Question 1

How do I read a frequency spectrum display?

Accepted Answer

The spectrum display plots frequency on the horizontal axis (left = low, right = high) and level on the vertical axis (bottom = quiet, top = loud). On a logarithmic frequency scale, each doubling of frequency (octave) takes equal horizontal space. The vertical axis is in dB (decibels), typically spanning 60-80 dB of range. Peaks in the spectrum indicate strong frequency content. A flat spectrum means equal energy at all frequencies. A downward slope from left to right indicates more low-frequency energy (common in music and environmental noise). Pink noise has a flat spectrum on a 1/3 octave display. SonaVyx uses standard audio measurement conventions: logarithmic frequency, dB magnitude, with grid lines at standard frequencies (125, 250, 500, 1k, 2k, 4k, 8k Hz) and 5 dB vertical increments.

Question 2

What does a flat frequency response look like on a spectrum analyzer?

Accepted Answer

A flat frequency response means equal output at all frequencies. On a spectrum analyzer with a logarithmic frequency axis and showing the FFT (per-bin) magnitude, a flat signal appears as a horizontal line. However, pink noise (the standard test signal for audio) has 3 dB less energy per octave than white noise, so pink noise played through a flat system shows as a line sloping downward at 3 dB per octave on an FFT display. On an octave-band (RTA) display, pink noise through a flat system appears flat because the octave bands compensate for the increasing number of FFT bins per octave at higher frequencies. SonaVyx RTA mode shows octave-band levels, which is the standard for evaluating tonal balance. The FFT mode shows the raw per-bin spectrum for detailed analysis.

Question 3

What do narrow peaks in the spectrum mean?

Accepted Answer

Narrow peaks (also called tonal components or discrete frequencies) indicate a single-frequency or near-single-frequency sound source. Common causes: electrical hum (50/60 Hz and harmonics), acoustic feedback (a single frequency amplified by the loop), resonance of a structural element (duct, panel, pipe), equipment tones (motors, compressors, cooling fans), and electronic interference (switching power supplies, LED dimmers). The narrower the peak relative to the surrounding noise, the more tonal it is. Tonal noise is more annoying than broadband noise at the same level, which is why many noise standards apply a 5 dB penalty for tonal content. SonaVyx problem detector identifies tonal components and the hum detector specifically flags 50/60 Hz and harmonics.

Question 4

Why does my spectrum show more energy at low frequencies?

Accepted Answer

Most real-world audio signals have more energy at low frequencies. Music has strong bass content. Environmental noise (traffic, HVAC, wind) is dominated by low-frequency energy. Room modes amplify low frequencies where the room dimensions are comparable to the wavelength. The proximity effect boosts bass when a directional microphone is close to a source. Additionally, on an FFT display with linear frequency bins, low-frequency content appears concentrated because each octave at low frequencies contains fewer bins than at high frequencies. This is why octave-band (RTA) analysis is more intuitive: it groups bins into perceptually meaningful octave bands. SonaVyx displays both FFT and RTA views so you can see the raw spectrum and the perceptually weighted version.

Question 5

What is a spectrogram (or spectrograph) and how is it different from a spectrum?

Accepted Answer

A spectrogram displays three dimensions: frequency (horizontal axis), time (vertical or horizontal scrolling axis), and amplitude (color). It shows how the spectrum changes over time, like stacking successive spectrum snapshots. Hot colors (red, yellow) represent high levels, cool colors (blue, green) represent low levels. This is invaluable for identifying: intermittent noise events (they appear as horizontal streaks at specific times), frequency sweeps (diagonal lines), feedback building up (a bright line growing in intensity), and the temporal pattern of music or speech. SonaVyx spectrograph view displays scrolling time-frequency data with configurable history depth (5 to 60 seconds) and a color palette from dark blue (quiet) through green and yellow to red (loud).

Question 6

How do I identify feedback from a spectrum display?

Accepted Answer

Acoustic feedback appears as a single narrow peak that grows in level over time. It starts as a small elevation at the most sensitive frequency (where the loop gain is highest) and rapidly becomes dominant, eventually overwhelming the program material. On a spectrum display, feedback shows as a sharp spike at one frequency, often 20-30 dB above the surrounding levels. On a spectrogram, it appears as a bright vertical line that grows wider (hotter colors) over time. The frequency depends on the speaker/microphone positions and room acoustics. Common feedback frequencies are 2-4 kHz (where many microphones are most sensitive) and room resonance frequencies. SonaVyx feedback detector identifies potential feedback frequencies before they become audible by tracking spectral peak growth rates.

Question 7

What does comb filtering look like on a spectrum display?

Accepted Answer

Comb filtering creates a pattern of evenly-spaced peaks and dips across the frequency spectrum, resembling the teeth of a comb. The spacing between dips equals 1/delay_time. For a 1 ms delay (34 cm path difference), dips appear at 500, 1500, 2500, 3500 Hz (every 1000 Hz). For a 5 ms delay (1.7 m path difference), dips appear at 100, 300, 500, 700 Hz (every 200 Hz). The dips can be 6 dB deep (for equal-level direct and reflected signals) up to theoretically infinite depth (for perfectly matched levels). On a smoothed display, comb filtering appears as a wavy response with regular undulations. SonaVyx comb filter detector automatically identifies periodic null patterns, reports the implied delay, and suggests the reflection source distance.

Question 8

How do I tell the difference between a room mode and a speaker resonance?

Accepted Answer

Room modes appear at frequencies determined by room dimensions: f = c/(2L) and harmonics, where c is the speed of sound and L is a room dimension. They are position-dependent: measuring at different locations shows different mode amplitudes, and nodes (where a mode is weakest) alternate with antinodes (strongest). Speaker resonances appear at the same level regardless of microphone position (as long as you maintain the same distance and angle to the speaker). A quick test: measure at 3-4 different positions. Peaks that remain constant are likely speaker characteristics. Peaks that shift in amplitude or disappear are room modes. Additionally, room modes are typically below 300 Hz (in rooms smaller than 100 m³), while speaker resonances can occur at any frequency. SonaVyx trace overlay makes this comparison straightforward.

Reading a Spectrum Display: Peaks, Dips, Feedback, and Comb Filtering Identification

Frequently Asked Questions