Feature Extraction and Performance Comparison for ANN-based Speaker Identification Systems