Recording Aperture and the Selection of Filter Bandwidth for bvas and bamp

As one might expect, there is a limit to how finely the spectrum may be investigated. This depends on the available aperture, which follows from the choice the user made in recording the field data. If there are N samples in each seismic trace, and if the sample interval is $\Delta t$ , then the filter bandwidth should not be less than $\frac{1}{N\Delta t}$ . Further, the frequency step size (independent of the bandwidth specification) should be no smaller than the bandwidth. Both programs, bvas and bamp, should be run with the same spectral bandwidth. However, the program cainv3.m is able to handle different total numbers of amplitude and velocity samples in the frequency domain. Thus, it is OK to be missing either some decay or velocity samples. Further, the sampled frequencies do not have to be identical for both velocity and decay. What should be the same is the bandwidth of the filters.

It is important to record enough data (large enough aperture), if the velocity measurements are to be considered phase velocity, and not group velocity. With hammer sources on soil, I recommend recording no less than 0.5 seconds of data (translates to a bandwidth of 2 Hz).