Before we jump into the specifics of each graph, please note that electronic audio equipments – such as sound cards, amplifiers, or audio players – tend to be indistinguishable in many cases on these tests. Thus, other measurements, such as THD, IMD, and Stereo Crosstalk are used whenever this happens (the frequency response is still relevant).

The original signal out of the Lynx L22 is provided for comparison in each measurement – the results out of the earphones or headphones should show how the signal has been modified through the device.

1. Frequency Response

The Frequency Response curve is arguably the most important graph among the many graphs pertaining to the character of the sound. Normally, any talks of the character of the sound could be considered an elaboration on the Frequency Response curve.


The horizontal axis of the graph represents the pitch of the signal, measured in Hertz (Hz), and the vertical axis represents the intensity of the sound at a given pitch, measured in decibels (dB). Thus, the right part of the graph represents the treble, while the left part represents the bass; a higher line represents a larger sound and vice versa.

For example, if the < 100Hz section of the frequency response graph were to be modified, this would change the sound of bass guitars or drums, whose sounds lie almost entirely within the deep bass range. If the graph is shifted upwards in this range, this results in a powerful, deep sound, but if this part of the graph is emphasized too much the sound tends to become muddy as the upper frequencies are masked by the bass. On the other hand, if the graph is lowered significantly here then this has the effect of having a sound that sounds ‘weak’ and empty. To put it positively, the bass becomes ‘tame and controlled’ – or negatively, ‘tinny and brassy’.

In the case of the midrange, the piano and the vocal sounds tend to be affected by the change. If the upper part of the midrange is relatively emphasized (higher up on the graph), the piano and vocals tend to sound more delicate and detailed, while if the lower part is emphasized the piano and the vocals become dull and dark.

Finally, the upper range of the frequency response changes the sounds of instruments such as the hi-hat – if the upper range starting at 5kHz has a significant presence, instruments like the hi-hat sound very bright and clear, but it might border on aggressive if the bass presence is lacking. On the other hand, the hi-hat sounds muffled and dull if the upper frequencies are lacking compared to their lower counterparts.

As you might expect, Hi-Fi in the sense of ‘faithfully presenting the recorded audio’ requires a flat response across the entire audible spectrum, which ranges from 20-20000Hz. Sometimes there are products whose limited response is advocated as having been by design; however, unless the product has been made for a very specific application (such as in-ear-monitors with a boosted midrange for the vocalist on stage), it is generally desirable to have a wider frequency response.

2. Step Response

The step response is the graph obtained by delivering a step signal (as shown in the Lynx graph) to earphones or headphones. The horizontal axis is the time; the vertical axis is the voltage – the graph shows the change of voltage over time.


Ideally, the signal should decay exponentially as it is in the original output, but most products tend to show some ripple within < 20mS and decay rapidly. As you will notice in the reviews, headphones such as Sennheiser HD600 are notable for rippling only a mildly.

The Step Response is related to how reactive the product is as well as its frequency response, and a response like that of the Lynx L22 that shows an exponential decay without rippling is desirable.

3. Impulse Response

The Impulse Response is the response shown by the earphones or headphones when an impulse signal (a very short signal with high intensity like a gunshot) is input. It is analogous to firing a gun and seeing how the intensity of the sound changes over time.


A desirable Impulse Response graph should have no ripples as in the case of Lynx L22, and if artificial sound effect (like DNSE or SRS or DSEE) to DAPs (Digital Audio Players) for ambiance sometimes reflected signals may be registered.

4. Cumulative Spectral Decay

The Cumulative Spectral Decay, abbreviated CSD, is different from other graphs in that it is presented in 3 dimensions.
The horizontal axis of the graph represents the frequency, and the vertical axis represents the intensity, much as in the frequency response graph. The difference is that there is an extra time axis out of the page. Thus, the CSD graph is used to understand how the sound decreases in intensity (or ‘decays’) over time.


The graph represents the change in intensity across the frequency over time, so a mark of a good product is a balanced decaying across all frequencies. If one part of the spectrum decays markedly slower than the rest, this implies that there is a resonance at that frequency – this makes CSD an indispensable tool in testing driver enclosures.

5. Impedance & Phase

The impedance is a measure of how much earphones or headphones resist an electric current (more specifically, a change in a current).

The impedance graph is related to the output impedance of DAPs and headphone amplifiers – for more information on the subject, consult ‘How the output impedance is related to the impedance graph’.