Describir: Application of the Vector Quantization Methods and the Fused MFCC-IMFCC Features in the GMM Based Speaker Recognition