Many passive headphones (those without internal amplification) have an electrical impedance that varies widely across the audible frequency range. This is particularly true of dynamic (moving coil) headphones, much less so of planar magnetic types whose impedance is usually almost constant. Unusually low impedance can tax the current capability of some headphone outlets (particularly on mobile devices); impedance that varies significantly with frequency will cause the headphone's frequency response to be modified if it is fed from a source with finite output impedance. Prominent ripples in the impedance versus frequency trace may be associated with resonances within the headphone structure or drive unit diaphragm (in which case ridges should be apparent in the CSD waterfall at the same frequencies).

Impedance modulus is measured using a take-off box connected in the headphone feed, in a which there is a precision series resistance. By measuring the signal voltage at either end of this resistance, the headphone impedance can easily be calculated. In HTL's take-off box the resistor is 102 ohms ±0.1%, 1W. The headphone is placed on the artificial ear for this test and each capsule measured separately, with the driving signal provided by a headphone amplifier rather than directly by the measurement computer's sound card. Different test signals can provide differing impedance estimates; our test uses a stepped-sine signal and coherent FFT analysis to ensure consistent results, without the low frequency ripples that can result from using a periodic noise test signal. (Note: older measurements on this site, including all the legacy measurements, were made using a noise signal and may evince this LF ripple effect. If a noise signal was used it will say so in the graph's information box.)

Test signal:                                                      stepped sine (1000 steps)

Sampling frequency:                                     96kHz

FFT length:                                                      variable, 64pt to 1,048,576pt

Frequency resolution (measurement):       variable, 0.092Hz to 1.5kHz

Frequency resolution (graph):                      1/100th octave

Measurement length per step:                    100,000 samples minimum