Zaslat SMS: Performance Analysis of Speaker Features Extracted from High-order Fractional Domains