Fix the event filter that filters on side-band acoustic activity #390
Labels
Comments
|
Closing as invalid. No details in body. I'm going to assume this is a mistake. |
towsey
added a commit
that referenced
this issue
Oct 15, 2020
Issue #390 Separate out the lower and upper side band activity so as to filter on the two separately rather than combined as previously.
towsey
added a commit
that referenced
this issue
Oct 16, 2020
towsey
added a commit
that referenced
this issue
Oct 17, 2020
Issue #390 Split the GenericRecognizer class into processing and post-processing classes. This was so I could more neatly deal with a change to the use of multiple decibel thresholds.
towsey
added a commit
that referenced
this issue
Oct 17, 2020
Issue #390 Added the analysis threshold as an argument to method dealing sideband activity. THis was an experiment. In the end did not use the argument but left , just in case might be used in future.
towsey
added a commit
that referenced
this issue
Oct 17, 2020
Issue #390 Add another means to filter events by removing an event which is entirely enclosed by another event. This circumstance now arises due to the new way of using decibel thresholds.
towsey
added a commit
that referenced
this issue
Oct 17, 2020
Issue #390 Due to implementation of a new way of finding events with multiple decibel thresholds, have had to change the Algorithms used by Generic recognizers. So far only done one of hte seven algorithms.
towsey
added a commit
that referenced
this issue
Oct 24, 2020
Issue #390 Moved some classes into separate folders, according to their functionality.
towsey
added a commit
that referenced
this issue
Oct 24, 2020
Issue #390 Write unit test for an event filter that removes events enclosed by another event.
towsey
added a commit
that referenced
this issue
Oct 24, 2020
Issue #390 Fix bug that occured in this method and for which new unit test was written.
towsey
added a commit
that referenced
this issue
Oct 24, 2020
Issue #390 A bug appeared when first implementing multiple decibel thresholds. The loop over an array of thresholdholds was implemented only around the profile matching and did not include the post processing. This caused the recognizer to miss many true calls. The corrections implemented here are to encusre that the lopp over all decibel thresholds includes both the profile detection and the post processing in each pass.
towsey
added a commit
that referenced
this issue
Oct 24, 2020
Issue #390 Following on from the previous commit, it was necessary to pass a decibel throld to isse run of a profile.
towsey
added a commit
that referenced
this issue
Oct 24, 2020
towsey
added a commit
that referenced
this issue
Oct 24, 2020
Issue #390 THe array of decibel thresholds is shifted up one level in the config hierarchy.
towsey
added a commit
that referenced
this issue
Oct 24, 2020
towsey
added a commit
that referenced
this issue
Oct 27, 2020
Issue #390 I changed the way the event filter is working that filters on acoustic activity in the sidebands. I changed the name of a variable and its function. The sidebnad filter now checks for two kinds of sideband activity - the average background decibels in the sidebands, and it looks for brief events in the sidebands.
towsey
added a commit
that referenced
this issue
Oct 27, 2020
Issue #390 I changed the scope of the loop over decibel threshold values. This was to fix a bug in how events were subsequently merged.
towsey
added a commit
that referenced
this issue
Oct 27, 2020
towsey
added a commit
that referenced
this issue
Oct 28, 2020
Issue #390 Change recognizer.config.yml files to reflect changes that allow user to set different threshold array for each profile.
towsey
added a commit
that referenced
this issue
Oct 28, 2020
Issue #390 The TestMultipleAlgoritnms unit test picked up a bug in the naming of events. This was fixed by shifting assignment of event names to inside the relevant for-loop.
towsey
added a commit
that referenced
this issue
Oct 30, 2020
Issue #390 While writing tests for unit methods made some changes to boundary conditions for hertz and duration values. More unit tests could be done.
towsey
added a commit
that referenced
this issue
Oct 31, 2020
Issue #390 Fixed bug which meant that plots were not being returned with events.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Separate out the lower and upper side band activity so as to filter on the two separately rather than combined as previously.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Split the GenericRecognizer class into processing and post-processing classes. This was so I could more neatly deal with a change to the use of multiple decibel thresholds.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Added the analysis threshold as an argument to method dealing sideband activity. THis was an experiment. In the end did not use the argument but left , just in case might be used in future.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Add another means to filter events by removing an event which is entirely enclosed by another event. This circumstance now arises due to the new way of using decibel thresholds.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Due to implementation of a new way of finding events with multiple decibel thresholds, have had to change the Algorithms used by Generic recognizers. So far only done one of hte seven algorithms.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Moved some classes into separate folders, according to their functionality.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Write unit test for an event filter that removes events enclosed by another event.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Fix bug that occured in this method and for which new unit test was written.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 A bug appeared when first implementing multiple decibel thresholds. The loop over an array of thresholdholds was implemented only around the profile matching and did not include the post processing. This caused the recognizer to miss many true calls. The corrections implemented here are to encusre that the lopp over all decibel thresholds includes both the profile detection and the post processing in each pass.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Following on from the previous commit, it was necessary to pass a decibel throld to isse run of a profile.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 THe array of decibel thresholds is shifted up one level in the config hierarchy.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 I changed the way the event filter is working that filters on acoustic activity in the sidebands. I changed the name of a variable and its function. The sidebnad filter now checks for two kinds of sideband activity - the average background decibels in the sidebands, and it looks for brief events in the sidebands.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 I changed the scope of the loop over decibel threshold values. This was to fix a bug in how events were subsequently merged.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Change recognizer.config.yml files to reflect changes that allow user to set different threshold array for each profile.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 The TestMultipleAlgoritnms unit test picked up a bug in the naming of events. This was fixed by shifting assignment of event names to inside the relevant for-loop.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 While writing tests for unit methods made some changes to boundary conditions for hertz and duration values. More unit tests could be done.
atruskie
pushed a commit
that referenced
this issue
Nov 1, 2020
Issue #390 Fixed bug which meant that plots were not being returned with events.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The operating system I'm using is:
The version of AnalysisPrograms.exe I'm using is:
The data used can be found:
The command I used:
The config file I used:
Screenshots
Additional Details
The text was updated successfully, but these errors were encountered: