We have applied a modern amplitude- and azimuth-preserving seismic data processing workflow to the SEG Advanced Modeling Program (SEAM) Phase II Barrett classic data set — an orthorhombic synthetic seismic model that has extremely dense sampling of all azimuths and offsets. We analyze the resulting prestack depth-migrated offset vector tiles with a variety of methods and software. Note that we only analyze the P-P wave mode, which is the focus of our study. We demonstrate that observed azimuthal changes cannot be correlated with the model's reservoir properties. We have made the migrated data available through SEAM. Compared to modeled data, real onshore seismic data have significantly lower amplitude fidelity, higher noise levels, and more uncertainty in the migration velocity field used for processing. Since we are unable to relate the anisotropy measured from the fully sampled clean SEAM Phase II Barrett synthetic seismic data to the model's known anisotropy, we conclude that it is highly unlikely that azimuthal variations observed on real onshore seismic data will be predictive of reservoir fracture properties.