Quality Paper

"What is Good Sound Quality as defined

by the Sound System Design Industry?"

by Stephen Edmonds

Date revised: October 16, 2000

     When taking a survey from people on what they consider "Good Sound Quality" to be, the answers received are often hard to measure with test equipment. For every person asked a slightly different definition results. Some want the sound to have a lot of "punch" in the low end. Others want the high end to "sparkle". Still others want the system to have "warmth" in the midrange. So how does the sound system design industry design auditorium sound systems that will satisfy all the different people that will come into the auditorium to listen?

     The first and most obvious approach to solving this problem is to design for the owner or person paying the bill for the system. This solution seems unsatisfactory for many situations because it relegates the rest of the audience to listening to the owner's definition of sound quality. If the owner is not present most of the time, or is a large group of people, as in the case of a city auditorium or a church sanctuary, this solution may yield a sound quality that is considered horrible by a large percentage of the listening group. If the simple approach of satisfying the person paying the bill is not chosen then what method is left for designing a quality sound system that will satisfy a multitude of people?

     The industry's answer to this question is to design a sound system to accurately reproduce the source being input into the sound system (such as a singer's voice or band's musical instruments) (1). The goal here is to make the sound system as transparent as possible between the listener and the performer. The audience's quality judgement of the experience is then not decided by the sound system's "Good Sound Quality" as defined by someone, but by the quality of the audio performance itself. If the sound system is as accurate as is technologically possible, then if the performance had "punch" it would be conveyed, if the performance "sparkled" it would be conveyed, and if it had "warmth" it would be conveyed.

     The professional sound industry has a few standard measurements that are made using test equipment to determine the quality of a system's accuracy.

     The first is the polarity relationship of the loudspeakers. This is a measure of the direction the speaker elements move when an electrical signal is applied. If all the loudspeakers elements are not all working in the same direction, odd frequency response problems develop throughout the listening area. These problems result from addition and/or cancellation of sound waves that are out of polarity with each other (2). The polarity of a speaker can be checked using a polarity checker. This device puts a forward moving pulse through the sound system and the direction of movement of the speaker is checked at each speaker with a microphone built into the receiving unit.

     Once the speakers are in the same polarity orientation the next important quality measurement for system accuracy is the measurement of a system's frequency response. This measurement is done by taking a pink noise generator (a test signal) and placing it as the input to the sound system. Once the sound from the generator is heard through the speakers a measurement device called a Real-Time Audio Spectrum Analyzer is used to "listen to the system" with its calibrated microphone.

Figure 1: Audio Control's Real Time Analyzer           

     This device displays the amplitude of thirty-one, one-third octave, bands in the audio range, The desired response on the display at the room's listening position (one hundred feet from the speaker enclosure) is a flat line to approximately one kilohertz, and should roll off at a rate of three decibels per octave above one kilohertz (1). The reason for the three decibel per octave roll off at the listening position is to compensate for a natural effect caused by the moisture content of air. It is the effect that makes close thunder crack and distant thunder rumble. If the system were tuned perfectly flat at the one hundred foot mark then the sound at that mark would be "brighter" than the unamplified sound at the same distance. The reason the measurement is made at the listening position is to account for the acoustical effect of the room itself on the reproduced system sound (1). If the analyzer is carried around the listening area, the frequency response should be similar to the response at the one-hundred-foot mark (1). The tolerance for the above frequency response measurements is plus or minus three decibels. The reasoning behind this tolerance decision is the sensitivity of the human ear based on hearing tests (1). The hearing tests show that the average human ear can detect that a change has occurred when the level is increased or decreased by three decibels (1). So, if the frequency response is within three decibels of the reference level, the human ear will not notice that it is not ideal.

     Another important aspect of a sound system being transparent is the amount of noise that is generated by the system itself. All electronic equipment has noise in it (3). This noise comes from the molecular activity of the components in any electronic device. The signal-to-noise ratio of the electronic devices is an important factor in how noisy a system is when there is no signal going through it. This signal to noise ratio measurement is usually made by the product manufacturer for each piece of installed equipment. Their measurement techniques conform to a specific set of rules for measuring signal to noise ratios. The higher the signal to noise ratio the quieter the component is. A typical value for this number is ninety-five decibels between the test signal and the noise in the circuit. One component with a bad signal to noise ratio is detrimental to the entire system's quality.

     The final measurement that can be made on a sound system to determine its accuracy is a measurement of the percentage of articulation loss of consonants in speech from the person on stage to a listener in the audience (1). This measurement tests not only the sound system but the room itself for accurate reproduction of the spoken word. The classical way of performing this test is to have several people spread out and sit in the audience area. Then a single person gets up on stage and reads a list of one hundred unrelated words. The people in the audience write down the words as the speaker says them. In a good quality situation the people in the audience should not have gotten more than ten words incorrect in the process. The outer acceptable limit for this test is missing or misunderstanding fifteen out of the one hundred words, I have heard of an electronic means for performing this test, known as RASTI. However I was not able to locate any information on this method of testing.

If a system passes all of the above tests then the sound design industry considers that system to be a good quality sound system. These criteria are, according to the industry, required for long term acceptance of a sound system

Works Cited

(1) Davis, Don & Carolyn. Sound System Engineering. (This is our sound system design Bible)

     Howard W. Sams & Co., Inc: Indianapolis Indiana, 1979.

(2) Blake, Roy. Basic Electronic Communication,

     West Publishing Company: St, Paul NIN, 1993,

(3) Bogart, Theodore F. Electronic Devices and Circuits.

     Macmillian Publishing Company: New York NY, 1993