UE5.4

Synavis: UE5.4 release

This release marks the version we used in the course "Virtual Worlds for Machine Learning" as well as our recent paper to the NeIC2024, available as preprint: Adapting Agricultural Virtual Environments in Game Engines to Improve HPC Accessibility, Baker et al. (2024), https://doi.org/10.34734/FZJ-2024-03386

Video on parallel execution of Unreal Engine for AI Image generation:

Video on texture mapping to map simulation data or labels onto materials:

SynavisUE was moved into the main repository, be aware that this causes some larger changes to the workflow

In the coming weeks, we will gradually make all workflows more concise, which in part involves the collection of all code pieces into one. This will also influence the example repositories. This release will still retain the same structure, while the next release will be restructured!

Other changes are:

1. More direct integration of texture and material handling into the workflow.
2. SynavisUE is now moving to UE5.4
3. Extended functionality for movement and encapsulating of behavior within UE
4. Synavis build workflow is now slightly lighter

Executables and installation guide will follow

v0.1.5

What's Changed

Current working version including implementation as preparation for upcoming features, and current publication in review.

• Will search for LibVPX for upcoming features (you can ignore this and it is off by default)
• Signalling Server will automatically seek out certain information on unix systems
• Improved compatability with different JSON messages in the data channel
• Connection classes now have an "initialize" method so that IP constraints can be relayed
• Data Connector has retry methods as well as improved robustness against missing packages. This is mainly due to the fact that packages might not get through when the data channel was used the other direction a lot.
• libdatachannel messages can be activated through Synavis for debugging purposes

Full Changelog: v0.1.0...v0.1.5

v0.1.0

Initial release for Synavis as a central framework. Releases will usually also be accompanied by fitting versions in SynavisUE and the example project.

This release of Synavis includes, in overview, the components:

• A seperate signalling server specifically for the communication wit the python interface
• A framework to couple a python backend to UE via DataConnectors or MediaReceivers
• Handling of the sending and reception of messages, specifically the sending of buffers and geometry
• A framework for the establishment of connections via a bridge

Note that specific use of the UE PixelStreaming Plugin and handling of its frame packages are not included in this framework. For information on those issues, please see the SynavisUE plugin: it contains the JSON commands and their usage.

If there are questions, please use the Issue Tracker.