JASA express letters最新文献

Echolocating big brown bats (Eptesicus fuscus) perceive a target's reflecting parts, or glints, from spectral peaks and nulls in echoes. Glints serve as cues for object size and shape. In this exploratory study, we investigated whether the scalp-recorded auditory brainstem response (ABR) carries information about glints. We show that ABR amplitude, latency, and root mean square energy do not differentiate between glints in a manner paralleling the bat's perception. ABR recordings, even though they encode some perceptually relevant features of echoes, may not be sensitive enough to distinguish these more subtle cues.

The deconvolved beamforming (dCv) improves spatial resolution without expanding the array aperture but fails for the shift-variant beam pattern and the real targets, which are not located on the sampling grids. To solve them, this Letter extends the off-grid sparse Bayesian learning (OGSBL) to dCv because the generalized convolutional model considers the beam pattern at each angle in beam domain. OGSBL reduces modeling errors by parameterizing sampled locations in coarse grids. Controlling the number of output beams from conventional beamforming to cover the spatial area of interest could accelerate convergence without sacrificing accuracy. The simulation results confirm the good performance.

We applied passive acoustic monitoring and Hierarchical Density-Based Spatial Clustering of Applications with Noise clustering to define spatiotemporally cohesive groupings of franciscana dolphin (Pontoporia blainvillei) echolocation click trains. This unsupervised, objective method identified biologically relevant click train clusters, offering rare insights into the species' social organization. The observed structure revealed consistent intra-cluster cohesion and inter-cluster separation, supporting the effectiveness of the approach. Our findings demonstrate that clustering acoustic detections can serve as a robust framework for delineating social groups and can be integrated into future density estimation protocols, enhancing the ecological understanding and conservation potential for this cryptic and vulnerable species.

Speech synthesis has improved dramatically over recent years, enabled by large datasets and advances in neural network architectures. Little is known, however, about how synthesised speech patterns are realized from a phonetic perspective. By synthesising speech in two languages with differing implementations of stop voicing, we observe that synthesised speech broadly follows expected patterns for each language, though partially diverges for specific segments. Synthesising speakers into the opposing language also results in stops similar to target language distributions. These findings demonstrate the capability of speech synthesis models to encode phonetic information and further motivate questions regarding the phonetics of synthesised speech.

Reverse correlation (RC) is an established method for reconstructing auditory representations and has recently emerged as a tool for characterizing the sounds experienced by tinnitus patients. Tod further optimize RC for auditory research, the present work investigated the influence of subject instructions on characterization quality of tinnitus-like sounds. A validation study was conducted in which 18 normal-hearing subjects were randomly assigned 1 of 3 candidate instruction sets, each inspired by the RC literature. Results show a significant effect of instruction set on characterization quality and reveal that instructing subjects to detect a hidden signal in the RC stimuli resulted in the best reconstruction.

This study investigates the feasibility of using non-native listeners to assess non-native pronunciation of different target languages. Native English-speaking students studying university-level German and French reproduced German, French, and Indonesian sentences based on native-speaker models. Students not pursuing university-level language study did the same. Other native English-speaking students served as non-native listeners, rating the similarity of elicited sentences to native-speaker models. Similarity ratings were higher for language learners compared to non-learners across languages. Articulation rate and interval-based rhythm measures were strong predictors of similarity ratings. Results parallel findings from second-language acquisition studies, wherein native listeners typically evaluate L2 pronunciation.

When stretch and vibration are applied, human vocal folds develop from an isotropic gel-like structure to a layered structure with specific fiber direction. A human vocal fold fibroblast cell line was used to seed polyurethane foams that were cultured in media and subjected to a variety of mechanical conditioning regimens over a 3-week period. A bioreactor mimicked forces experienced by the human vocal folds to develop a vocal ligament. Fiber orientation was studied qualitatively for combinations of strain and vibration. While collagen type 1 fibers were randomly organized under static conditions, a combination of vibration and strain produced some fiber orientation.

This study investigated how variation in vocal timbre (e.g., neutral and twangy) influences intelligibility and listening effort in traffic-shaped noises. Word identification and subjective listening effort ratings were used to assess timbre effects. Intelligibility was significantly higher in twangy speech than in neutral speech. Twangy timbre had lower listening effort compared to neutral timbre for female speech only. The twang advantage might arise from higher perceived loudness, higher F1 and F2, and less breathy and hoarse voice quality. The result suggests that vocal timbre adjustments could be a strategy talkers employ to improve intelligibility in noise.

Acoustic differences between homophonous morphemes raise concerns for mainstream models of language production. But evidence for such differences so far comes almost exclusively from word final s in English. The current study analyzes homophonous va affixes and corresponding non-morphemic syllables in a corpus of Bora, an under-resourced and morphologically complex Amazonian language. Results from a statistical model that controls for final lengthening and other factors known to influence syllable duration indicate that non-morphemic vas are on average 8.2 ms longer than va affixes. This adds urgency to reconsidering the role of morphology in language production.

This paper describes Omura's whale vocalizations recorded off northern Australia. Most common was a two-unit "doublet" (17-50 Hz, peak 25.5 Hz, duration ∼15 s). This was rhythmically repeated for extended periods of time, in typical baleen-whale song structure, suggesting it is a male breeding display. A one-unit "singlet" occurred in the lower-latitude locations off the northwest (13-72 Hz, peak 26.6 Hz, 8 s). In the Great Barrier Reef, similar doublets were detected, identifying a possible population of the species in the Southwest Pacific. Such geographic variation of song may benefit passive acoustic monitoring of this species for conservation management.