Phoneme-Dependent Speech Enhancement
نویسندگان
چکیده
The majority of current speech enhancement systems are based on generalized signal-to-noise ratio dependent weighting rules and do not take into account the characteristics of the actual speech sound being processed. The following contribution is concerned with phoneme-specific speech enhancement methods that apply specially tailored signal processing methods. The first signal processing algorithm proposed in this work – fricative spreading – enhances high frequency unvoiced sounds for bandlimited speech transmission. The spreading algorithm detects different fricatives using a vector quantization codebook and then a suitable spectral compression function is applied to map high frequency energy from above the transmission bandwidth threshold into lower frequency regions still within the transmission bandwidth. A second approach – formant boosting – provides enhancement for voiced speech. Utilizing the codebook classification from fricative spreading, voiced speech phonemes are identified and accentuated by boosting formant regions and attenuating in between the formant frequencies.
منابع مشابه
Allophone-based acoustic modeling for Persian phoneme recognition
Phoneme recognition is one of the fundamental phases of automatic speech recognition. Coarticulation which refers to the integration of sounds, is one of the important obstacles in phoneme recognition. In other words, each phone is influenced and changed by the characteristics of its neighbor phones, and coarticulation is responsible for most of these changes. The idea of modeling the effects o...
متن کاملPhoneme-Dependent NMF for Speech Enhancement in Monaural Mixtures
The problem of separating speech signals out of monaural mixtures (with other non-speech or speech signals) has become increasingly popular in recent times. Among the various solutions proposed, the most popular methods are based on compositional models such as non-negative matrix factorization (NMF) and latent variable models. Although these techniques are highly effective they largely ignore ...
متن کاملMinimum cost based phoneme class detection for improved iterative speech enhancement
It is known that degrading acoustic noise innuences speech quality across phoneme classes in a non-uniform manner. This results in variable quality performance for many speech enhancement algorithms in noisy environments. To address this, a hidden-Markov-model phoneme classiica-tion procedure is proposed which directs single channel speech enhancement across individual phoneme classes. The proc...
متن کاملMarkov Model Based Phoneme Class Partitioning for ImprovedConstrained Iterative Speech
Research has shown that degrading acoustic background noise innuences speech quality across phoneme classes in a non-uniform manner. This results in variable quality performance of many speech enhancement algorithms in noisy environments. A phoneme classiication procedure is proposed which directs single-channel constrained speech enhancement. The procedure performs broad phoneme class partitio...
متن کاملSpeech Enhancement using a Deep Mixture of Experts
In this study we present a Deep Mixture of Experts (DMoE) neural-network architecture for single microphone speech enhancement. By contrast to most speech enhancement algorithms that overlook the speech variability mainly caused by phoneme structure, our framework comprises a set of deep neural networks (DNNs), each one of which is an ‘expert’ in enhancing a given speech type corresponding to a...
متن کامل