Speaker Indentification Based On An Integrated System Combining Cepstral Feature Extraction And Vector Quantization