A new method for Q estimation : the generalized central frequency shift

Quality factor Q is a useful tool for hydrocarbon detection and reservoir identification. However, popular methods for Q estimation have their respective limitation in processing real seismic data, for example, the spectral ratio method is lack of stability and the frequency shift methods exist systemic error since the assumption of wavelet types. In this article, we first derive a kind of new Q estimation methods named the generalized central frequency shift (GCFS) without assuming the types of source wavelet based on divergence. Information divergences are introduced in our methods to measure the difference between the amplitude spectrum of the source wavelet and the the amplitude spectrum of the attenuation wavelet. Q estimation formulas based on different divergence(Jeffreys divergence, Renyi divergence) indicate that Q-factor is not only related to divergence, but also to center frequency deviation. The test of seismic synthetic records show the GCFS method has better accuracy and higher robustness compared to other methods.