Comparing models of the combined-stimulation advantage for speech recognition.
Ontology highlight
ABSTRACT:
SUBMITTER: Micheyl C
PROVIDER: S-EPMC3356316 | biostudies-other | 2012 May
REPOSITORIES: biostudies-other
Similar Datasets
Project description:Infants appear to learn abstract rule-like regularities (e.g., la la da follows an AAB pattern) more easily from speech than from a variety of other auditory and visual stimuli (Marcus et al., 2007). We test if that facilitation reflects a specialization to learn from speech alone, or from modality-independent communicative stimuli more generally, by measuring 7.5-month-old infants' ability to learn abstract rules from sign language-like gestures. Whereas infants appear to easily learn many different rules from speech, we found that with sign-like stimuli, and under circumstances comparable to those of Marcus et al. (1999), hearing infants were able to learn an ABB rule, but not an AAB rule. This is consistent with results of studies that demonstrate lower levels of infant rule learning from a variety of other non-speech stimuli, and we discuss implications for accounts of speech-facilitation.
| S-EPMC3399874 | biostudies-literature
Project description:Purpose This study investigated methods used to simulate factors associated with reduced audibility, increased speech levels, and spectral shaping for aided older adults with hearing loss. Simulations provided to younger normal-hearing adults were used to investigate the effect of sensation level, speech presentation level, and spectral shape in comparison to older adults with hearing loss. Method Measures were assessed in quiet, steady-state noise, and speech-modulated noise. Older adults with hearing loss listened to speech that was spectrally shaped according to their hearing thresholds. Younger adults with normal hearing listened to speech that simulated the hearing-impaired group's (a) reduced audibility, (b) increased speech levels, and (c) spectral shaping. Group comparisons were made based on speech recognition performance and masking release. Additionally, younger adults completed measures of listening effort and perceived speech quality to assess if differences across simulations in these outcome measures were similar to those for speech recognition. Results Across the various simulations employed, testing in the presence of a threshold matching noise best matched differences in speech recognition and masking release between younger and older adults. This result remained consistent across the other two outcome measures. Conclusions A combination of audibility, speech level, and spectral shape factors is required to simulate differences between listeners with normal and impaired hearing in recognition, listening effort, and perceived speech quality. The use of spectrally shaped and amplified speech in the presence of threshold matching noise best provided this simulated control. Supplemental Material https://doi.org/10.23641/asha.13224632.
| S-EPMC8608154 | biostudies-literature
Project description:Although the default state of the world is that we see and hear other people talking, there is evidence that seeing and hearing ourselves rather than someone else may lead to visual (i.e., lip-read) or auditory "self" advantages. We assessed whether there is a "self" advantage for phonetic recalibration (a lip-read driven cross-modal learning effect) and selective adaptation (a contrastive effect in the opposite direction of recalibration). We observed both aftereffects as well as an on-line effect of lip-read information on auditory perception (i.e., immediate capture), but there was no evidence for a "self" advantage in any of the tasks (as additionally supported by Bayesian statistics). These findings strengthen the emerging notion that recalibration reflects a general learning mechanism, and bolster the argument that adaptation depends on rather low-level auditory/acoustic features of the speech signal.
| S-EPMC6440388 | biostudies-literature
Project description:BackgroundThe gold standard in dysarthria assessment involves subjective analysis by a speech-language pathologist (SLP). We aimed to investigate the feasibility of dysarthria assessment using automatic speech recognition.MethodsWe developed an automatic speech recognition based software to assess dysarthria severity using hidden Markov models (HMMs). Word-specific HMMs were trained using the utterances from one hundred healthy individuals. Twenty-eight patients with dysarthria caused by neurological disorders, including stroke, traumatic brain injury, and Parkinson's disease were participated and their utterances were recorded. The utterances of 37 words from the Assessment of Phonology and Articulation for Children test were recorded in a quiet control booth in both groups. Patients were asked to repeat the recordings for evaluating the test-retest reliability. Patients' utterances were evaluated by two experienced SLPs, and the consonant production accuracy was calculated as a measure of dysarthria severity. The trained HMMs were also employed to evaluate the patients' utterances by calculating the averaged log likelihood (aLL) as the fitness of the spoken word to the word-specific HMM.ResultsThe consonant production accuracy reported by the SLPs strongly correlated (r = 0.808) with the aLL, and the aLL showed excellent test-retest reliability (intraclass correlation coefficient, 0.964).ConclusionThis leads to the conclusion that dysarthria assessment using a one-word speech recognition system based on word-specific HMMs is feasible in neurological disorders.
| S-EPMC6449601 | biostudies-literature
Project description:The presented paper introduces principal component analysis application for dimensionality reduction of variables describing speech signal and applicability of obtained results for the disturbed and fluent speech recognition process. A set of fluent speech signals and three speech disturbances-blocks before words starting with plosives, syllable repetitions, and sound-initial prolongations-was transformed using principal component analysis. The result was a model containing four principal components describing analysed utterances. Distances between standardised original variables and elements of the observation matrix in a new system of coordinates were calculated and then applied in the recognition process. As a classifying algorithm, the multilayer perceptron network was used. Achieved results were compared with outcomes from previous experiments where speech samples were parameterised with the Kohonen network application. The classifying network achieved overall accuracy at 76% (from 50% to 91%, depending on the dysfluency type).
| S-EPMC8749906 | biostudies-literature
Project description:OBJECTIVES:Combined electric-acoustic stimulation (EAS) is a well-accepted therapeutic treatment for cochlear implant (CI) users with residual hearing in the low frequencies but severe to profound hearing loss in the high frequencies. The recently introduced SONNETeas audio processor offers different microphone directionality (MD) settings and wind noise reduction (WNR) as front-end processing. The aim of this study was to compare speech perception in quiet and noise between two EAS audio processors DUET 2 and SONNETeas, to assess the impact of MD and WNR on speech perception in EAS users in the absence of wind. Furthermore, subjective rating of hearing performance was registered. METHOD:Speech perception and subjective rating with SONNETeas or DUET 2 audio processor were assessed in 10 experienced EAS users. Speech perception was measured in quiet and in a diffuse noise setup (MSNF). The SONNETeas processor was tested with three MD settings omnidirectional/natural/adaptive and with different intensities of WNR. Subjective rating of auditory benefit and sound quality was rated using two questionnaires. RESULTS:There was no significant difference between DUET 2 and SONNETeas processor using the omnidirectional microphone in quiet and in noise. There was a significant improvement in SRT with MD settings natural (2.2 dB) and adaptive (3.6 dB). No detrimental effect of the WNR algorithm on speech perception was found in the absence of wind. Sound quality was rated as "moderate" for both audio processors. CONCLUSIONS:The different MD settings of the SONNETeas can provide EAS users with better speech perception compared to an omnidirectional microphone. Concerning speech perception in quiet and quality of life, the performance of the DUET 2 and SONNETeas audio processors was comparable.
| S-EPMC6402691 | biostudies-literature
Project description:Cochlear implant users report difficulty understanding speech in both noisy and reverberant environments. Electric-acoustic stimulation (EAS) is known to improve speech intelligibility in noise. However, little is known about the potential benefits of EAS in reverberation, or about how such benefits relate to those observed in noise. The present study used EAS simulations to examine these questions. Sentences were convolved with impulse responses from a model of a room whose estimated reverberation times were varied from 0 to 1 sec. These reverberated stimuli were then vocoded to simulate electric stimulation, or presented as a combination of vocoder plus low-pass filtered speech to simulate EAS. Monaural sentence recognition scores were measured in two conditions: reverberated speech and speech in a reverberated noise. The long-term spectrum and amplitude modulations of the noise were equated to the reverberant energy, allowing a comparison of the effects of the interferer (speech vs noise). Results indicate that, at least in simulation, (1) EAS provides significant benefit in reverberation; (2) the benefits of EAS in reverberation may be underestimated by those in a comparable noise; and (3) the EAS benefit in reverberation likely arises from partially preserved cues in this background accessible via the low-frequency acoustic component.
| S-EPMC3283901 | biostudies-other
Project description:Automated speech recognition (ASR) systems, which use sophisticated machine-learning algorithms to convert spoken language to text, have become increasingly widespread, powering popular virtual assistants, facilitating automated closed captioning, and enabling digital dictation platforms for health care. Over the last several years, the quality of these systems has dramatically improved, due both to advances in deep learning and to the collection of large-scale datasets used to train the systems. There is concern, however, that these tools do not work equally well for all subgroups of the population. Here, we examine the ability of five state-of-the-art ASR systems-developed by Amazon, Apple, Google, IBM, and Microsoft-to transcribe structured interviews conducted with 42 white speakers and 73 black speakers. In total, this corpus spans five US cities and consists of 19.8 h of audio matched on the age and gender of the speaker. We found that all five ASR systems exhibited substantial racial disparities, with an average word error rate (WER) of 0.35 for black speakers compared with 0.19 for white speakers. We trace these disparities to the underlying acoustic models used by the ASR systems as the race gap was equally large on a subset of identical phrases spoken by black and white individuals in our corpus. We conclude by proposing strategies-such as using more diverse training datasets that include African American Vernacular English-to reduce these performance differences and ensure speech recognition technology is inclusive.
| S-EPMC7149386 | biostudies-literature
Project description:In the real world, human speech recognition nearly always involves listening in background noise. The impact of such noise on speech signals and on intelligibility performance increases with the separation of the listener from the speaker. The present behavioral experiment provides an overview of the effects of such acoustic disturbances on speech perception in conditions approaching ecologically valid contexts. We analysed the intelligibility loss in spoken word lists with increasing listener-to-speaker distance in a typical low-level natural background noise. The noise was combined with the simple spherical amplitude attenuation due to distance, basically changing the signal-to-noise ratio (SNR). Therefore, our study draws attention to some of the most basic environmental constraints that have pervaded spoken communication throughout human history. We evaluated the ability of native French participants to recognize French monosyllabic words (spoken at 65.3 dB(A), reference at 1 meter) at distances between 11 to 33 meters, which corresponded to the SNRs most revealing of the progressive effect of the selected natural noise (-8.8 dB to -18.4 dB). Our results showed that in such conditions, identity of vowels is mostly preserved, with the striking peculiarity of the absence of confusion in vowels. The results also confirmed the functional role of consonants during lexical identification. The extensive analysis of recognition scores, confusion patterns and associated acoustic cues revealed that sonorant, sibilant and burst properties were the most important parameters influencing phoneme recognition. . Altogether these analyses allowed us to extract a resistance scale from consonant recognition scores. We also identified specific perceptual consonant confusion groups depending of the place in the words (onset vs. coda). Finally our data suggested that listeners may access some acoustic cues of the CV transition, opening interesting perspectives for future studies.
| S-EPMC3834087 | biostudies-literature
Project description:This paper describes open source software (available at https://github.com/robotology/natural-speech) to build automatic speech recognition (ASR) systems and run them within the YARP platform. The toolkit is designed (i) to allow non-ASR experts to easily create their own ASR system and run it on iCub and (ii) to build deep learning-based models specifically addressing the main challenges an ASR system faces in the context of verbal human-iCub interactions. The toolkit mostly consists of Python, C++ code and shell scripts integrated in YARP. As additional contribution, a second codebase (written in Matlab) is provided for more expert ASR users who want to experiment with bio-inspired and developmental learning-inspired ASR systems. Specifically, we provide code for two distinct kinds of speech recognition: "articulatory" and "unsupervised" speech recognition. The first is largely inspired by influential neurobiological theories of speech perception which assume speech perception to be mediated by brain motor cortex activities. Our articulatory systems have been shown to outperform strong deep learning-based baselines. The second type of recognition systems, the "unsupervised" systems, do not use any supervised information (contrary to most ASR systems, including our articulatory systems). To some extent, they mimic an infant who has to discover the basic speech units of a language by herself. In addition, we provide resources consisting of pre-trained deep learning models for ASR, and a 2.5-h speech dataset of spoken commands, the VoCub dataset, which can be used to adapt an ASR system to the typical acoustic environments in which iCub operates.
| S-EPMC7805979 | biostudies-literature