Method for enhancing clarity of speech signals

Abstract

The invention relates to a method for enhancing the clarity of speech signals, which comprises the steps of performing frame expansion and windowing processing on audio signals according to an audio coding and decoding format in a CRH train driver control information analysis system, estimating the location of the speech by using a dynamic speech detection algorithm based on the short time energydifference, performing noise estimation by using the non-speech audio signals, finally performing filtering processing by using a wiener filter, carrying out frame cutting on the filtered audio signals to recover back to the original frame length, and performing noise reduction processing in the real-time playing process of the audio, thereby providing a clear and fluent speech analysis environment for the related personnel, and providing convenient conditions for driver control analysis. The method is applied to the CRH train driver control information analysis system, can perform enhancementon the speech signals in the system, has a good suppression effect for noise signals, well restores the frequency spectrum, the tone and the pitch of the speech signals, and keeps the speech playingcontinuity and the timeliness and the correlation of dotting information.

Description

technical field [0001] The invention belongs to the application field of analysis of control information of railway train drivers, in particular to a method for enhancing the clarity of voice signals in the analysis of control information of train drivers. Background technique [0002] With the rapid development of high-speed railways, the management of EMU drivers is becoming more and more difficult. The analysis of the control information of EMU drivers is not only related to the regulation of driver behavior, but also related to the safety of train operation. When analyzing the voice signal in the control information, in order to improve the voice clarity, an effective and feasible voice enhancement technology is urgently needed. [0003] During the operation of the train, affected by the noise generated by the mechanical movement and the complex electromagnetic environment, the recorded voice signal is accompanied by great interference, and the voice in the cab is subme...

AI Technical Summary

Problems solved by technology

[0003] During the operation of the train, affected by the noise generated by the mechanical movement and the complex electromagnetic environment, the recorded voice signal is accompanied by great interference, and the voice in the cab is submerged in the noise and becomes blurred, which brings great difficulties to the analysis of the driver's control information. inconvenient

Generally, the circuit noise reduction filter technology is effective for electromagnetic or high- and low-frequency noise interference, and is ineffective for noise interference in the same frequency band as voice; the method of reducing the gain reduces the gain of the voice signal while reducing the noise, making the voice become very weak

Images