s-sh Confusion Test Audio Files

Listed in Datasets

By Joshua M Alexander

Purdue University

Audio files for words in isolation and in two carrier phrases ("Pick the word" and "Say the word"), including word lists and calibration tone. Child version is a subset of word pairs from the adult version within a 5-year old...

Version 1.0 - published on 14 Jul 2021 doi:10.4231/AG8V-DM52 - cite this Archived on 15 Aug 2021

Licensed under CC0 1.0 Universal

Description

Sixty-six minimal pair /s/ and /ʃ/ words were selected, including proper nouns (e.g. Sam to contrast with sham). Forty-five of these pairs had /s/-/ʃ/ in the initial position, eighteen in the final position, and three in the medial position. Six female talkers were recorded speaking these words in with the carrier phrase “pick the word ____ now,” in order to encourage a natural speaking rate and to discourage emphasizing the /s/-/ʃ/ portion of the word. Native geographical regions of the talkers included the Midwest (Illinois, Indiana, and Ohio), the South (Tennessee), and the Northeast (New Jersey). Recordings were obtained in a sound-treated booth using a Shure model SM48 microphone positioned at a 45-degree angle and 4-5 inches from the talker’s mouth. Using a computer outside the booth, Audacity open source software version 2.1.2 was used to record the speech at a rate of 44,100 Hz with 32-bit depth. During recording, slips of paper were printed with each word, and the order was randomized by shuffling the slips. This was done to help avoid the tendency to over-enunciate the difference between the paired set, as might occur if the minimal pair was spoken sequentially. To extract words from the longer phrases, lab assistants used Audacity to examine the time waveform to mark the keyword, and then to listen to it in isolation to further refine their selection. The waveform was then marked at the zero amplitude crossings. These were extracted to an Audacity sound file and saved. Then using similar methods, the start and end time for the contrastive /s/ and /ʃ/ sounds were marked in the Audacity file for later acoustic analysis. For quality control purposes, three lab assistants listened to the recordings individually in order to make judgments about sound quality and naturalness when played in isolation. For a further check of quality control, the assistants listened to word pairs sequentially to ensure that the pronunciation was similar in most respects, except for the /s/ and /ʃ/. Words and word pairs that did not meet this requirement were re-recorded and evaluated again, and if necessary recorded yet again. Keywords and the calibration tone are scaled to equal RMS amplitude (-20 dB full scale). Carrier phrases ("pick the work . . .," "say the word . . .," ". . . now" are scaled to the same amplitude as the keywords. Keywords were inserted into the carrier phrases with 50 ms of silence before and after the keyword.

Alexander, J.M. (2019).  “The s-sh confusion test and the effects of frequency lowering,” J. Speech Lang. Hear. Res., 62, 1486-1505.

Cite this work

Researchers should cite this work as follows:

Tags

The Purdue University Research Repository (PURR) is a university core research facility provided by the Purdue University Libraries and the Office of the Executive Vice President for Research and Partnerships, with support from additional campus partners.