The specific mechanical constant of a material can be determined by analyzing its dynamic resonance or dynamic pulse response. This paper proposes the development and the application of an alternative statistical analysis method called Multilevel Integrated Kurtosis Algorithm with Z-notch Filter (I-kaz Multilevel) signal analysis used on dynamic resonance signals to predict the value of specific mechanical properties of the test materials. Four rectangular bars from four different materials were excited in an anechoic room by impact test which produce transient sound radiation and vibration responses. The signals were recorded using microphone and accelerometer sensor respectively. An alternative Z-stem filtering technique was used on the captured signals to remove noise. The filtered signals were analyzed using I-kaz Multilevel signal analysis to determine the pattern of the signal and to estimate the significance differences between the four materials. The experimental power curve equations of mechanical loss coefficient vs. linear coefficient and compressive strength vs. linear equation can be used to predict the unknown values of mechanical loss coefficient and compressive strength of a material.