Several months later, having received no complaints, I went back, mostly just to enjoy the fair, but also to check out the sound systems, half expecting to find my equalizers bypassed or replaced by third-octave graphics.
Instead, everything was as I had left it. I suspect that there were several attempts to get better results with graphics, but apparently everyone decided that my equalizers sounded better. (I didn’t press the issue!)
However, I inspected the console EQ settings closely. Both music channels had only a slight boost in both the bass and treble.
The live announce channel, however, showed strong settings in all three bands.
Rather than try to figure out how these settings and the mic response might combine, I decided to just measure the overall result.
Figure 3. Frequency response measured with a flat-response microphone (with no equalization) and the announcer’s microphone (with mixer equalization settings preferred by the system operators) at the same location. (click to enlarge)
My flat-response mic was placed at an arbitrary location in the seating area and recorded the response with the console EQ bypassed. Then I substituted the announce mic in the same location, aimed at the loudspeakers, with its console EQ settings restored, and ran another response. (Figure 3)
Several features are noteworthy. First, although not noted on the chart, it’s apparent that I made these measurements with my equalizer bypassed, as the responses are very close to those of the original condition.
A more surprising observation, though, is how closely these two traces track together. This means that the console EQ settings nicely flattened the response of the mic cartridge.
It also means that this flat response is what was perceived and preferred by the system operators. Interesting!
What does all of this teach us?
—Valid measurements can yield very useful information. This information can accurately describe the timbre/tonal balance of a system, and provides this description in at least as much detail as our ears can detect.
It shows how and how much the performance deviates from the ideal, and further, reveals this at any given location, or many locations, detailing how the performance varies from one location to another.
The information also suggests specific physical causes for performance deficiencies and the means to correct or reduce them. It verifies corrective efforts and provides new information for further improvement, and distinguishes between multiple shortcomings and causes, which may be impossibly confusing to the ear.
Finally, it provides documentation that can be used for future reference or comparison. This can be of a quality and quantity far exceeding what can be obtained just by listening.
And used properly, the information significantly improves the competence and effectiveness of any sound professional.
—Valid measurement results can be costly to obtain. The necessary test equipment costs thousands of dollars, and it’s surprisingly difficult to find equipment capable of running all tests needed to provide the desired information.
Significant learning and experience are necessary for competent results. It’s easy to fail to obtain the needed information or to obtain incorrect or inaccurate information.
Even if the correct information is obtained, it’s just as easy to misinterpret it or overlook some of its significance - or to misunderstand, or fail to understand at all - what corrective actions should be taken.
Any of these mistakes can lead to incorrect actions that result in system performance that is insufficiently better, or even worse, than before.
Such an experience can lead to the conclusion that measurements are worthless, and, therefore, to the failure to utilize a valuable resource.
On the other hand, using this resource can result in costly damage to cherished opinions and beliefs, and even to the artistic ego. Be warned!
—Many other measurements need to be made. While the overall frequency response is almost certainly the characteristic most important to perceived sound quality, there are many others that should not be ignored.
These include maximum sound level, overall coverage uniformity, freedom from feedback, and factors affecting reliability - all warrant extensive discussions of their own.
In addition, the characteristics examined here are not fully measured by the means discussed; new techniques are always needed. Most particularly, there is currently no realistic way to measure the nature and uniformity of the direct sound field in a reverberant environment.
And it’s not at all clear that we’re correctly interpreting or understanding the measurement results. Sadly, little real research has been done in this area.
However, there is one thing that’s pretty clear: the flatter or smoother or more uniform the measurement results, the better the system will sound. Measure well and prosper.
Final note - I still have, and occasionally use, the GenRad gear used to make the measurements shown in this article. Even though it dates back half a century, this unit is still capable of making many important measurements accurately. But it’s also large and heavy. For the past 18 years, I’ve employed a Neutrik Audiograph, which performs very comparable tests and is an order of magnitude smaller and lighter. I’ve yet to find a computer-based configuration that can adequately perform all tests I consider essential.
Bob Thurmond began his professional audio career in 1959 as a recording technician. He became a consultant in 1965, specializing in sound reinforcement and formed his own company, G.R. Thurmond and Associates, in 1971.
Using instinctive knowledge to judge the loudness of a system
