Abstarct: In recent years, methods of operation, processing and interpretation of seismic data has improved tremendously and allows accurate visualization of complex geological structures have been provided. Sometimes the seismic response of geologic events in a wide range of frequency bandwidth is hidden. In this context, the spectral analysis method enables the interpreter that the frequency components of the seismic interpretation and understanding to collateralize elaborate underground. Spectral analysis, which refers to all ways that provide analysis of the time - frequency. Since seismic signals are unstable, this method as a powerful tool in the analysis of these signals is widely used. Spectral decomposition of seismic data, seismic amplitudes as a function of time and place are converted to spectral ranges as a function of frequency, time and location. The tools in the late 90s and since it was introduced as a means of interpretation in many areas, such as laxyer thickness, show a stratigraphic facies, described the reservoir characteristics and hydrocarbon exploration has been used to direct. In recent decades, various algorithms in order to display time - frequency from astable signals are presented. The most important of these algorithms can be named short time Fourier transform, continuous wavelet transform, S transform and …, that each of them have own advantages and disadvantages. In this paper introduces the time – frequency transform and review of advantages and disadvantages after that time – time transforms are introduced that the inverse calculation of Fourier transform from the S transform central frequency are obtained. It's become a one-dimensional signal to the two-dimensional signal which is a function of time. This method rather than other time - Frequency methods have high temporal resolution. The frequency changes in this transform is measured by qualitative. The purpose of introducing this transformation is identify capabilities and use it to improve the processing and interpretation of seismic data, In this thesis with using empirical mode decomposition in the time – time domain to undermine effect of ground roll from real seismic data is studied.