How do I get the highest quality MVN data?

The data quality of the MVN systems can be optimized by taking some important factors into account: posture, environment and body dimensions. If these factors are accounted for during calibration, like making sure all segments are in a correct orientation during calibration, you can expect a very accurate system without offsets or drift in orientation. After the Npose (or Tpose) + Walk calibration you can even go into ANY environment with MVN Animate/Analyze 2018 due to its full magnetic immunity (new feature! Read more about it in this post).

First, make sure that you know how to place the hardware. Please refer to the following videos to do so:
Placing MVN straps
Preparing hardware MVN Awinda
Preparing hardware MVN Link

Also enter the body dimensions of your subject. These are used to determine in-body distances and they eventually aid in global positioning. Please refer to the following video on how to accurately measure body dimensions. 
Measuring body dimensions

Quality calibration posture:

The Npose + Walk is the recommended calibration procedure, since the Npose relates closely to a natural standing position. Next to watching the tutorial video about calibration (for MVN Link or MVN Awinda) the next few statements will help you improve your data quality even further:

  • The joint center of the ankles should be located straight below the respective hip joints. As a rule of thumb you can use the method as shown in the picture below to get the correct orientation and approximate distance in between the feet:

  • Extend the knees straight, while making sure not to overstretch/overextend.
  • Stand up straight like you normally do. Do not overextend your spine (pushing your chest forward). After calibration if you see the avatar bending a bit forward while you stand normally, you were standing a bit overextended during calibration. 
  • The thumb is pointing forward and the center of the wrist joint straight underneath the center of the shoulder joint. Make sure that the arm is straight in all axes: prevent overstretching/overextending and sideways (lateral) bending in the elbow. The anatomy of the elbow in some people can cause a natural offset in a sideways bend. Try to minimize this effect as much as possible. 
    As a rule of thumb for calibrating an 'average person' you can touch the hips slightly with the wrist. Commonly, this aligns the arms with a straight and vertical line from the shoulders. This will also help with keeping the arms still for a moment while calibrating. 
  • Have the palm of the hand in a straight line with the forearm. 
  • Stand still during the Npose calibration. This takes approximately 5 seconds. 
  • Next, the MVN software will prompt you to start walking. During walking, walk normally and casually swing the arms as you walk. Walk 7-8 meters, turn around and walk back. Ideally, continue walking until the calibration capture finishes. 
  • After the processing of the calibration is done stand in Npose when clicking on 'Apply'.

Quality data throughout the recording:

The MVN system will perform very well after a high quality calibration. To make this high data quality last throughout all sessions there are a few steps you can do. 

  • Warm up the data filters. The systems work with data filters. They use data from the (near) past as a reference to correct and predict recent and future data. Immediately after calibration the filters start building their data memory. Feeding the trackers for the first 10-20 seconds after calibration with low-dynamic 2D/3D data will secure a high quality and stable performance of your MVN system.

Quality data after the recording:

The 'No level' and 'Single level' scenario provide the possibility to run the HD reprocessor (MVN 2018 and up). As opposed to the real-time engine or (automatic) reprocessor of MVN, where the data is processed frame by frame, progressing forward in time, the HD reprocessor adds the feature to process data over a larger time window. In this way, the HD reprocess engine is able to include information from the past, present, and future to get an optimal (and more consistent) estimate of the position and orientation of each segment.

This will raise the data quality by extracting more information from the same data set and modelling for skin artifacts, with the purpose to ensure the best results within the chosen type of environment or movements. Click this button (see picture), activate the check for the HD Reprocessing in the drop down menu and run the HD reprocessor! -->





Was this article helpful?
4 out of 4 found this helpful
Do you have a question? Please post your question in our Community Forum