Distortion Measurement
Canonical Definition
Distortion measurement quantifies output components that are not part of the intended linear response under a stated stimulus, level, bandwidth, load, and measurement method. THD, THD+N, individual harmonics, intermodulation products, and residual noise are related but distinct results.
Core Metrics
| Metric | Includes | Main caution |
|---|---|---|
| Individual harmonic | One integer multiple of the excitation frequency | Must be above the measurement noise floor and within bandwidth |
| THD | Selected harmonic components relative to the fundamental | Result depends on which harmonics are included |
| THD+N | Harmonic distortion plus noise and non-harmonic energy in the stated bandwidth | Strongly affected by measurement bandwidth and noise floor |
| IMD | Products created by interaction of two or more input frequencies | Depends on stimulus and calculation method |
| Absolute harmonic level | Acoustic or electrical level of the distortion component | Often more useful than percent when the fundamental response is irregular |
Measurement Methods
Logarithmic sweep
Fast and useful for magnitude, impulse, phase, harmonics, and broad acoustic testing. Longer sweeps improve harmonic separation and signal-to-noise ratio. Short sweeps or high-distortion systems can produce misleading overlap between components.
Stepped sine
Slow but often more accurate for very low electronic distortion, high-frequency harmonics, level sweeps, compression, and detailed output behaviour. Sustained tones require conservative level and thermal monitoring.
RTA spot check
Useful for a quick harmonic view at one frequency, but the source purity, FFT settings, window, input range, and noise floor must still be documented.
Electronic Versus Acoustic Distortion
Electronic measurements can use a controlled load and direct voltage reference. Acoustic measurements also include the loudspeaker's frequency response, microphone response, room or cabin, distance, reflections, background noise, and sensor distortion. Do not reuse amplifier thresholds as universal speaker targets.
Speaker distortion should be reported with fundamental SPL, distance, position, drive level, crossover, enclosure, frequency range, and thermal state. A percentage without those conditions is not a portable specification.
Normalised Percent Versus Absolute Level
When the fundamental becomes weak, the same absolute harmonic level can become a much larger percentage. That is why normalised speaker-distortion graphs often rise at frequency-response rolloffs. Inspect both absolute harmonic level and percent relative to the fundamental when possible.
Isolate The Measurement Chain
- Verify the generator or file is clean.
- Measure interface loopback and input headroom.
- Confirm the microphone and preamp do not distort at the test level.
- Test the device under test at multiple levels.
- Repeat suspicious features with a different method or lower level.
A microphone, interface, or source can create the same harmonic pattern that might otherwise be blamed on the amplifier or loudspeaker.
No Universal Pass Table
Acceptability depends on device type, frequency, harmonic order, output level, bandwidth, listening context, and the purpose of the test. Establish project limits from the relevant standard, product requirement, controlled comparison, or measured failure onset rather than one generic THD table.
Minimum Reporting Set
- stimulus, frequency range, level, sweep length or step settings, sample rate, and averaging
- load, distance, microphone position, environment, crossover, and operating temperature
- measurement bandwidth, harmonics included, calibration, and noise floor
- fundamental level, absolute harmonic levels, normalised values, and method
- repeat measurements at multiple levels to show onset and compression