主权项 |
1. A method of generating an excitation class, the method including:
receiving audio from an input device; first determining, based on a result of signal classification, whether a current frame of the audio corresponds to a speech signal; when the current frame does not correspond to the speech signal, second determining, based on a tonal characteristic, whether the current frame corresponds to one of a noisy signal and a tonal signal; and generating, based on a result of the first determining and the second determining, an excitation class for the current frame, wherein the excitation class for the current frame is one of a plurality of excitation classes, wherein the plurality of excitation classes include a speech class, a tonal music class, and a non-tonal music class. |