To measure the performance of headphones accurately and relevantly requires the use of costly test equipment. But simple subjective assessments using special test signals can give some quick insight into a headphone's performance. Here are some you may find useful.
A slow exponential sine sweep (sometimes termed a logarithmic or log sine sweep) is a means of assessing a headphone's tonal balance. The MP4 video below includes an initial 1kHz tone to help you set an appropriate, comfortable replay level – begin with the volume set low – before sweeping from 10Hz to 10kHz over a period of 40 seconds, equivalent to a sweep rate of one octave every four seconds. In the course of the sweep, the cursor displays instantaneous frequency, so if you hear an anomaly you can pause the video and see at what frequency it occurs.
You should not expect to hear a constant output level versus frequency with this signal. If you use a typical replay volume then a headphone with a flat frequency response below 1kHz will deliver gradually increasing perceived output up to that frequency because of the loudness characteristics of the ear.
Beyond this – between 1kHz and 10kHz – the output level should appear to be about constant, and the swept tone should remain centrally positioned rather than move from side to side. With many modern headphones, which have a treble-light tonal balance, the output level will seem to reduce somewhat above 1kHz but should not fluctuate much. Peaks in treble output at around 8kHz will typically result in sibilant emphasis, while large shifts in image position indicate significantly different frequency responses at the two ears. If you hear similar shifts whatever headphone you use, it may be that you have different hearing acuity in your left and right ears and this is responsible rather than the headphone. Try wearing the headphones the other way round and see if the same effects persist.
If you hear output before the sweep reaches 20Hz you may be experiencing harmonic distortion products if the headphone has poor low-frequency linearity.
For private, non-commercial use you can download the MP4 video via the right mouse button menu.
As every headphone user knows, headphones 'image' quite differently to loudspeakers when replaying conventional stereo recordings. The loudspeaker sound stage, which covers an arc from 30 degrees left to 30 degrees right with a conventional equilateral triangle layout of speakers and listening position, with headphones becomes a line of image positions stretching from the left ear to the right ear. Ideally what are equally spaced positions in the loudspeaker soundstage should translate into equally spaced image positions in the head of the headphone listener. Or, if externalisation of the headphone image is achieved (usually by electronic processing), it should behave similarly to loudspeaker imaging.
This test track helps determine whether this is the case. It is constructed using a mono recording of violin made in an anechoic chamber. It was created for auralisation purposes (the simulation of acoustic spaces) whereas here it is used to step round the loudspeaker image from 30 degrees left to 30 degrees right in 5 degree increments, which is equivalent over headphones to full left to full right in 13 steps. It begins with the violin playing centrally, to allow level setting, then moves progressively from left to right, the green circle showing which image position is currently playing.
Ideally the 13 image positions should be regularly spaced, the image should always be narrow, with no sign of spreading, the central image should be exactly central, and the tonal balance should remain consistent. What you will sometimes hear, because the violin recording has wide bandwidth and strong partials (harmonics), is that the full left or full right signals excite diaphragm resonances that are less apparent elsewhere, and/or that the tonal balance differs at either end of the range because of differences in frequency response at the left and right ears. If the central image is offset, there may be a disparity in the sensitivity of the headphone's left and right capsules. As noted with the Exponential sine sweep test, if you hear this effect whatever headphone you use, it may be that you have different hearing acuity in your left and right ears and this is responsible rather than the headphone. Try wearing the headphones the other way round and see if the same effects persist.
For private, non-commercial use you can download the MP4 video file via the right mouse button menu.
Many headphones suffer from headband resonance but it is rarely commented on in reviews and even less frequently measured. But is is surprisingly easy to test for headband vibration and resonance subjectively, using the audio file below.
It comprises 26 successive one-second bursts of pink-spectrum periodic noise – which has a line spectrum from 20Hz to 20kHz at 1Hz intervals – in the left channel only. Because the signal is confined to the left channel, it should ideally be heard to come from the headphone's left capsule only. Whereas commonly what you hear instead is sound carried over towards the right side of the head, sometimes accompanied by obvious coloration resulting from headphone structural resonance. You may find that you can hear this most clearly if, while the signal is playing, you remove the headphone from your right ear and then replace it.
Hint: if you can contrive it, I've found that headband resonances are often more readily audible if the headphone is fed from a high-resistance source. Try 100 ohms or thereabouts.
For private, non-commercial use you can download the audio file via the right mouse button menu.
Headband resonance test