Perceptual grouping in the cocktail party: contributions of voice-feature continuity

Kreitewolf, Jens; Mathias, Samuel R.; Trapeau, Régis; Obleser, Jonas; Schönwiesner, Marc

doi:10.1101/379545

Cited by 7 publications

(10 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Similar to several of our previous studies [34][35][36], we followed an iterative model-fitting procedure: Starting with the minimal model that only included the subjectspecific random intercepts, we first added fixed-and then random-effects terms in a step-wise fashion. Fixedeffect terms were added in the order of their clinical importance (i.e., diagnosis, vestibular symptoms, surface, eyes, and interactions between these factors).…”

Section: Discussionmentioning

confidence: 99%

Subdiagnosis, but not presence of vestibular symptoms, predicts balance impairment in migraine patients – a cross sectional study

et al. 2020

Self Cite

View full text Add to dashboard Cite

Background: Vestibular symptoms and balance changes are common in patients with migraine, especially in the ones with aura and chronic migraine. However, it is not known if the balance changes are determined by the presence of vestibular symptoms or migraine subdiagnosis. Therefore, the aim of this study was to verify if the migraine subdiagnosis and/or the presence of vestibular symptoms can predict balance dysfunction in migraineurs. Methods: The study included 49 women diagnosed with migraine with aura, 53 without aura, 51 with chronic migraine, and 54 headache-free women. All participants answered a structured questionnaire regarding migraine features and presence of vestibular symptoms, such as dizziness/vertigo. The participants performed the Modified Sensory Organization Test on an AMTI© force plate. The data were analysed using a linear mixed-effect regression model. Results: The presence of vestibular symptoms did not predict postural sway, but the subdiagnosis was a significant predictor of postural sway. Migraine with aura patients exhibited more sway than migraine patients without aura when the surface was unstable. Additionally, we found high effect sizes (ES > 0.79) for postural sway differences between patients with chronic migraine or with aura compared to controls or migraine without aura, suggesting that these results are clinically relevant. Conclusions: The subdiagnosis of migraine, instead of the presence of vestibular symptoms, can predict postural control impairments observed in migraineurs. This lends support to the notion that balance instability is related to the presence of aura and migraine chronicity, and that it should be considered even in patients without vestibular symptoms.

show abstract

Section: Discussionmentioning

confidence: 99%

Subdiagnosis, but not presence of vestibular symptoms, predicts balance impairment in migraine patients – a cross sectional study

et al. 2020

Self Cite

View full text Add to dashboard Cite

show abstract

“…We propose that in the context of naturally-varying voices, using such spectral representations as acoustic measures is preferable to other measures that have frequently been used in voice research, such as the mean F0 and/or measures derived from formants, such as vocal tract length (Baumann & Belin, 2010;Gaudrain, Li, Ban & Patterson., 2009;Kreitewolf, Mathias, Trapeau, Obleser & Schönwiesner, 2018;Lavner et al, 2000; see Mathias & von Kriegstein, 2014, for a review). First, the spectral representations we describe are richer signals that include -but are not restricted to -F0 and formant-related acoustic information within the same measure.…”

Section: Experiments 1: Voice Identity Discriminationmentioning

confidence: 99%

Familiarity and task context shape the use of acoustic information in voice identity perception

Lavan¹,

Kreitewolf²,

Obleser³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Familiar and unfamiliar voice perception are often understood as being distinct from each other. For identity perception, theoretical work has proposed that listeners use acoustic information in different ways to perceive identity from familiar and unfamiliar voices: Unfamiliar voices are thought to be processed based on close comparisons of acoustic properties, while familiar voices are processed based on diagnostic acoustic features that activate a stored person-specific representation of that voice. To date no empirical study has directly examined whether and how familiar and unfamiliar listeners differ in their use of acoustic information for identity perception. Here, we tested this theoretical claim by linking listeners’ judgements in voice identity tasks to complex acoustic representations—spectral similarity of the heard voice recordings. Participants (N=150) who were either familiar or unfamiliar with a set of voices completed an identity discrimination task (Experiment 1) or an identity sorting task (Experiment 2). In both experiments, identity judgements for familiar and unfamiliar voices alike were guided by spectral similarity: Pairs of recordings with greater acoustic similarity were more likely to be perceived as belonging to the same voice identity. However, while there were no differences in how familiar and unfamiliar listeners used acoustic information for identity discrimination, differences were apparent for identity sorting. Our study therefore challenges proposals that view familiar and unfamiliar voice perception as being at all times distinct and suggests a critical role of the listening situation in which familiar and unfamiliar voices are evaluated.

show abstract

“…This phenomenon is known as the cocktail party effect (CPE) and can be difficult for machines to imitate [1], [2], [3], [4]. The CPE has been the subject of many studies [5], [6], [7], [8], [9], [10], [11]. Researchers in the fields of psychophysics and neuroscience have discovered that besides binaural spatial cues, monaural cues also play an important role in CPE tasks [7], [12].…”

Section: Introductionmentioning

confidence: 99%

“…In particular, evidence suggests that listeners can rely solely on monaural cues to recognize a target speech in the presence of a masking speech. Past experiments showed that differences in the spectral envelope and base frequency-also referred to as the fundamental frequency (F0)-between concurrent speech streams are the two most influential monaural cues [7], [8], [9], [13]. F0 is determined by vocal cord vibration while the configuration of the vocal tract determines the spectral envelope [7], [14].…”

Section: Introductionmentioning

confidence: 99%

Effects of Base-Frequency and Spectral Envelope on Deep-Learning Speech Separation and Recognition Models

Jian¹,

Wang²,

Chen³

et al. 2019

Interspeech 2019

View full text Add to dashboard Cite

Base-frequencies (F0) and spectral envelopes play an important role in speech separation and recognition by humans. Two experiments were conducted to study how trained networks for multi-speaker speech separation/recognition are affected by difference of F0 and spectral envelopes between source signals. The first experiment examined the effects of natural F0/envelope on the performance of speech separation. Results showed that when the two target signals differed in F0 by ±3 semitones or more or differed in the envelope by a scaling factor larger than 1.08 or less than 0.92, separation performance improved significantly. This is consistent with human listeners and is the first finding for deep learning-network (DNN) models. The second experiment tested the effect of F0/envelope difference on multi-speaker automatic speech recognition(ASR) system's performance. Results showed that multi-speaker recognition result also significantly rely on F0/envelope differences. The overall results indicated that the dependency of the existing automatic systems on monaural cues is similar to that of human, while automatic systems still perform inferior than human on same tasks.

show abstract

Perceptual grouping in the cocktail party: contributions of voice-feature continuity

Abstract: 23

Cited by 7 publications

References 37 publications

Subdiagnosis, but not presence of vestibular symptoms, predicts balance impairment in migraine patients – a cross sectional study

Subdiagnosis, but not presence of vestibular symptoms, predicts balance impairment in migraine patients – a cross sectional study

Familiarity and task context shape the use of acoustic information in voice identity perception

Effects of Base-Frequency and Spectral Envelope on Deep-Learning Speech Separation and Recognition Models

Contact Info

Product

Resources

About