Open-back headphones are, by their nature, subject to acoustical crosstalk: some of the sound emitted by one capsule will find its way across to the opposite ear at low level. But that is not what this measurement is primarily intended to capture. Of principal interest here is not air-borne but structure-borne crosstalk, via the headphone’s headband, which can introduce coloration because of resonances within the headphone structure. If such resonances are present they reveal themselves as peaks in the acoustic crosstalk trace, usually – but not always – at low frequencies.
The headphone’s right capsule is placed on the artificial ear for this test, and its one-shot frequency response measured in the normal way. Then its response is measured a second time but with the test signal played over the left capsule instead, with the right capsule unmoved. Because the acoustical crosstalk signal can be many tens of dB lower in level than the acoustical output of the active capsule, this measurement requires a high degree of averaging to obtain consistent results, particularly at low frequencies. It is also essential that the artificial head be compliantly isolated from external vibrations, which is achieved by mounting it on four coil springs that provide a suspension natural frequency of ~6Hz. To prevent electrical crosstalk affecting the measurement, the acoustical crosstalk signal is measured with the right capsule open-circuited by means of an adaptor box connected in series with the headphone.
Test signal: pink-spectrum periodic noise
Sampling frequency: 96kHz
FFT length: 32,768pt
Frequency resolution (measurement): 2.93Hz
Frequency resolution (graph): 1/100th octave
Measurement averages: 100
Trace smoothing: none