Abstract Most hydrocarbon fields found within the Ordovician Trenton and Black River Groups of the eastern United States and eastern Canada are associated with basement-related faults. These faults are best imaged by 3D seismic technology. Because of environmental conditions, seismic data often are contaminated by noise that masks fault terminations and reduces signal-to-noise ratio. As a result, seismic horizons are discontinuous, calculated coherence values are affected adversely, and horizon interpretation and fault identification are difficult, if not impossible. Poststack processing is required to attenuate this noise before an optimal interpretation can be done. We conducted a three-step poststack processing flow to attenuate noise and highlight fault terminations. Noise-reducing algorithms consist of frequency-space (f-x) deconvolution, zero-phase filtering, and τ-p filtering. The structural grain of major faults identified with these techniques agrees with the regional strike of major faults previously defined in the area. These faults were confirmed by drilling results. Our calculated semblance of cosine of phase provides better fault definition than does regular coherency and serves as an alternative attribute for mapping faults. The processing sequence could apply in areas with similar geologic settings and noise problems.