This project will be maintained on the new repository at: https://github.com/yallah-team/YALLAH.

The new Wiki address is: https://github.com/yallah-team/YALLAH/wiki.

This repository will not be updated further.

(June 26th, 2019)

YALLAH stands for Yet Another Low-Level Agent Handler.

The goal of YALLAH is to allow 3D content creators to generate, customize, animate, dress, and deploy a virtual human in a Game Engine in a few hours of work. We aim at providing a common framework and API for the creation of multi-functional virtual humans that can be used in different application domains: video games, embodied conversational agents, virtual assistants, front-end for chat-bot systems, and the like.

YALLAH is composed of:

1. A Blender plugin (yallah) that takes as input virtual characters created with ManuelBastioniLab and quickly (procedurally) prepares them to run in real-time environments.
2. A Unity package (under the folder YALLAH) containing several motion controllers (Gaze, Locomotion, Text-to-Speech, ...) for the configuration of interactive virtual charcaters.

The following people can take advantage of project YALLAH:

• For 3D Authors (Blender/Unity artists) YALLAH is a reference for the development of virtual humans that have to be employed in real-time engines. They can create a custom virtual character and have it ready in a real-time game engine like Unity in less than an hour of work. Only a minimal knowledge of Blender and Unity is required;
• For 3D Software Developers (Unity scripters) YALLAH is an open platform that can be taken as reference to implement new functionalities for improved interactive virtual humans. They can insert a new character in a virtual environment and control it with a high-level API (e.g., LookAt, SpeakText, PlayAnimation, SetFacialExpression, ...);
• For Developers and Researchers YALLAH is a starting point for the development of new motion controllers for Virtual Humans.

• YALLENS Using a YALLAH-generated character with Hololens Youtube Video
• DECAD (DFKI Embodied Conversational Agent Demo) at http://decad.sb.dfki.de demonstrates an online running demo of a virtual human created with YALLAH.

The official YALLAH documentation is maintained on the online github wiki pages.

There are video tutorials on YouTube at the following playlist: https://www.youtube.com/playlist?list=PL92sVNf_0lio6ilwFiAesUJ1AQO1B4ILU

1. Character Creation in Blender https://youtu.be/wB9baKwU9dw
2. Import the Character in Unity https://youtu.be/bWzclpmE3ag
3. Configure the Motion Controllers in Unity https://youtu.be/dD2oNAgDN7k

All downloads here: https://cloud.dfki.de/owncloud/index.php/s/WxjR8X4MH7qGtxn

• YALLAH-SDK-YYMMDD.zip contains:
  • yallah.zip - the Blender add-on needed to author a new virtual character.
  • YALLAH_Unity-YYMMDD.unitypackage - all the scripts and basic assets to run the virtual character in Unity.
  • Wiki.zip - all the documents: a dump of the online Wiki.
• YALLAH-Assets-YYMMDD.zip: a collection of assets to dress and animate your characters.

• The character generated by YALLAH can be used in the Unity game engine. The characters can be integrated as Asset in any Unity project and deployed on most desktop, mobile, and web platforms.
• Features, like eye-blinking, eye gaze, animated Text-to-speech (and others), can be enabled or disabled according to the needs (application domain, performance, ...)
• A set of dedicated scripts allow to build a Unity project for WebGL and use the high-level API (LookAt, SpeakText, ...) directly from the web pages via JavaScript code.
• If a developer wants to implement new motion controllers (e.g., locomotion), a set of guidelines and design patterns allow him/her to inject the new feature into the system without breaking the compatibility with existing ones.

• The 3D authoring pipeline relies on Blender.
• The character generation heavily relies on the excellent MB-Lab character generator, an addon for Blender.
  • Hystorical note! The MB-Lab addon is a continuation of the excellent ManuelBastioniLab, that closed twards the end of 2018. Version 1 of YALLAH still relies on it. The old MB-Lab v1.6.1a, currently used for this project, is available for direct download at https://cloud.dfki.de/owncloud/index.php/s/558mWyCq7BDS58t.
• The real-time 3D rendering targets Unity.
• The text-to-speech synthesis relies on the MARY Text-to-Speech System (MaryTTS)
• The Haxe meta-programming language is used to develop some of the code.
• This project is developed at the Affective Computing group, a research group of the German Research Center for Artificial Intelligence (DFKI),
  • The development of YALLAH started at the Sign Language Synthesis and Interation group, a research group of the German Research Center for Artificial Intelligence (DFKI) and funded by the Cluster of Excellence on Multimodal Computing and Interaction (MMCI) in Saarbrücken, Germany.

• Nunnari, Fabrizio, and Alexis Heloir. “Yet Another Low‐level Agent Handler.” Computer Animation and Virtual Worlds 30, no. 3–4 (June 6, 2019): e1891. https://doi.org/10.1002/cav.1891.

  @article{nunnari_yet_2019,
      author = {Nunnari, Fabrizio and Heloir, Alexis},
      title = {Yet another low-level agent handler},
      journal = {Computer Animation and Virtual Worlds},
      volume = {30},
      number = {3-4},
      pages = {e1891},
      keywords = {agent framework, character creation pipeline, virtual human},
      doi = {10.1002/cav.1891},
      url = {https://onlinelibrary.wiley.com/doi/abs/10.1002/cav.1891},
      abstract = {Abstract YALLAH is a framework for the creation of real-time interactive virtual humans. Its production pipeline supports the continuous, parallel development of both the character and the software, and allows users for the deployment of a new character in a few hours of work. YALLAH is based on freely available software, mostly open-source, and its modular software architecture provides a framework for the seamless integration of new features. Finally, thanks to transpilation, the whole framework is conceived to accommodate multiple game engines.},
      year = {2019}
  }

• Fabrizio Nunnari (Research & Development)
• Alexis Heloir (Research, Testing)
• Daksitha Withanage Don (Development, Testing)

• Kiarash Tamaddon (Development)
• Timo Gühring (Research & Development: text to speech, Haxe ports)