Abstract
Mobile devices are often used in our daily lives for the purposes of speech and communication. The speech quality of mobile devices is always degraded due to the environmental noises surrounding mobile device users. Regretfully, an effective background noise reduction solution cannot easily be developed for this speech enhancement problem. Due to these depicted reasons, a methodology is systematically proposed to eliminate the effects of background noises for the speech communication of mobile devices. This methodology integrates a dual microphone array with a background noise elimination algorithm. The proposed background noise elimination algorithm includes a whitening process, a speech modelling method and an H2 estimator. Due to the adoption of the dual microphone array, a low-cost design can be obtained for the speech enhancement of mobile devices. Practical tests have proven that this proposed method is immune to random background noises, and noiseless speech can be obtained after executing this denoise process.
Highlights
The speech communication function of mobile devices has been well-designed and widely used as a convenient tool for contacting others due to its portable characteristics
The speech communications of mobile devices are inevitably affected by varying environment noises
The dual microphone array constructed with an omnidirectional microphone (Omni-Mic) and unidirectional microphone (Uni-Mic) is used for collecting all possible sound sources
Summary
The speech communication function of mobile devices has been well-designed and widely used as a convenient tool for contacting others due to its portable characteristics. The steady state form of the Kalman filter known as the H2 estimator combined with a colored noise whitening process is studied for the speech enhancement problem in this investigation By integrating this modified H2 estimator with a dual microphone array, an effective background noise reduction design that has low calculation power consumption and can effectively mitigate the non-stationary background noises is developed. The overall speech enhancement process of this investigation can be summarized as the following: Step 1, two measured data pairs of microphone channels are whitened as a white noise driven speech sequence This speech sequence is mathematically expressed as a corresponding difference equation which contains a set of identifiable parameters in Step 2 and further formulated as a vector-matrix form in time domain for conveniently deriving the related H2 estimator in Step 3. This proposed method can effectively remove the unknown background noises and deliver promising speech enhancement performances in severe scenarios
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
