The basic problem of speech is a complete separation of the natural noise which arise from correct articulation of voiced and unvoiced consonants from noise and disturbance environment. Objective of this master’s thesis is to find an effective method that could identify the pauses without speech activity, which can identify the properties of noise and disturbance. Once the noise is correctly identified, it is already possible to use different methods for its removal. The master’s thesis describes two methods of pauses identification. These methods are programmed in Matlab and tested on nine speech recordings. Methods analysis of the results was performed using the ROC (Receiver Operating Characteristic) curves. In the end are summarized results analysis of created methods.
Identifer | oai:union.ndltd.org:nusl.cz/oai:invenio.nusl.cz:218837 |
Date | January 2011 |
Creators | Kepák, Petr |
Contributors | Míča, Ivan, Smékal, Zdeněk |
Publisher | Vysoké učení technické v Brně. Fakulta elektrotechniky a komunikačních technologií |
Source Sets | Czech ETDs |
Language | Czech |
Detected Language | English |
Type | info:eu-repo/semantics/masterThesis |
Rights | info:eu-repo/semantics/restrictedAccess |
Page generated in 0.0787 seconds