The effects of acoustic degradation on lexical processing were investigated in children ranging from 12 to 31 months of age. Using a visual fixation technique, familiar target words (in infant-directed speech) were presented either acoustically unaltered, time compressed (50%) or low-pass filtered (1.5kHz). The children's ability to correctly identify the respective target was assessed by exami...

Speech perception relies on a listener's ability to simultaneously resolve multiple temporal features in the speech signal. Little is known regarding neural mechanisms that enable the simultaneous coding of concurrent temporal features in speech. Here we show that two categories of temporal features in speech, the low-frequency speech envelope and periodicity cues, are processed by distinct neu...

Speech output, including modified and synthetic speech, is used increasingly in natural settings where message reception might be affected by noise. Recent evaluations have demonstrated the effect of different speech styles on intelligibility for native listeners, but their impact on listening in a second language is less well-understood. The current study measured the intelligibility of four s...

Speech comprehension relies on temporal cues contained in the speech envelope, and the auditory cortex has been implicated as playing a critical role in encoding this temporal information. We investigated auditory cortical responses to speech stimuli in subjects undergoing invasive electrophysiological monitoring for pharmacologically refractory epilepsy. Recordings were made from multicontact ...

This paper concerns the problem of automatic speech recognition in noise-intense and adverse environments. The main goal of the proposed work is the definition, implementation, and evaluation of a novel noise robust speech signal parameterization algorithm. The proposed procedure is based on time-frequency speech signal representation using wavelet packet decomposition. A new modified soft thre...

We present a description of the development and evaluation of a first South African broadcast news transcription system. We describe a number of speech resources which have been collected in the resource-scarce South African environment for system development purposes: a 20 hour corpus of South African English (SAE) broadcast news; a 109M word corpus of South African newspaper text collected fo...

Speech data for NIST speaker recognition evaluations has traditionally been distributed in compressed, telephone quality form, even for microphone data that was originally recorded at higher quality. We evaluate the effect that improved audio quality has for speaker verification performance, using a recently released full-bandwidth version of microphone data from the SRE2010 evaluation. Remarka...

Stochastic speech recognition has been cast as a natural realization of the compressive sensing problem in this work. The compressed acoustic observations are subword posterior probabilities obtained from a deep neural network. Dictionary learning and sparse recovery are exploited for inference of the high-dimensional sparse word posterior probabilities. This formulation amounts to realization ...

We present the first unsupervised LSTM speech segmenter as a cognitive model of the acquisition of words from unsegmented input. Cognitive biases toward phonological and syntactic predictability in speech are rooted in the limitations of human memory (Baddeley et al., 1998); compressed representations are easier to acquire and retain in memory. To model the biases introduced by these memory lim...

Stochastic speech recognition has been cast as a natural realization of the compressive sensing problem in this work. The compressed acoustic observations are sub-word posterior probabilities obtained from a deep neural network. Dictionary learning and sparse recovery are exploited for inference of the high-dimensional sparse word posterior probabilities. This formulation amounts to realization...