Discrimination of aspiration noise in breathy vowels
Bibliographic record
Abstract
The amount of noise in a speech signal is typically associated with the perception of breathiness (Hillenbrand, 1998; Kreiman and Gerratt, 2005). Previous studies estimated difference limens (DLs) for discriminating a change in the amount of noise in a vowel using either adaptive procedures (Shrivastav and Sapienza, 2006) or a same-different task (Kreiman and Gerratt, 2005). In these studies, DLs for the amount of noise in a vowel were found to be about 11 to 21 dB, and depended on the noise level of the standard vowel (smaller DLs for larger amounts of noise in the standard). While previous studies co-varied the amount of noise with changes in other glottal parameters (e.g., open quotient, spectral tilt), in the present study we investigated DLs for aspiration noise by keeping other glottal parameters constant. Vowels were synthesized using the implementation of the Klatt synthesizer within the Praat software (Boersma and Weenink, 2015). Six-formant vowels /ae/ and /i/ were modeled after utterances of three male speakers selected from the Hillenbrand vowel database (Hillenbrand, Getty, Clark and Wheeler, 1995). There were six vowels in total (two vowels by three speakers). Formant and fundamental frequency estimates at eight time points within each vowel were used to synthesize the experimental stimuli using the parallel branch of the Klatt synthesizer. Glottal waveform settings for all stimuli included an open quotient of 0.4. Voicing amplitude was set to reproduce the intensity contour of the original vowels, and ranged from 67 to 73 dB (average = 70 dB). All other glottal parameters were set to default values, except for the level of the aspiration noise (AH), which was set to either 35 (standard stimuli), 37, or 39 dB. Vowel duration ranged from 276 to 302 ms (average = 290 ms). Stimuli were presented using a two-alternative forced choice (2AFC) procedure. Within each trial, listeners heard a sequence of two intervals, one containing the standard stimulus (35-dB AH level) and the other containing the comparison vowel (AH level of either 37 or 39 dB). Naive listeners participated in a short training session prior to taking part in two experimental sessions (one for /ae/, and one for /i/ stimuli). Listeners were asked to select the interval corresponding to the stimulus that sounded “breathier”. Feedback was provided after each response. Stimuli were presented binaurally through headphones at a level of 72 dBA. Responses were converted into d’ scores, a sensitivity measure based on signal detection theory (Macmillan and Creelman, 2005). As expected, sensitivity was consistently higher for the 39 dB than for the 37 dB stimuli. Differences in d’ scores were observed between the two vowels, and across the three speakers. By taking a d’ value of 1 (moderate sensitivity) as the empirical threshold estimate for a change in AH, DLs for aspiration noise were found to be 2-4 dB across the six vowels. These DLs are much smaller than those reported by previous studies. The possible reasons for the large differences in reported DLs among different studies, and the acoustic cues associated with breathiness discrimination with the present vowels, will be discussed. Boersma, P., and Weenink, D. (2015). Praat: doing phonetics by computer [Computer program]. Version 5.4.19, retrieved 22 September 2015 from http://www.praat.org/ . Hillenbrand, J., Getty, L. A., Clark, M. J., and Wheeler, K. (1995). Acoustic characteristics of American English vowels. The Journal of the Acoustical Society of America, 97 (5), 3099-3111. Kreiman, J. and Gerratt, B. R. (2005). Perception of aperiodicity in pathological voice. The Journal of the Acoustical Society of America, 117(4 Pt 1), 2201–2211. Macmillan, N. A. and Creelman, C. D. (2005). Detection Theory: A user’s giude (2 nd ed.) Mahwah, New Jersey: Lawrence Earlbaum Associates. Shrivastav, R., and Sapienza, C. M. (2006). Some difference limens for the perception of breathiness. The Journal of the Acoustical Society of America, 120 (1), 416-423.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.000 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".