Frequency-wavenumber velocity filtering is often applied to prestack seismic data for the attenuation of coherent noise. Although the process often gives excellent results, it can sometimes result in signal smoothing and distortion and poor attenuation of coherent noise. A slowness adaptive f-k filter reduces signal distortion and improves the attenuation characteristics of the filter. The technique uses a time- and space-variant narrow reject-band f-k filter. Optionally, coherent noise is compressed before application of the filter.The apparent slowness of coherent noise events is estimated using local t-x slant stacks weighted by coherence. A two-dimensional (2-D) window is moved across the shot record, and at each point on the record slant stacks are taken through the central sample of the window. The slowness value that produces the maximum stack is assigned to the central sample of the window. In this way, an instantaneous slowness image of the shot record is produced. A one-dimensional (1-D), high-pass, finite-duration impulse-response (FIR) filter is applied in a spatially and temporally varying way across the record on the basis of the instantaneous slowness values. Before filter application, trace-to-trace static and amplitude effects are estimated and removed from the data. This results in compression of coherent noise and improved attenuation after filtering.The filtering process has been applied to low-fold prestack dynamite data from the Surat Basin, Australia. The results indicate that the technique has good attenuation characteristics and produces minimal distortion of seismic signal. The process, however, is computationally expensive.