Technical Report: What is a Sound Level Meter? (Part 5)

Figure 8-1 below shows a block diagram of the electrical circuit of a sound level meter. Pay particular attention to the relative positions of the AC output (AC out) and DC output (DC out) in the block diagram. Understanding the difference between these two outputs is important for understanding and effectively using a sound level meter.

High-end stereo systems faithfully reproduce the original sound from soft to loud, and from deep bass to high frequencies. To faithfully reproduce the original sound, the original sound must be captured accurately; otherwise, the reproduced sound will differ from the original. A microphone accurately captures this sound and converts it into an electrical signal, and a preamplifier amplifies this picked-up minute electrical signal to a certain level and performs impedance conversion.

Sound is a wave phenomenon in which alternating periods of high and low air density are transmitted. The microphone and preamplifier of a sound level meter are sensors that convert this air density (instantaneous sound pressure) into an electrical signal, and therefore require high sensitivity and good frequency characteristics.

There are three main types of microphones: condenser, dynamic, and ceramic. For sound level meters, condenser microphones are used because they can be made into a small diameter shape which is acoustically advantageous, have a flat frequency response over a wide frequency range, and are more stable than other types. The structure of a condenser microphone is illustrated above. There are two types of condenser microphones: bias type and back electret type. The difference is mainly whether a DC voltage is applied to the diaphragm or whether a permanently electrically polarized polymer film is used instead of applying a voltage. Generally, bias type microphones have the advantage of higher sensitivity. In addition, self-noise and temperature stability are important factors for microphones used in sound level meters.

Measuring noise requires absolute measurement of sound pressure levels within the audible frequency range. However, the sensitivity of the ear varies with frequency, and simply taking the RMS value of sound pressure does not represent the perceived loudness of the sound.

Chapter 6, Section 4 describes the equal-loudness curves (ISO 226), which represent the sound pressure levels at which sounds are perceived as having the same loudness at different frequencies. The A-weighted curve in Figure 8-3 approximates the perceived loudness of quiet sounds, and the C-weighted curve approximates the perceived loudness of loud sounds. In the past, these characteristics were used interchangeably depending on the loudness of the sound. Subsequent research has shown that while these characteristics represent the perceived loudness of sound, they are not suitable for representing the loudness of noise, and it is better to use the A-weighted curve even for loud sounds.

Modern sound level meters are equipped with A-weighting, C-weighting, and Z (or FLAT) weighting, which provides a flatter frequency response. In particular, A-weighting is always used for measuring noise levels. Since fluctuating noise cannot be accurately measured in a single measurement, the equivalent sound level LAeq (calculated by taking a long-term energy average) or the time-weighted sound level LX (calculated from the cumulative frequency distribution) are used. C-weighting, with its relatively flat frequency response, is used when recording the AC output of the sound level meter or when measuring impact sounds (which have a wider frequency range). Z (or FLAT) weighting provides an even flatter frequency range than C-weighting, allowing the sound level meter to be used as a general-purpose acoustic sensor by utilizing its AC output and acting as a microphone and amplifier with guaranteed frequency response.

Conventional sound level meters defined a so-called "flat" frequency response without frequency weighting as the "FLAT characteristic." However, the specific specifications, such as the frequency range, were left to the manufacturers, so the FLAT characteristics of commercially available sound level meters were not, in practice, unified. Therefore, the new standard defines a new "Z characteristic," specifying its frequency range as "flat from 10 Hz to 20 kHz." However, considering tolerances and other factors, in practice, it is probably acceptable to consider the FLAT characteristic and the Z characteristic as equivalent, as before.

Time-weighted characteristics refer to specifications concerning the movement of indicator meters (including digital displays). Specifically, they correspond to the time constant for averaging in the RMS detection circuit shown in Figure 8-4. These dynamic characteristics (time constants) are described in detail in JIS C 1509. There are fast dynamic characteristics (F [Fast]: 125 ms) and slow dynamic characteristics (S [Slow]: 1 s). The Fast characteristic approximates the time response of the ear, while the Slow characteristic is used to indicate the average level of fluctuating noise. Normally, the fast dynamic characteristic (Fast) is used for measuring noise. Furthermore, since the magnitude of impact sounds cannot be accurately measured even with Fast, a time-weighted characteristic I (Impulse = 35 ms <rise>, 1.5 s <fall>) is provided for such cases. However, recent research has shown that this weighted characteristic I (Impulse) is "not very suitable for evaluating impactful sounds," and it has been removed from the IEC (and JIS) standards. Instead, there is a tendency to use the peak value of instantaneous sound pressure as a parameter for evaluating impact noise.

Figure 8-5 shows the level changes when a 1 kHz tone burst signal (burst duration 200 ms, repetition time 3 s) is input to the time constants Fast and Slow.

The RMS detection in Figure 8-4 is equivalent to passing the squared signal of the instantaneous sound pressure through the first-order RC low-pass filter shown in Figure 8-6. The time constants of the dynamic characteristics (Fast and Slow) described above correspond to the time constant of this low-pass filter (τ = RC). Figure 8-7 shows the rising and falling edge response waveforms when a tone burst signal is input to this equivalent circuit in the same way as in Figure 8-5. As shown in this figure, if the RMS value of the steady-state input sine wave is e _i, then the time constant τ is:

It will be ;

Note that Figure 8-5 uses dB units for the Y-axis, so it may look different from Figure 8-7. However, if you draw it on a linear scale in the same way (Figure 8-8), you will be able to better understand the difference between the response waveforms for fast and slow dynamic characteristics.