摘要 |
PURPOSE: A method and apparatus for extending a bandwidth of a voice signal are provided to perform a different operation for each sound by determining a voiced sound and an unvoiced sound, thereby reducing an operation quantity and improving voice quality. CONSTITUTION: A bandwidth expansion device determines a voiced or unvoiced section from a received voice signal(S130). Based on the voiced section, the device extracts a first section which includes a harmonic content of the voiced section from a normalized voice signal(S140). Based on a correlation between the unvoiced section and the normalized voice signal, the device extracts a second section from the normalized voice signal(S135). The device outputs a wideband voice signal by the synthesis of a generated high-band voice signal and the transformed voice signal(S170,S180). [Reference numerals] (AA) Start; (BB) End; (S100) Receive a narrow band voice signal; (S110) Convert into a frequency domain; (S120) Normalize the converted voice signal; (S130) Determine a voiced/unvoiced section(voiced/unvoiced section?); (S135) Extract a second section according to correlation; (S140) Extract a first section including a harmonic section; (S150) Amplify and transit a bandwidth; (S160) Control energy scale; (S170) Obtain a wideband signal by synthesis of a low band signal and a high band signal; (S180) Output a wideband voice signal |