What a spectrum shows

Any repeating signal can be built from sine waves of different frequencies added together. A spectrum shows how strong each of those sine waves is.

• A pure sine has one frequency, so one peak.
• A square wave is a sine at its own frequency plus weaker ones at 3, 5, 7 and more times it, its odd harmonics. So it shows a row of peaks getting smaller.
• Distortion adds harmonics to a signal that should be a pure sine.
• Noise spreads across all frequencies as a rough floor.

Decibels

Levels are shown in decibels relative to 1 volt RMS (dBV). Each 20 dB is a factor of ten in voltage: 0 dBV is 1 V RMS, -20 dBV is 0.1 V, -40 dBV is 0.01 V. This lets a small harmonic and a large fundamental share one screen.

The level of each frequency is its RMS value, the steady voltage that would deliver the same power. For a sine, that is its peak divided by the square root of 2, so a 1 V peak sine reads 0.707 V RMS, or -3 dBV. Other units, dBu, dBm and volts, are in Spectrum.

Harmonics and distortion

A circuit that should pass a sine unchanged, such as an amplifier, adds harmonics when it distorts: a little at 2, 3, 4 times the frequency. Their total against the fundamental is the total harmonic distortion (THD). Its significance depends on the signal and the circuit being tested. Noise also limits how small a signal can be distinguished. The spectrum’s Distortion figures measure both; see Distortion.

Resolution

The duration of the samples used determines how closely spaced tones can be separated, and the window widens their peaks. The frequency step is the sample rate divided by the transform length. A gated capture may be padded with zeros to give finer steps, but that adds no resolution: the shorter gate still gives wider peaks. See Window. A lower sample rate gives a smaller frequency range; with the same sample count it also gives a longer capture.