Technical Report: What is Sound Quality Evaluation? 1

In household appliances such as air conditioners and refrigerators, office machines like copiers, and automobiles, there has been a demand for comfort through low noise levels alongside the pursuit of basic performance. Traditionally, noise evaluation for such devices relied on physical quantities like sound pressure levels and acoustic power; however, since the final judgment of noise is left to human hearing, psychological acoustic evaluation metrics related to these physical quantities have started to be used. Psychological acoustic evaluation metrics quantify how humans feel when they hear sounds, making it easier to gain insights into how to improve sound quality. The method of using psychological acoustic technology to provide feedback into the design of devices is called acoustic design.

Sound quality evaluation is a technology that uses psychoacoustic techniques to quantitatively analyze sound according to human perception. In contrast to traditional mechanical evaluations based on physical quantities, this sound quality evaluation technology is believed to add humanity to future living environments. For example, current air conditioners have become much quieter than before, but their sound is still mechanical. However, if sound quality is analyzed using sound quality evaluation technology and advancements in sound improvement techniques are made, it could be possible to transform it into pleasant, gentle natural sounds. This sound quality evaluation technology is also expected to have future applications in various fields, such as vehicles and office equipment.

Sound quality evaluation based on psychoacoustic technology was initially used to assess the interior sound of automobiles.

Loudness, which is a measure of sound intensity, is standardized by ISO. Psychoacoustic evaluation metrics are gaining attention not only in the automotive industry but also in other sectors. In 1998, standards were established in the home appliance industry regarding noise labeling, which specifies the use of loudness in addition to noise levels. Furthermore, it is also recognized as a quantitative evaluation metric for noise control in industries such as office equipment.

Psychoacoustic technology as an academic discipline has been researched since the 1930s, as shown in the table on the next page. Loudness, one of the psychoacoustic evaluation metrics, has been standardized by ISO (for steady sounds only) and has begun to be widely used.

This loudness has been developed through the accumulation of research by various people and has been refined to its current state by Zwicker in Germany. Many other psychological acoustic evaluation metrics have been devised relatively recently, most of which have been proposed by researchers associated with Zwicker's laboratory. Currently, several parameters are still being researched, and this technology is expected to continue to develop further.

Psychoacoustic evaluation metrics are derived from studies of the structure of the human ear and the functioning of auditory nerves, as well as from conducting listening experiments with many individuals. Listening experiments involve playing sounds to people and asking them how they felt about those sounds, thereby examining their responses. Human responses to sound can vary slightly based on factors such as age, past experiences, physical condition at the time, temperature, and surrounding environment. Therefore, psychoacoustic evaluation metrics have been established by conducting numerous tests with many individuals and statistically processing the results.

So, why is psychoacoustic technology necessary? Traditionally, sound generated by equipment has been evaluated and addressed based on physical quantities (such as sound pressure level, noise level, power level, 1/3 octave spectrum, FFT spectrum, etc.). However, it is humans who actually hear the sound. Moreover, while physical sound mitigation generally involves reducing these physical quantities, there are limitations to this approach. Of course, it is not impossible to completely eliminate sound by covering it with very expensive, large, and heavy materials, but this is not practical. Additionally, eliminating sound does not necessarily create a pleasant environment. Therefore, instead of simply lowering physical quantities like sound pressure, what is quantified is how humans feel when they hear the sound, which is referred to as psychoacoustic evaluation metrics. Since psychoacoustic evaluation metrics are related to physical quantities, it becomes easier to gain insights into how to improve sound quality from these metrics. This method of providing feedback into the design of equipment is known as acoustic design.

To help you understand sound quality evaluation, we will present an actual example of analysis. The figure below shows an analysis of six machine sounds with varying volumes.

We conducted an analysis of the characteristic sound pressure level and loudness of this sound, similar to the previous example.

In the upper line (green) analyzed by sound pressure level (noise level), the values for the two motorcycles are the same. In contrast, in the lower line (pink) analyzed by 'loudness', the sound of the second motorcycle passing is louder. When actually listening to these sounds, the second motorcycle is indeed perceived as louder, consistent with the loudness results.

This time, there are two machine sounds.

The second sound is heard as louder than the first sound. However, when examining the A-weighted sound pressure levels, the value for the latter sound is smaller. This is an example of a situation where the relationship of loudness assessed by A-weighted sound pressure levels is reversed compared to what is actually heard.

So, why does the value of sound pressure not correspond to the actual perceived loudness of sound?