Optimization of the Dutch Matrix Test by Random Selection of Sentences From a Preselected Subset

Abstract

Matrix tests are available for speech recognition testing in many languages. For an accurate measurement, a steep psychometric function of the speech materials is required. For existing tests, it would be beneficial if it were possible to further optimize the available materials by increasing the function’s steepness. The objective is to show if the steepness of the psychometric function of an existing matrix test can be increased by selecting a homogeneous subset of recordings with the steepest sentence-based psychometric functions. We took data from a previous multicenter evaluation of the Dutch matrix test (45 normal-hearing listeners). Based on half of the data set, first the sentences (140 out of 311) with a similar speech reception threshold and with the steepest psychometric function (≥9.7%/dB) were selected. Subsequently, the steepness of the psychometric function for this selection was calculated from the remaining (unused) second half of the data set. The calculation showed that the slope increased from 10.2%/dB to 13.7%/dB. The resulting subset did not allow the construction of enough balanced test lists. Therefore, the measurement procedure was changed to randomly select the sentences during testing. Random selection may interfere with a representative occurrence of phonemes. However, in our material, the median phonemic occurrence remained close to that of the original test. This finding indicates that phonemic occurrence is not a critical factor. The work highlights the possibility that existing speech tests might be improved by selecting sentences with a steep psychometric function.

The Benefits of Bimodal Aiding on Extended Dimensions of Speech Perception: Intelligibility, Listening Effort, and Sound Quality

“…Sentence intelligibility in noise was tested with the optimized version of the Dutch Matrix test ( Houben & Dreschler, 2015 ) within a set-up of spatially separated speakers. The Dutch Matrix material is based on a closed speech corpus of sentences with an identical syntactical structure: “name, verb, numeral, adjective, object,” for example, “Mark gives five large flowers.” The accompanying Dutch Matrix noise is a stationary noise with an average power spectrum equivalent to that of the sentences.…”

“…More information on the development of the Dutch Matrix material is available in Houben et al. (2014) and Houben & Dreschler (2015) . This corpus of sentences has low linguistic complexity without redundancy, making it well suited for repetitive testing ( Houben et al., 2014 ).…”

Evaluating hearing performance with cochlear implants within the same patient using daily randomization and imaging based fitting - The ELEPHANT study

“…Outcomes Extensive outcome testing will be performed over the course of the study period ( [49] . In addition, the Dutch Matrix Sentence test, which has been validated for repeated measures in CI-recipients [50,51], will be used through three listening conditions: in quiet, with background noise, and with spatially separated background noise. The tests will be administered using the Oldenburg measurement applications software package developed by Hörtech gGmbH Germany.…”

The principle of inverse effectiveness in audiovisual speech perception

“…The audiovisual speech material is based on an existing auditory-only matrix sentence test for Dutch native speakers [30,37]. It is not immediately clear whether the observed results hold specifically for the Dutch language, or whether it is immaterial for which language this test has been developed.…”