| The recording procedure will unavoidably be interfered by surrounding environment, because of the existence of environmental noise, the speech recorded in the storing media is no longer the original pure speech signal, and it has become noisy speech being contaminated by noise. In many occasions of voice monitoring, the contamination of background noise is very serious, and human speech takes only a small part of the storage space. If noise suppression measures not taken, not only voice definition will suffer remarkable decrease, storage space will also be wasted significantly. So it is very necessary to suppress noise in recording system, so as to improve recording timbre and storing efficiency.Towards the specific application field of digital recording, on the basis of the analyses of speech and noise characteristics, this dissertation proposes a refinement algorithm which aims at detecting silence voice, and it has solved the implementation problem of a preferable algorithm in speech enhancement. Finally, this dissertation designs an intelligent recording system based on ARM embedding system.The first chapter of this dissertation gives an introduction of research and application situation at home and abroad, and discusses the significance of the application of this technology in engineering areas. The second chapter studies various characteristics of speech and noise and various methods in speech signal processing, it also studies how to build a simulation system to analyze and process speech signal. On the basis of the study on the existing silence detection algorithms, the third chapter proposes an improved algorithm. The forth chapter studies the speech enhancement algorithm and its refinement algorithm towards noisy speech. The fifth chapter studies the systems solutions design, hardware design and software design of an intelligent recording system based on ARM embedding system. The sixth chapter summarizes the dissertation, and puts forward the reformative job latter. |
PDF Full Text Request