Stress has effect on speech characteristics and can influence the quality of speech. In this paper, we study the effect of Sleep-Deprivation (SD) on speech characteristics and classify Normal Speech (NS) and Sleep Deprived Speech (SDS). One of the indicators of sleep deprivation is ‘flattened voice’. We examine pitch and harmonic locations to analyse flatness of voice. To investigate, we compute the spectral coefficients that can capture the variations of pitch and harmonic patterns. These are derived using Two-Layer Cascaded-Subband Filter spread according to the pitch and harmonic frequency scale. Hidden Markov Model (HMM) is employed for statistical modeling. We use DCIEM map task corpus to conduct experiments. The analysis results show that SDS has less variation of pitch and harmonic pattern than NS. In addition, we achieve the relatively high accuracy for classification of Normal Speech (NS) and Sleep Deprived Speech (SDS) using proposed spectral coefficients.