Opal magnetometer data are subject to non-uniformities in the magnetic field occurring in the recording environment. These disturbances or non-uniformities can sometimes affect the processing algorithms, resulting in metric data or joint angles that seem incorrect or impossible. If you are encountering these issues, it may be worth preprocessing these recordings without the magnetometer data to improve the reported metric results.
The following instructions outline how to make copies of a recording so that you can compare the original recording against an updated version that has been reprocessed without the magnetometer, or with a constant value magnetometer.
- First, export .h5 copies of any file you would like to reprocess. It might be a good idea to put these exports in folders by subject to keep them organized depending on how many subjects/recordings you would like to reprocess.
- Change the file naming convention as shown below. In this example, I added “_No_magnetometer” to the file naming convention. We will assume that the default file naming convention (%D_%T), was used for the original recordings with the “variable magnetometer” approach to orientation estimation.
- Navigate a subject and choose "create trial from recording" from the cog icon in the upper-right-hand corner of the window. Choose an *.h5 file that belongs to this subject to be reimported into Moveo Explorer. The file will be renamed according to the new file name convention when copying it into the Moveo workspace.
- Repeat this process by updating the naming convention again to append “_Fixed_magnetometer” instead of “_No_magnetometer” and re-creating the trials from recording for each subject.
- At this point, there should be 3 total trials present in Moveo Explorer for each original file. Each one will have a different backing file name, but they are all still identical at this point regarding the magnetometer, so we will address that next.
- Switch to the Motion Studio view by clicking on View->”Motion Studio View” in the top menu bar.
- Expand the “MobilityLabProject” folder and the monitorData folder within. You should see the 3 copies for each file you imported, each with a different name.
Right-click on the file with the “_No_magnetometer” extension on select the “Reprocess Orientation” option. Then select “No magnetometer” from the options and select “Reprocess”.
Right-click on the file with the “_Fixed_magnetometer” extension on select the “Reprocess Orientation” option. Then select “Constant weight magnetometer” from the options and select “Reprocess”.
- Switch back to Moveo by clicking on View->”Mobility View” in the top menu bar.
- At this point, we have 3 Moveo trials, each with different backing files that have different magnetometer processing options, but the joint angle estimates have not been re-calculated yet.
- Click on Tools->”Reanalyze Data”. Select the “Force update even if versions match” option, as shown below, and then the “Continue” button.
- You can now select these trials within Moveo to view their reports. Note that the “previous trial” and “next trial” chevrons in the report do not work correctly in the current release, so it is recommended that you return to the list of tests from the calendar view to access and compare the results. This issue has been resolved in a soon-to-be-released update.
- Repeat this process for any additional files as needed
- Make sure to reset your file naming convention to remove the extensions we added. The original format that should be used is “%D_%T”.
If this process doesn't seem to improve your results, please feel free to reach out through a support request with any follow-up questions you may have.
0 Comments