ERIC Number: ED023209
Record Type: RIE
Publication Date: N/A
Reference Count: 0
Experiment in Learning to Discriminate Frequency Transposed Speech.
Ahlstrom, K.G.; And Others
In order to improve speech perception by transposing the speech signals to lower frequencies, to determine which aspects of the information in the acoustic speech signals were influenced by transposition, and to compare two different methods of training speech perception, 44 subjects were trained to discriminate between transposed words or syllables. Since the subjects had normal hearing, a hearing less was simulated by including a low-pass filter and white noise in the circuit for transposing speech signals. The consonant sounds most easily identified at the first test session were /g/, /k/, /sp/ and /sk/, with /d/ and /c/ the most difficult. The subjects learned to rely in their discriminations on the code dimensions voicing, duration, and cluster consonant but not on place of articulation. Neither of the two training methods compared, anticipation and recall for paired associate learning, was found more effective. It did not seem to matter which of the two training materials, words or syllables, was used. Six figures, 18 tables, and six references are included. (GD)
Publication Type: N/A
Education Level: N/A
Sponsor: Swedish Council for Social Science Research, Stockholm.; Tri-Centennial Fund, Stockholm (Sweden).
Authoring Institution: Uppsala Univ. (Sweden). Inst. of Education.