-3dB Points, Bandwidth, and Octave Relationships
Frequency response describes how a sound system or component responds to input signals across the frequency spectrum. It indicates which frequencies the system reproduces at what relative levels, revealing the system's tonal character and capabilities. A perfectly flat frequency response reproduces all frequencies equally, while variations from flat response create the tonal coloration that distinguishes different sound systems.
Frequency response is measured by applying a test signal (pink noise, sine sweep, or impulse) to the system and measuring the output with a microphone or measurement system. The resulting data shows how output level varies with frequency, typically displayed as a graph with frequency on the horizontal axis (logarithmic scale) and level in decibels on the vertical axis.
No real sound system achieves perfectly flat response in all conditions. Room acoustics, speaker design limitations, crossover interactions, and numerous other factors create response variations that must be understood and managed. The goal is not necessarily flat response in every installation, but rather appropriate response for the application and consistent, predictable behavior.
The bandwidth of a system describes the frequency range over which it operates effectively. The standard definition of bandwidth uses the -3 dB points—the frequencies where output power drops to half the level at the reference frequency.
-3 dB represents the point where output power falls to approximately 50.1% of the reference level (or voltage to approximately 70.7%). This level difference is generally considered the limit of perceptible flat response—deviations beyond -3 dB are typically noticeable as reduced bass or treble response compared to the reference level.
Low-frequency limit (-3 dB point) indicates how low the system can reproduce bass. A system specified as extending to 50 Hz (-3 dB) produces useful output at 50 Hz but rolls off progressively below that frequency. A system that reaches 30 Hz (-3 dB) provides deeper bass extension.
High-frequency limit (-3 dB point) similarly indicates the upper frequency boundary. Human hearing extends to approximately 20 kHz, and systems that extend well beyond this limit may not provide audible benefit. However, some rolloff within the audible range (such as -3 dB at 18 kHz) may be perceptible as reduced sparkle or air.
Bandwidth calculation uses the ratio between upper and lower -3 dB frequencies divided by the center frequency to determine fractional bandwidth. A system with 100 Hz lower limit and 10 kHz upper limit has a bandwidth ratio of 100:1, while a system from 200 Hz to 8 kHz has a 40:1 ratio. These ratios help compare systems and understand relative broadband versus narrowband behavior.
| Specification | Meaning | Perceptual Effect |
|---|---|---|
| 50 Hz - 15 kHz (-3 dB) | Extended bass, limited treble | Full but slightly dull |
| 45 Hz - 18 kHz (-3 dB) | Extended both ends | Full and detailed |
| 80 Hz - 12 kHz (-3 dB) | Bass-limited, moderate treble | Lacks deep bass, clear mids |
| 20 Hz - 20 kHz (±3 dB) | Full range reference | Flat and accurate |
An octave represents a doubling or halving of frequency—the musical interval between notes that sound most alike, separated by eight scale degrees. Octave relationships are fundamental to understanding acoustic behavior, filter slopes, and frequency response specifications.
Octave numbering starts at A4 = 440 Hz as the reference. A3 is 220 Hz (one octave below), A5 is 880 Hz (one octave above). Each octave span represents the same musical interval and similar acoustic wavelength ratios, making octave-based analysis intuitive for musical and acoustic applications.
Octave band filters divide the frequency spectrum into bands one octave wide, with center frequencies at 31.5, 63, 125, 250, 500, 1000, 2000, 4000, 8000, and 16000 Hz. These standardized bands approximate how human hearing groups frequencies and are the basis for most acoustic measurement and noise rating systems.
Third-octave bands provide three times the resolution of full octaves, with center frequencies at 50, 63, 80, 100, 125, 160, 200, 250, 315, 400, 500, 630, 800, 1000 Hz and continuing similarly to 16 kHz. This finer resolution reveals more detail about frequency response behavior and is standard for professional audio analysis.
Slope specifications use octaves to describe how quickly response falls off. A crossover slope of 12 dB/octave means response decreases by 12 dB for each octave increase in frequency away from the crossover point. A 24 dB/octave slope drops twice as fast. Understanding octave-based slopes helps predict system behavior in crossover regions and cutoff frequencies.
Human hearing typically spans from approximately 20 Hz to 20,000 Hz (20 kHz), though this range narrows with age, exposure to noise, and individual variation. Understanding the hearing range helps put frequency response specifications into perceptual context.
Low-frequency hearing below 20 Hz is felt more than heard. While we cannot consciously perceive discrete tones below approximately 12-16 Hz, we clearly feel the impact of very low-frequency sounds in our bodies. Bass guitars, kick drums, pipe organs, and explosions all produce significant energy below the audible threshold that affects our perception of the sound.
High-frequency hearing above 10 kHz is typically reduced in most adults and significantly diminished in many people due to noise exposure, aging (presbycusis), and other factors. The commonly cited 20 kHz upper limit is rarely achieved in practice, and the ability to perceive tones above 15-16 kHz varies significantly between individuals.
Sensitivity variation across the hearing range means we don't hear all frequencies equally loudly even when they're at the same SPL. The ear is most sensitive between 1-4 kHz, where speech consonants and musical presence reside, and progressively less sensitive at frequency extremes. This variation is why A-weighting and other frequency weightings are used to correlate measurements with perceived loudness.
Accurate frequency response measurement requires appropriate equipment and technique to capture meaningful data that can guide system optimization.
Measurement microphone quality significantly affects measurement accuracy. Measurement-grade microphones (often called "measurement mics") have flat, calibrated frequency response and consistent polar patterns. Consumer microphones may have significant response variations that invalidate measurements. Professional acoustic measurement systems use Type 1 or Type 2 measurement microphones meeting IEC 61672 standards.
Test signals for frequency response measurement include pink noise (equal energy per octave band, approximating music's spectral distribution), swept sine tones (pure tones varying continuously through the frequency range), or maximum length sequences (MLS) that provide impulse response data. Each method has advantages for specific measurement scenarios.
Measurement position must be consistent for valid comparisons. Near-field measurement close to the speaker driver captures direct sound with minimal room influence, suitable for individual driver evaluation. Far-field measurement at the listening position captures the complete system behavior including room effects but with greater position sensitivity.
Averaging multiple positions provides more representative data for room acoustics than single-point measurement. A typical approach measures at the primary listening position plus several nearby positions, then averages the results to identify consistent features versus position-dependent anomalies.
Frequency response specifications appear on equipment from every manufacturer, but their meaning and accuracy vary dramatically.
±3 dB tolerance specifications indicate the maximum deviation from the reference level across the stated frequency range. A system specified as 50 Hz - 15 kHz (±3 dB) may have up to 6 dB of variation (3 dB above and 3 dB below the reference) at any frequency within that range. Some manufacturers use looser tolerances (±5 dB, ±10 dB) that make their specifications essentially meaningless.
On-axis versus off-axis response may be specified differently. On-axis response measures directly in front of the speaker, where response is typically smoothest and flattest. Off-axis response at angles reveals how sound spreads and whether high frequencies beam excessively. Complete specifications should indicate measurement angle.
Anechoic versus in-room measurements differ dramatically below approximately 300-400 Hz where room acoustics dominate. Anechoic measurements (in free space or using absorption to simulate free space) show the speaker's direct response without room influence. In-room measurements include room reflections and boundary reinforcement that dramatically affect low-frequency response.
Optimizing frequency response involves identifying deviations from target and applying appropriate corrections through EQ, positioning, or processing.
Room-related problems below approximately 300 Hz typically cannot be fixed with EQ alone because they result from standing waves and boundary interactions. Bass trapping, subwoofer positioning, and acoustic treatment address room-based response problems more effectively than equalization.
System EQ can address response irregularities above approximately 100-200 Hz where room effects are less dominant. Parametric or graphic EQ applied to the system output can smooth response peaks but cannot fill deep nulls caused by acoustic interference. Conservative EQ is preferable—aggressive boosts consume headroom and may drive amplifiers into clipping.
Speaker placement significantly affects frequency response, particularly in the low frequencies where room boundaries strongly influence output. Small adjustments in speaker positioning (inches can matter at low frequencies) can dramatically improve or degrade response at the listening position.
Explore these helpful resources for your audio system: