v1.1.0

• Instrument response correction will now always try to use poles/zeros to remove
  response, and only use the simpler sensitivity method if no poles/zeros are available.
• The pattern matching on excluding FDSN records is improved to better handle wildcards,
  and can be applied to channel, network, station, and location codes.
• Fixed a bug where zeros could be returned by the Konno-Omachi smoothing algorithm.
  Now it will return nans if there's no point within the smoothing window, and there's
  an option to zero-pad the data to ensure that there are points in all windows.
• Fixed bug in Raoof et al. attention model (that is used for fitting a Brune spectra
  to the signal spectra)
• Added a goodness-of-fit measure of the Brune spectra to the signal spectra.
• Fixed bug in station map that allowed the extent to be too small if there's only one
  station and it is very close to the earthquake.
• Fixed bugs in code for reading the workspace file into Matlab that arose because it
  was not kept up to date with other changes in the code.
• Relaxed the restrictions placed on the allowed IMT/IMC combinations.
• Fixed a bug in how the upsampling was done when computing high-frequency response
  spectra. The code now does the full upsampling method as recommended by Boore and
  Goulet (2014; DOI 10.1007/s10518-013-9574-9), and we added a test of low sample rate
  records against the high-frequency response spectra reported by NGA East for the same
  record.