Calibration of the TDH-39 Earphone

Correction factors and thus accurate calibration of TDH-39 earphone mounted on MX-41/AR and Auraldome cushions are difficult to achieve.  The reasoning behind this is because different methods of measuring the performance of each device produce different results for each piece of hardware.  When psychoacoustic studies are conducted comparing the two earphone cushion combinations, both seem to suggest equivalent performance.  That is, threshold data obtained using both the supra-aural and circumaural cushions implies that equivalent SPL’s are produced at the tympanic membrane.  Consider the following data obtained from Ciechanowski et al. (1976). 

Ciechanowski et al. (1976) conducted similar measurements using an electroacoustic technique, and found a very different result.  What they found was that less SPL is generated by the TDH-39 using a circumaural cushion on a flat plate coupler compared to  the same earphone mounted in a MX 41/AR in a NBS-9A coupler.

In conducting an electroacoustic comparison, the TDH-39 earphone was mounted in the MX-41/AR cushion and the performance was measured in a NBS-9A coupler illustrated above.  The Auraldome coupled to the TDH-39 earphone was measured using a flat plate coupler with a 1-inch microphone.  

Consider the data and graph below.  According to this graph, the two earphone cushion combinations differ significantly.  The largest discrepancy between the two earphones occurs at 125 Hz.  The puzzling question now is why do the two earphone cushion combinations perform similarly under psychoacoustic testing but not under electroacoustic testing?  

Audiologists may find it interesting to know that the physical construction of the circumaural earphone has a specific impact on the behaviour of sound transmission.  Generally, the greater volume of air enclosed, the greater the reduction in SPL in the low frequencies.  This may account for the large discrepancy at 125 Hz in the above graph.   

When the TDH-39 MX 41/AR and Auraldome combinations are fixed onto a KEMAR manikin, different results occur again.  Consider the following graph, which presents a mean difference between the SPL’s developed in the Zwislocki coupler mounted in the manikin.  In this set up, the TDH-39 earphone is set in the MX 41/AR and Auraldome cushions, and in the NBS 9/A coupler by the same earphone in the MX/41 AR cushion.

What is clear in the manikin experiment is that the earphone performance with the MX-41/AR cushion is different from that measured on the NBS/9A coupler, especially in the lower frequencies.  Some of these differences can be easily accounted for.  The large reduction in SPL at the low frequencies on the KEMAR manikin is likely due to an acoustic leak around the earphone.  The greater response at 8000 Hz on the KEMAR manikin is most likely due to ear canal resonance effects.  

At this point, you may be confused as to which set of results are most relevant to your practice.  Take the following example into consideration: 

The KEMAR manikin model (containing a Zwislocki coupler) more closely resembles the human ear than the NBS-9A coupler which is basically a hard walled cylinder.  Therefore the increased SPL measured at 8000 Hz in the manikin experiment may very well be more representative a clinical situation.     

According to Zwislocki 1971, data on the performance of circumaural earphones on the Zwislocki coupler have shown that the coupler more accurately represents real ear data.  The drawback however is that no psychoacoustic data has been reported with circumaural earphones (as of 1976) and the ear-like coupler in a manikin. 

Zwislocki coupler

The problem with earphone calibration using a manikin is that psychoacoustic data obtained on a TDH-39 earphone with MX-41/AR and Auraldome cushions are very similar, however, electroacoustic responses of the two seem to differ consistently across cushions and across couplers.  As a result, Audiologists need to be aware that an accurate method of calibrating supra-aural or circumaural earphones on a manikin is not yet available.  

What was demonstrated in the Ciechanowski et al 1976 article is that data on the MX-41/AR and the Auraldome differ across many different testing conditions, be it on a manikin, on the flat plate/NBS-9A coupler or on actual people.  Ciechanowski et al 1976 also state that  until an accurate electroacoustic calibration method is available, the best means of assessment should be a biological calibration.  

Despite the recommendation by Ciechanowski et al. (1976), the current ANSI calibration method and standard should suffice and are the current accepted methods of calibration.   

According to ANSI S3.6 1996, calibration of the TDH39 supra-aural earphones needs to be applied to the coupler with a static force of 4.5 N.  The RETSPL values for the TDH39 are as follows:

data from ANSI S3.6 1996.

To calibrate the headphones, vary the audiometer output until it matches the NBS 9A coupler values at the specified frequency.   

Note:  This site is best viewed using IE 4.X or higher under 800x600 resolution