Bandwidth Enhancement - Inverse Q Deconvolution or Time-varying Wiener Filtering?

Both dispersion and attenuation corrections can improve the resolution of seismic data. This significantly facilitates interpretation. In principle, both inverse Q deconvolution and time-varying Wiener filtering can achieve this. Inverse Q deconvolution is a deterministic process that requires knowledge of the quality factor Q, whereas time-varying Wiener filtering is a statistical approach based on the estimation of the nonstationary propagating wavelet. Dispersion corrections based on phase-only inverse Q deconvolution is an inherently stable method that is robust in the presence of noise. Attenuation corrections via amplitude-only inverse Q deconvolution on the other hand is likely to lead to noise amplification as well as bandwidth enhancement. Dispersion corrections via time-varying Wiener filtering are challenging since this requires estimation of a nonstationary, frequency-dependent, nonminimum-phase wavelet. Fortunately attenuation corrections via Wiener filtering need only estimation of a zero-phase time-varying wavelet for which robust methods exist. The optimum procedure for combined dispersion and attenuation correction is thus comprised of first applying dispersion corrections using phase-only inverse Q deconvolution, followed by zero-phase time-varying Wiener filtering.