In my last QC-related post, I looked at response files and the particular problem of supplying an incorrect gain. But as we know, it’s not just the gain that can be wrong in a response file, other culprits abound.
In this post, I’ll look at two other aspects of a response file that must be right in order to produce expected and accurate results when using them: the poles and zeroes.
But first, a reminder of what the PSD PDF looks like when the response file is correct:
What happens when a required pole is missing? Removing the first pole of the required four, in this plot, a clear trend of slanting to the upper right can be seen:
And when we remove the first zero of the required two? Here we see a very obvious slanting trend toward the bottom left:
Traditionally, both the HNM and LNM baselines are used to assess noise levels and whether or not a station’s performance conforms to the expected natural results. In this case, however, we can use these two baselines somewhat differently by asking the question: does the overall shape of our resulting PSD PDF match that of the HNM and LNM baselines?
Because when the shape as well does not conform, then we know something with the meta-data must be wrong. In these two cases, either a missing pole, or a missing zero from the response file.
More QC issues to come…