Bruel & Kjaer: Electro Acoustic Measurements


Bruel & Kjaer - Electro Acoustic Measurements (AN16-035).pdf (


Why Make Electro Acoustic Measurements?

Frequency response, phase response, harmonic distortion, intermodulation distortion, transient intermodulation distortion, directivity index, efficiency, sensitivity the list of possible measurements is almost endless Too often the mass of data assembled leads to confusion, and not insight about the quality of the product being measured Therefore, the fundamental problem for the electro-acoustic engineer is to select the appropriate measurements and evaluate the results by giving the proper weight to the various measurements He must then translate the objective results into the subjective domain using understandable and relevant terms However, it is an art to do this It requires a great deal of talent, insight, instrumentation, and time Only after a wide variety of measurements have been made is it possible to obtain "the complete picture."

In practice, the objective and subjective evaluations must work together to improve the final product Objective measurements alone are not enough, subjective judgments alone do not suffice They must be combined Then the objective measurements will lead to an improved product, and the subjective judgments will suggest improved, perhaps even new objective measure ment techniques The purpose of all these measurements and subjective judgments is an improved product in practice, that means higher fidelity, a better loudspeaker, a superior microphone, an improved tape recorder And in the final analysis, whether or not the product has been improved will be judged subjectively But objective measurements will have played an important role in achieving the desired end result

These notes will concentrate on a broad range of measurements and will emphasize the dual application of most techniques, to acoustical and electrical measurements The notes are also structured along a line beginning with simple and then moving to more complex instrumentation The Bruel & Kjaer system is a building block system, where the addition of one or more instruments to the basic combination permits new measurements This avoids duplication of equipment and permits advanced measurements to be obtained for a relatively small marginal cost.


What Measurements are Relevant?

Only those measurements that in some way, directly or indirectly af feet the subjective quality evaluation can be considered relevant Here we will deal with the primary ones in brief before discussing detailed instrumentation techniques Frequency Response The most common electrical and acoustical measurement is frequency response The pure sine frequency response is relevant for the electrical part of the system But when loudspeakers or other acoustical transducers are measured in an anechoic room, the pure sine excitation is not sufficient since the influence of the acoustical load, the listening room, is ignored.

Therefore, additional measurements such as sound power and measurements in a "typical" room should be made Recent investigations have shown that frequency response measurements in the actual listening room using third octave, pink weighted, random noise, along with power response and phase response give the best correlation to subjective judgments The main problem is defining a typical room, and it is doubtful whether this will ever be solved because of the wide variety in room dimensions depending on building type and also the country.

However, if a typical room could be agreed upon, it would permit comparison of measurements from different manufacturers, and more important, would make it easier to predict the performance of a loudspeaker in a given consumer's room. Lacking this standard room, it would stili be valuable for a manufacturer to make his own standard room with well-defined acoustical characteristics. However, anechoic facilities are still necessary to evaluate the many different phases of loudspeaker design. Fortunately, it has recently become possible to synthesize anechoic conditions in a normal reflective room by the use of gating techniques (see section 5). Thus the combination of anechoic and normal room measurements should yield valuable results.

Phase Response Frequency response alone tells us little about the transient response of the system. But when combined with the phase response, the transient response can be predicted. Traditionally, transient response has been measured — or rather evaluated — using tone bursts, but now a more compiete picture can be obtained since phase can be measured as easily as frequency response. The more linear the phase response, the better the response of the system to transients such as found in percussion instruments, the attack of a trumpet, the pizzicato of violins, etc. Distortion These notes will also describe measuring techniques for various types of distortion, all of which have audible relevance. There is total harmonic distortion, individual harmonics measured separately, intermodulation and difference frequency distortion, and transient intermodulation distortion (TIM). Since distortion measurement at a single or at spot frequencies may give non-representative results, all of the measurement techniques described here will permit swept measurements of most types of distortion over the 2 Hz to 2OOkHz range.

Frequency response, phase response, and distortion: these are probably the three most important categories of measurements. But in reality, all three categories are various types of distortion. If the frequency response is not flat, the amplitude will be distorted, and if the phase response is not linear, a time distortion will result. In fact, virtually all measurements seek to detect distortion of the original waveform in some manner or another.

For example, wow and flutter results in frequency distortion, and noise adds to the waveform, thus distorting it. As the various techniques for measuring these types of distortion are described in the following chapters it is hoped that insight will be gained, not only in better measurement technique, but in the art of combining the mass of objective measurements in a sensible manner and transforming them to the subjective domain. If this goal is attained, the ultimate purpose of obtaining better high fidelity can be approached.