Vicon Shogun banner

About Vicon Shogun 1.7

Shogun 1.7 is the seventh point release of Vicon’s entertainment market software.

In this latest release, Shogun Live introduces the ability to capture uncompressed MOV files, Z-up labeling and solving skeleton templates, and Y-up origin support. These enhancements are in addition to multiple improvements made to subject calibration and Vicon Solving Setup (VSS) files.

Shogun Post has also received improvements to its offline subject calibration, with Auto-Skeleton now working with the new Z-up solving templates, and support has been added for VST 3.6.

These improvements enhance the performance and flexibility of tools introduced in Shogun 1.6, in response to direct customer feedback. Non Virtual Production users will also enjoy higher levels of precision in their subject calibration, with greater flexibility for non-standardized marker placements.

For a more detailed description, see:

This release also benefits from ongoing maintenance, with a number of issues having been addressed (see Addressed issues in Vicon Shogun 1.7).

For information on requirements for installing and running Shogun, see PC requirements in Installing and licensing Vicon Shogun.

For detailed information on PC requirements, visit the Vicon website FAQs and select Operating systems and PC or contact Vicon Support.



New features in Vicon Shogun 1.7

For descriptions of the new features in this release of Shogun, see:

Shogun Live 1.7 new features

These are the main new features in this release of Shogun Live:

Capture to uncompressed MOV

You can now save uncompressed video directly to a .mov file so that you don't have to capture to Vicon Video file format (.vvid) and then convert to .mov in Vicon Video Converter.

To capture to uncompressed MOV format:

  1. Ensure your Vicon system, including your video camera(s), is connected to Shogun Live.
  2. On the System tab, select the video camera, and in its Properties section below, go to the Capture section and:
    1. In the Container Format field, select MOV.
    2. In the Video Payload field, select Uncompressed .


  3. Ensure the other settings are as required and then capture the required movement.
    A MOV file is created in the specified location (either in the default Capture Folder, or in the custom Device Capture Folders).

For more information on setting up video cameras, see Prepare video cameras in Getting started with Vicon Shogun.

Y-up origin control

For productions and projects working in Y-up coordinate systems, an advanced control in the Camera Calibration panel puts Shogun into Y-up mode.

To set the Shogun Live scene to Y-up mode:

  • In the Camera Calibration pane, ensure the Advanced properties are displayed and in the Set Volume Origin section, select Legacy Y-Up Scene Orientation.

     

Z-up skeletons

The labeling and solving skeleton templates that are supplied with Shogun 1.7 are now Z-up, instead of Y-up. This means that the subject models are now defined in the same axis as Shogun (Z-up).

Note

VSS live calibration and traditional offline calibration in Shogun 1.7 is only compatible with 1.7 templates. VSS calibration only works with subjects that use solving skeleton templates (VSS) version 3.6 and later. These are the templates that are installed with Shogun 1.7. If you try to calibrate with a subject using an earlier version of the templates, an error message is displayed as a reminder. Labeling calibration continues to work with older versions of the labeling skeletons.

Auto-Skeleton is compatible with both Shogun 1.6 and 1.7 VSS templates.

You can continue to use calibrated VSS files from Shogun 1.6 or earlier for solving in Shogun Post.

If you have scripted the importing of templates and aren't yet ready to use the new Z-up versions of the templates, modify your script to point to the older, Y-up versions. They can be found in:
C:\Program Files\Vicon\ShogunLive1.6\Configuration\ModelTemplates

In addition, the foot bones have been simplified, so that the leftForeFoot and rightForeFoot bones have been removed, leaving only Foot and ToeBase.

Shogun Post 1.6 template, showing the ForeFoot bone:

 

Shogun Post 1.7 template, showing the simplified foot bones:


Improvements have also been made to the neck and spine, resulting in skeletons that more accurately fit the data and provide a smoother solve.

These enhancements also contribute to a better overall calibration.

Shogun Post 1.7 new features

There are no new features for Shogun Post with this release.



Improvements in Vicon Shogun 1.7

Shogun Live 1.7 improvements

These are the main improvements in this release of Shogun Live:

Improvements to subject calibration

Shogun 1.7 introduces improved Live subject calibration, enabling you to calibrate subjects in Shogun Live in real time with greater speed and accuracy.

The standard real-time subject calibration process (see Create and calibrate a subject in Getting started with Vicon Shogun) in Shogun 1.7 offers the following improvements:

  • Support for the new skeleton templates (VST 3.6): calibration now handles more natural spine and neck animation for the solving skeletons.
    • The skeleton model has been reparameterized (waist, shoulders, spine, neck, feet).
    • Kinks in the poses have been greatly reduced, which overall gives smoother solves than in previous releases of Shogun.

For more information, see VST 3.6 and the new skeletons.

  • Lower latency during calibration: more frequent refinements during the calibration process for both the labeling skeleton and the solving skeleton, synchronous booting of the labeling and the solving skeletons.
    • Priors have less importance, which gives a better fit.

    • The Range of Motion can be short, but must be comprehensive.

    • System is less tolerant of ranges of motion that are incomplete or lack the full range of movement required.

    • Improved finger calibration.

  • Skeleton booting from A-pose has been fully revised.
    • Increased robustness with non-standard marker placements (for example, for larger subjects, or thigh marker placements near the knees).
      • A number of booting issues have been fixed.
    • Higher precision and recall when classifying A-poses as valid or invalid.
      • Prevents bad poses from initializing the subject calibrator.

      • Requires the subject to stand in a valid A-pose (for example, T-poses are no longer valid).

      • The A-pose booting threshold gives flexibility if needed.

    • More effective booting of finger skeletons

This table gives further details of differences between Live subject calibration in Shogun 1.7 and previous versions.

ChangeLive subject calibration in earlier versionsLive subject calibration in 1.7
Latency between booting labeling skeleton and solving skeletonA few secondsImmediate
Booting qualityMarker re-calibration damped by target priors. This could lead to tracking failures.
  • Enhanced marker re-calibration.
  • Labeling and solving skeletons initially have identical scales.
  • Hands are initially approximately scaled. Functional calibration produces more accurate scaling later.
Solving skeleton calibration
  • Most calibration parameters inherited from labeling skeleton.
  • Optimization more stable, but less dependent on the data.
  • Fully optimized skeleton.
  • Optimization more dependent on the data (and therefore fits better). When the results are unstable, particularly at the start of a ROM, more data is needed for a good calibration estimate.
Labeling clustersAttachment occurred sooner but was less reliable.
  • Attachment occurs after 10 calibration frames.
  • The same rule applies to marker-only recalibration: the subject must not remain static.

VST 3.6 and the new skeletons

The new skeletons:

  • Are tagged with 3.6 as the VST version number.

    VST 3.6 skeletons support new A-pose constraints that apply to limbs such as the neck and the spine.
  • Are supported by the Auto-Skeleton feature in Shogun Post.
  • Have new symmetric population statistics, for both targets and joints.

    Tip
    Correct population statistics are key to labeling quality. Manually changing the population statistics may result in labeling failures as well as booting failures across live subject calibration and the normal processing graph.

Custom skeletons:

  • Should ideally be derived from one of the supplied Shogun templates.
    • We recommend that you keep the statistics intact for human skeletons. For each joint and for each target, import the following fields from new templates:
      • MEAN
      • COVARIANCE
      • RANGE-CENTRING-STATE
      • RANGE
      • TPOSE-MEAN
      • TPOSE-COVARIANCE
    • Statistics generally remain appropriate if there is no significant change in joint pre-rotations.

Legacy skeletons:

  • Are supported if:
    • VST version number is set to 3.6 (see above)
    • Statistics are updated using one of the new templates
  • May struggle to boot in Shogun Live if statistics are not updated.

If you need more help with A-pose booting and calibration, see Troubleshoot subject calibration in Getting started with Vicon Shogun.

Upgrade a 3.5 subject to 3.6 

To upgrade a subject that uses a custom skeleton saved in versions of Shogun earlier than 1.7, do one of the following:

  • Use the Export preferences in Shogun Post:
    1. Import the VST into Shogun Post, and in the Preferences dialog box, click the Export tab and in the File Format field, select the required option (VST or VSK).
    2. In the Format options area, change the Version to 3.6 [Shogun 1.7] and close the dialog box.

    3. Export the file.

or

  • Use a text editor:
    • In a text editor, open the VST file and change the VERSION to 3.6.

Shogun Post 1.7 improvements

This is the main improvement in this release of Shogun Post:

Auto-Skeleton

Auto-skeleton has been updated to work with the new, Z-up solving templates (see Labeling and solving template improvements). All VST and VSS files installed to C:\Program Files\Vicon\ShogunPost1.7\ModelTemplates are now Z-up skeletons.

Auto-skeleton now imports the solving template with the simplified foot bones, without the leftForeFoot and rightForeFoot bones, leaving only Foot and ToeBase. Auto-Skeleton calibrates a 1.7 template with the new foot setup.

In addition, the solving skeleton templates changes for 1.7 remove the need for the more complex constraint setup that was previously necessary. In the 1.7 the waist markers constrain only the waist, the neck and back markers constrain only Spine3, and the head markers only constrain the Head bone. The spine and neck bones are unconstrained.

An additional Auto-Skeleton option, Save VSS after calibration, which is selected by default, enables you to save the solving skeleton.


New tutorial videos

Since the previous release, the following educational materials have been produced:


How do I optimize my live solve using Auto-Skeleton?



How do I get a good A-pose to boot from?



How do I capture uncompressed mov footage?



How do I set up a Vicon-driven cine camera in Unreal Engine with lens distortion?