similar manner. As an example, two different two-way loudspeakers, each with a 12-inch woofer and a high-frequency horn with a 1.4-inch compression driver, will probably have similar maximum output capabilities. Sure, there might be some differences. However, if the maximum SPL of loudspeaker A is more than about 2-3 dB higher than loudspeaker B, chances are that loudspeaker A was specified differently; especially if the sensitivity of both loudspeakers is about the same. So, use the Nominal GLL (mode) for maximum SPL comparison with loudspeaker A.
The Gory Details
The maximum SPL output capability of a GLL is determined by basically three things; the loudspeaker’s sensitivity/frequency response, the maximum input voltage of the drivers used in the loudspeaker, and the spectral content of the input signal. This is a bit of an oversimplification, but it should suffice for this discussion. It is this second item, the specification of maximum input voltage of the drivers, that is most often the reason for GLLs of very similar loudspeakers to show very different maximum SPL capabilities.
Historically, the maximum SPL has been calculated for most loudspeakers by adding the power handling of the loudspeaker (after converting to dB) to the sensitivity. An example of this for a 200 W loudspeaker with a 90 dB sensitivity would be as follows.
90 dB SPL + 10*log(200) = 90 dB SPL + 23 dB = 113 dB SPL
This is not entirely accurate, but it can provide an easy way to get an answer. Many manufacturers use this same method when creating their GLLs. For multi-source GLLs, they will take the power handling of each driver (or pass band), calculate the equivalent voltage, and enter them into the GLL. For a single-source GLL, they will use the power handling of the loudspeaker system as a whole and enter that equivalent voltage into the GLL. Doing this presents two major problems.
Problem 1 – Actual Performance at High Input/Output
The first is that the method for determining the power handling of drivers and systems is based only on survivability (the DUT, device under test, can’t fail/be destroyed during testing) and then a comparison of the performance (most often frequency response & impedance) pre and post testing, after a cool down time of course. This test method is completely blind to the performance of the DUT during the test.
A better way to rate the MIV (maximum input voltage) of a driver or system is to monitor the response of the DUT during the test. When the response has changed by a certain amount, usually about 3 dB, the voltage and SPL are recorded and the test is halted. This is the method prescribed in the AES2-2012 [1] standard; specifically, section 5.13 Method of measurement for maximum usable continuous output sound-pressure level. The results using this method are
typically 2-5 dB lower than the results of a power handling test. This is one reason it might not get used much. However, employing this method for determining the MIV of drivers or systems yields maximum SPL simulation results that correlate much better with the measured maximum SPL of loudspeaker systems when in use with actual or simulated program material.
Problem 2 – Which Power Handling: Continuous/Program/Peak?
The second problem is which power handling number gets used in the GLL. As we detailed in Problem 1, typically none of the power handling numbers are the right ones to use. However, if one were to be used, the Continuous power handling would be the least worst.
Modeling programs that use GLL files are intended to simulate the continuous (rms) SPL capabilities of a system. For this reason alone, it is more appropriate to use the Continuous or AES power handling value. Using a Program power rating, which is typically twice the Continuous value, will result in an over estimation of the maximum SPL by another 3 dB.
There are even some GLLs that use the Peak power handling value. This causes the maximum SPL to be inflated by 6 dB over the Continuous rating. This would seem to be quite misleading.
So, how can one tell if a GLL is using something other than a Continuous power handling value? If it’s a single-source GLL this can be straightforward, but a bit of math is required. Open the GLL in the GLL Viewer [2] (free from AFMG). Calculate the GLL using “IEC 60268-1” as the Input Signal selection. Then click on Graphs > Input Voltage as shown in Figure 1. The upper-right corner of the graph will show the maximum input voltage level. Convert this to voltage and then to power following the example equations below.
For this example, the GLL shows a maximum voltage level equivalent to about 500 W. If the Program power handling on the spec sheet is 500 W the GLL will overestimate the maximum SPL by at least 3 dB, maybe more.
A multi-source GLL might not have the exact same equivalent voltage corresponding to a power handling rating shown on the loudspeaker’s spec sheet. This is due to the more accurate use of separate sources in the GLL to model the different drivers in the real loudspeaker. Examples of a multi-source GLL are shown for both the Nominal and Advanced modes of the GLL for the Desono EX-S8 loudspeaker.
The table shows that the Advanced mode maximum voltage is the same as the spec sheet rating, within rounding errors. The Nominal mode, based on the power handling of the individual LF and HF drivers, shows that the maximum voltage (and equivalent power) of the system is a bit higher than the values published on the spec sheet. It is reasonably close to the Continuous power handling but not exactly the same. Note that a different input signal for the
GLL might result in a different maximum voltage, and subsequently different maximum SPL.
Summary
The maximum SPL output capability predicted by a GLL will be dependent on how the maximum input level is specified within the GLL. If the right values are entered for the maximum input voltage, determined using modern methods and measurement techniques, then the maximum SPL output prediction should be reasonably accurate. If the values are based on Continuous power handling, no matter how well-intentioned, the maximum SPL capability will most likely be overestimated. If the values are based on Program or Peak power handling, the maximum SPL capability shown by the GLL is blatantly incorrect. The actual SPL from the loudspeaker will never reach these levels. The loudspeaker’s peak sound level might hit them but that will have little correlation to the subjective loudness of the loudspeaker system.
This table gives an overview of the inflated-SPL landscape for GLLs.
References
[1] Audio Engineering Society, "AES2-2012: AES standard for acoustics - Methods of measuring and specifying the performance of loudspeakers for professional applications," [Online]. Available: https://www.aes.org/publications/standards/search.cfm?docID=12.
[Accessed 10 Feb 2022].
[2] Ahnert Feistel Media Group, "GLL Viewer Download," [Online]. Available: https://www.afmg.eu/en/gll-viewer.
[Accessed 10 Feb 2022].
“EASE®, FIRmaker® and AFMG® are registered trademarks of AFMG Technologies GmbH.”