How to Read a Spectrum Analyzer: A Practical Guide

Start with the two axes

A spectrum analyzer turns a moment of audio into a frequency map. The horizontal axis moves from low frequencies on the left to high frequencies on the right. The vertical axis shows signal level, usually in dBFS for digital audio. A point near 0 dBFS is strong; a point near −90 dBFS is very quiet.

The graph is a snapshot that updates during playback. It answers a precise question: which frequencies are present now, and how strong is each one? It does not show when a sound happened over the full recording—that is the job of a spectrogram.

• 20–250 Hz: sub-bass, bass fundamentals, hum, and low mechanical energy.
• 250 Hz–4 kHz: most vocal information, instrument fundamentals, and intelligibility.
• 4–20 kHz: brightness, sibilance, cymbals, hiss, and fine detail.

Read peaks as relationships

A narrow peak often represents a stable tone. If the strongest peak is at 220 Hz and smaller peaks appear near 440, 660, and 880 Hz, you are probably seeing a fundamental plus harmonics. The fundamental often determines perceived pitch, while the relative strength of the harmonics shapes timbre.

The tallest peak is not always the fundamental. A small speaker, room resonance, filter, or instrument construction can make a harmonic stronger than the note’s base frequency. Look for evenly spaced families of peaks before deciding what the source is.

Understand the floor between peaks

The broad layer beneath distinct peaks is the noise floor. A clean electronic tone may show thin lines over a very low floor. Speech, percussion, wind, and distorted sounds naturally spread energy across wider bands. A wide shape is not automatically a defect.

Use a reference trace when making changes. Save the first spectrum, adjust the recording or processing, and compare the new curve against the reference at the same playback position and level. This is more reliable than comparing from memory.

Choose linear or logarithmic frequency

A logarithmic scale gives more horizontal room to bass and midrange frequencies and resembles how musical pitch is organized. It is the best default for music, speech, and general inspection. A linear scale gives equal width to equal frequency intervals, which makes harmonic spacing, modulation products, and encoded spectrogram art easier to measure.

Finally, check the FFT size. Larger FFT sizes separate nearby steady tones more clearly but blur fast timing changes. Smaller sizes respond faster to transients. Use 4096 as a balanced starting point, then change it only when the task calls for a different tradeoff.