Short Interval Recordings

Written by William F. Carver, Ph.D., FAAA, FASHA, CCC(A)ret.

A little more about live voice vs. recorded testing. One of the major reasons that I hear from those who defend the practice of using live voice is that recorded tests take too long.  Typically, we use four seconds as the interstimulus interval.  This period of silence between stimuli appears to be sufficient for most people to hear the word, decide what it is and repeat it.  I have and I’m sure that most audiologists have, experienced extremely slow responders where even four seconds is inadequate. On the other hand, we experience the quick responders for whom a couple of seconds is sufficient.  Auditec has, therefore, produced recordings with shortened interstimulus intervals of 2.5 seconds.  I know of many audiologists who have taken advantage of these recordings quite successfully.  (For the occasional slow responder, one can employ the pause button on their playback device.)

Short interval recordings are available from Auditec, Inc.  See a sample comparison in the video below. The most popular version is the NU-6 Ordered by Difficulty Version II.


Written by William F. Carver, Ph.D., FAAA, FASHA, CCC(A)ret.

You receive a referral from an audiologist.  An audiogram which includes a word recognition score is sent with the patient.  What can you tell from the word recognition score?  Nothing!   What list was used; W-22, NU-6, or PBK lists?  Did the audiologist use live voice, or was it from a recording?  Was it from a commercially available recording?  What is the articulation function (performance/intensity function) of the recording or of the audiologist’s voice? These variables can have a profound effect on  a word recognition score.

Presentations of cases (Grand Rounds) at conventions and meetings usually include an audiogram, SRT and word recognition scores…again what can you tell from the word recognition score?  If the presentation does not include information relative to how it was obtained with what materials, you have no real idea of the patient’s ability to discriminate speech.

To convey the crucial information about a person’s ability to communicate verbally, one must specify, not only the word recognition score, but must list: the list (i.e. W-22 or NU-6), the sensation level at which the test was administered, whether live voice or recorded.  If live voice, what is “normal?”  If from a recording, who’s recording (i.e. Auditec, dubbed from Technisonics, LAFO, QMass, etc.) ?

What variables control a word recognition score?  1.  The talker (the primary source of variation), 2. The presentation level, and 3. The list employed.

Word recognition (nee speech discrimination) is a slippery aspect of auditory tests.  Attempts have been made to standardize word recognition testing, but results have been disappointing.  Ideally, recorded tests that have been used throughout the area should be used.  And, ideally, one should not rely on a single score.  It has been shown that some sensory-neural patients will exhibit an articulation function that rises slowly and then curls over at higher levels.  Thus if one measures word recognition at a comparatively low level, it may be missed that a patient’s word recognition gets worse at higher levels.  A significant finding.

The point is, one should take the time to used recorded materials when testing for word recognition and ideally, especially in sensori-neural cases, obtain at least two measures at medium and high levels.

Adult word recognition lists like the NU-6 and W-22 and child word recognition lists like the PBK are available from Auditec, Inc.