3. Testing And Evaluation Methods

As with any audio equipment, the listening experience is what counts most of all, but there are pitfalls inherent in using that type of evaluation exclusively. Using electroacoustic measurements has its uses. Accordingly, we used our DAAS32 audio analysis system with a Neutrik artificial ear and microphone to record the response curves. This system is perfectly suited to, and calibrated for, testing traditional earphones.

For earbuds and intraaural earphones, we attempted to make an adapter for our artificial ear, but the results weren't very good because it was difficult to get proper acoustic coupling. So we constructed our own coupling system out of sound-deadening material - somewhat analogous to a plane-wave tube - and the results proved very satisfactory. In fact, we made several models to suit the different types of earphone. In all cases, the correlation between the response measurements and the actual listening experience was all but perfect.

Note that our measurement system is calibrated, and that the volume values in dB SPL are absolute levels (re 20 micropascals) for a voltage of 1 V at the earphone input. That means you can also compare the sensitivity of the headphones; keep in mind, however, that impedance is not taken into consideration. There are several ways of measuring the efficiency of a set of earphones. The simplest approach (and in our opinion also the best) consists of taking only the electrical level at the input into account. This is the application of basic audio amplification theory, where the source is considered a theoretically perfect voltage generator. Thus, we express sensitivity in dB SPL per volt. This is also the method used by some manufacturers, such as AKG and Sennheiser, and they're certainly no lightweights.

The other approach is to look at the transfer of power between the source and the receiver. This means that impedance is taken into consideration, and the result is a measurement of sensitivity in dB SPL per unit of power; that unit, in practice, is the milliwatt. Unfortunately, given the diversity of impedances found on earphones - typically from a few ohms to more than 1,000! - and the existence of "active" models containing electronic circuitry, this measurement has no real meaning with some products. It's meaningful only for passive, low-impedance models. Still, we show these values in the specifications when the manufacturer supplies them.

It's difficult to compare the different ways of expressing sensitivity simply. If you take the standard impedance value of 32 ohms, the sensitivity in mW is equivalent to a sensitivity in dB SPL for 0.18 V, which works out to approximately 15 dB less. Finally, certain manufacturers give a sensitivity specification, but don't say how they arrived at it...

We also recorded impedance curves to verify the values we actually obtained, since we were aware that they would be only be problematic if they were too low. Because of the test configuration we use, very high values aren't precise, but that's unimportant here. Low values are what we need to measure precisely!