• Corpus ID: 31754948

Real-Time Wind Noise Detection and Suppression with Neural-Based Signal Reconstruction for Mult-Channel, Low-Power Devices

@article{Rhodes2017RealTimeWN,
  title={Real-Time Wind Noise Detection and Suppression with Neural-Based Signal Reconstruction for Mult-Channel, Low-Power Devices},
  author={Anthony D. Rhodes},
  journal={ArXiv},
  year={2017},
  volume={abs/1710.00082}
}
  • A. Rhodes
  • Published 29 September 2017
  • Computer Science
  • ArXiv
Active wind noise detection and suppression techniques are a new and essential paradigm for enhancing ASR-based functionality with smart glasses, in addition to other wearable and smart devices in the broader IoT (Internet of things. [] Key Method In the first case, we advance a real-time wind detection algorithm (RTWD) that uses two distinct sets of low-dimensional signal features to discriminate the presence of wind noise with high accuracy.
View PDF on arXiv

Figures from this paper

References

SHOWING 1-10 OF 39 REFERENCES

Wind Noise Detection : Signal Processing Concepts for Speech Communication

Three new wind detection concepts are presented based on the shortterm mean introduced by wind noise in a recorded signal while the characteristic spectral shape and energy distribution of wind noise are exploited in the second and third approach.

Single microphone wind noise PSD estimation using signal centroids

Investigations with measured audio data show that the proposed method can cope with the non-stationary characteristics of wind noise and enables a sufficient reduction ofWind noise, and the proposed algorithm has low complexity and low memory consumption.

Coherence-based Dual Microphone Wind Noise Reduction by Wiener Filtering

A two-step method to reduce wind noise in dual microphone environments by exploiting the coherence of input signals and using a Wiener filter to noise frequency regions is proposed.

Dual MicrophoneWind Noise Reduction by Exploiting the Complex Coherence

This contribution presents methods for detecting, estimating and reducing the effect of wind noise picked up by a dual microphone communication device and an approach for the enhancement of distorted speech which uses the phase variance and the magnitude of the complex cross power spectral density (PSD).

Wind Noise Reduction using Non-Negative Sparse Coding

A new speaker independent method for reducing wind noise in single-channel recordings of noisy speech based on non-negative sparse coding and relies on a wind noise dictionary which is estimated from an isolated noise recording is introduced.

Phase Processing for Single-Channel Speech Enhancement: History and recent advances

This work focuses on single-channel speech enhancement algorithms which rely on spectrotemporal properties, and can be employed when the miniaturization of devices only allows for using a single microphone.

Neural network based spectral mask estimation for acoustic beamforming

A neural network based approach to acoustic beamforming is presented, used to estimate spectral masks from which the Cross-Power Spectral Density matrices of speech and noise are estimated, which are used to compute the beamformer coefficients.

Robustness of SVD-based optimal filtering for noise reduction in multi-microphone speech signals

It is shown that the SNR improvement provided by the SVD-base d optimal filtering technique is better than the improvement o btained with standard beamforming techniques.

Combining spectral feature mapping and multi-channel model-based source separation for noise-robust automatic speech recognition

This work explores the effectiveness of first applying a model-based source separation mask to the output of a beamformer that combines the source signals recorded by each microphone, followed by a DNN-based front end spectral mapper that predicts clean filterbank features.

A Regression Approach to Speech Enhancement Based on Deep Neural Networks

The proposed DNN approach can well suppress highly nonstationary noise, which is tough to handle in general, and is effective in dealing with noisy speech data recorded in real-world scenarios without the generation of the annoying musical artifact commonly observed in conventional enhancement methods.