Guided waves have proven their sensitivity to changes in thickness and porosity in the cortical bone. The soft tissue (ST) layer covering the bone can distort the wave pattern making it difficult to unambiguously interpret the results. Thus, the search for invariant parameters is relevant. Ultrasonic signals at 100 and 450 kHz were acquired by surface profiling of PMMA bone phantoms simulating the bone condition in the scale “norm (N) – osteoporosis (OP)” [1]. 2D waveform signals profiles were formed after low pass filter enveloping (Figure 1). The method for estimating bone condition in the scale N-OP accounting ST thickness is based on the use of the discrete Fourier transform (DFT) of the propagated signals. The resulting complex signal used magnitude values in the selected frequency range. According to them, the following statistical features were calculated: 1) frequency range in which the magnitude value is greater than the average; 2) ratio of maximum envelope values for the magnitudes; 3) the normalized estimate of the derivative of the envelope. The obtained signs allow us to give an approximate assessment of the new pattern using the pattern recognition methods.