Earthquake accelerograms are needed for dynamic analysis of structures. Preliminary data obtained from recording instruments may include some errors that their sources and values should be evaluated and corrected appropriately. In this study, at first step, earthquake recording instruments are introduced and explained. Then the noise (error) sources of accelerograms and processing methods of earthquake data are reviewed. Filtering techniques are also assessed from the view point of Earthquake Engineering. For this purpose, different methods for the determination of correction frequency are reviewed and finally the standard model introduced by Brune (1970) is selected. In this study, 78 earthquake records are selected from Mahdavian s study. In this selection, two parameters are considered:
(1) Causative earthquake of the record and (2) The soil type of at recording station. A computer program is developed to calculate Fourier amplitude spectra of three recorded components and to plot them. By comparing the calculated spectra and the standard -2 model, correction frequency is determined and then records are corrected using SWS processing software. Effect of different parameters like magnitude, epicentral distance and local site conditions are evaluated. Correction frequencies are estimated and Fourier amplitude spectra are plotted before and after the correction. Finally, several correction frequency bands are presented for different site conditions of recording stations.
)