Sound Express™: Assess Module

This module is one of the special module developed for Sound Express. This module is a speech test only module, which does not include training function in the program. Currently, there are two groups in this module. One group is the speech test in quiet. There are four speech tests in this group, including vowel recognition in quiet, consonant recognition in quiet, melodic contour identification, and concatenated sentence recognition in quiet. The other group is the speech test in noise,. There are two speech tests in this group, including digits recognition threshold and concatenated sentence recogntiion threshold.

Group 1: Speech Test in Quiet

This group is to evaluate speech recognition performance in quiet. The tests include vowel recognition, consonant recognition, melodic contour identification, and concatenated sentence recognition. Here is the more detailed description for each speech test:

1. Vowel recognition test

Vowel recognition test will be assessed in a 12-alternative identification paradigm. The vowel set included 10 monophthongs and 2 diphthongs, presented in a /h/-vowel-/d/ context. The tokens for vowel recognition test were digitized natural productions from 2 men and 2 women, resulting in a total of 48 tokens. Here shows the main interface for this test.

2. Consonant recognition test

Consonant recognition was measured in a 20-alternative identification paradigm. The consonant set included /b d g p t k m n l r y w f s S v z D tS dZ/, presented in an /a/-consonant-/a/ context. Consonant tokens consisted of digitized natural productions from 2 men and 2 women, for a total of 80 tokens. All consonant tokens were recorded at the House Ear Institute. Here shows the main interface for this test.

3. Melodic Contour Identification

For the MCI task, test stimuli were melodic contours composed of 5 notes of equal duration whose frequencies corresponded to musical intervals. The interval between successive notes in each contour was varied between 1 and 5 semitones; the “root note” of the contours was also varied. Nine distinct musical patterns were generated for each interval and root note conditionModule. Here shows the screen shot of the displaying window.

4. Concatenated Sentence Recognition

This speech test uses Concatenated speech training protocols. The sentence includes five words as the following order: name, verb, number, adj., and objects. Each cateagory includes 10 words, resulting in 100,000 random sentences. The user has to select the words from the screen (10x5) to have the whole sentence. The program will compare the whole sentnece and determine the percentage points. Here shows the interface for this training group.

Group 2: Speech Tests in Noise

This group uses several Advanced Training Protocols, including Open-set Recognition Training Protocols, Speech Synthesis based on Concatenation, and Trial-based Adaptive Speech in Noise Training. The task in this group will assess/train the listener's ability to recognize words, sentences, or digits in the presence of background noise. The background will use six-talker speech babble. Speech recogition threshold, the signal-to-noise ratio that produces 50% of words, sentences or digits recognition, will be measured. The digit string has 3 digits. The main interface will be same as the above groups.