Sound Measurement and Noise Metrics

Quantifying noise pollution requires measurement techniques and metrics that capture both the physical characteristics of sound and its potential impact on human recipients. The decibel scale, logarithmic by nature, provides the foundation for most noise measurements, but various weighting networks and time-averaging procedures adapt basic sound pressure level measurements to better predict human response to different types of acoustic environments.

A-weighted sound levels (dBA) represent the most commonly used noise metric because the A-weighting network approximates human hearing sensitivity across different frequencies. The A-weighting curve reduces the contribution of low and very high frequencies while emphasizing the midrange frequencies where human hearing is most sensitive:

LA = 10 log₁₀(∫[p(f)²WA(f)²df]/p₀²)

Where p(f) is the frequency-dependent sound pressure, WA(f) is the A-weighting function, and p₀ is the reference pressure (20 µPa). This approach provides better correlation with perceived loudness than unweighted measurements, though it may underestimate the impact of low-frequency noise that can be particularly annoying even when not perceived as loud.

Equivalent continuous sound level (Leq) addresses the temporal variation in noise levels by computing the constant sound level that would contain the same acoustic energy as the varying noise over a specified time period:

Leq = 10 log₁₀[(1/T)∫₀ᵀ (p(t)/p₀)² dt]

This metric proves essential for evaluating environmental noise where levels fluctuate significantly due to intermittent sources like traffic, aircraft, or construction activities. Typical integration periods range from 1 minute for detailed analysis to 24 hours for community noise assessment.

Statistical noise metrics capture the distribution of noise levels over time using percentile values. L₁₀ represents the sound level exceeded 10% of the time (typically associated with peak noise events), L₅₀ represents the median level, and L₉₀ represents the background level exceeded 90% of the time. These metrics help characterize different aspects of the acoustic environment:

- L₁₀: Peak events that may cause maximum annoyance - L₅₀: Typical exposure levels for most activities - L₉₀: Background levels that affect quiet activities and sleep

Day-night average sound level (Ldn) accounts for increased sensitivity to noise during nighttime hours by applying a 10 dB penalty to noise occurring between 10 PM and 7 AM:

Ldn = 10 log₁₀[(15 × 10^(Ld/10) + 9 × 10^((Ln+10)/10))/24]

Where Ld is the daytime average level and Ln is the nighttime average level. This metric reflects research showing that people are more sensitive to noise disruption during sleeping hours, when ambient levels are typically lower and sleep disturbance can have significant health impacts.

Noise exposure level (SEL) quantifies the total acoustic energy in discrete noise events by normalizing them to a one-second duration:

SEL = 10 log₁₀[∫₀ᵀ (p(t)/p₀)² dt]

This metric proves particularly useful for evaluating individual aircraft flyovers, vehicle pass-bys, or other transient noise events that can be characterized by their total energy content regardless of duration. Multiple SEL values can be combined to predict cumulative exposure from repeated events.

Specialized metrics address specific noise characteristics that affect human response. Tonality indices measure the prominence of pure-tone components that can be particularly annoying even at moderate levels. Impulsiveness metrics quantify the suddenness of noise events that create startle responses. Fluctuation indices describe the temporal variability that can affect subjective annoyance and interference with activities like conversation or concentration.