The lack of full body posture synchronization has always been a sore point which is in urgent need to be solved for the VR applications. Recently, REALIS, a company from China, announced that they would release two multi-person full-body motion capture software products: Tenoomi and Rovonga.
RTS-Tenoomi
RTS-Tenoomi is a rigid body based full-body motion capture software product, suitable for real-time interaction. Rigid bodies, bound on performer’s head, hands, feet and waist, will be captured by RTS-Tracker and then transferred to RTS-Tenoomi. The latter combines the original data of the rigid bodies (position and orientation) with IK technology and HSM model to calculate the full-body skeletons.
According to information, RTS-Tenoomi has the following characteristics:
High performance on data solving
With Onekey obtaining, user can transfer the data collected by RTS-Tracker to RTS-Tenoomi. It only takes 6 rigid bodies on each actor to solve a full-body skeleton. As long as the skeleton model remains the same, no geometrical dimension re-calculating is needed when user alters the human object, the model or the both of them. The retargeting function is useful in practice as it generates adjustments according to the morphological differences caused when altering human objects, to adapt the virtual model with different proportions and skeletons. In other words, no matter how diverse the human object’s physique will be, the solver is able to calculate smooth and natural skeletons.
Pic1: RTS-Tenoomi Real-time 3D-engine Mocap software
Real-time multi-person motion capture
Traditionally, motion capture is about tracking one single actor. Real-time multi-person motion capture is relatively more difficult since the number of actor increases, the total amount of rigid bodies increases accordingly.
Several optimizations of data processing have been in Tenoomi. First of all, the number of rigid bodies on each human object has been reduced to 6, which cuts down the volume of original data. Secondly, in order to guarantee the calculation of full-body skeleton, particular algorithms patented by Realis are integrated in Tenoomi to optimize the problems of occlusions, as well as to make sure of every object’s physical attitude being recorded simultaneously, even with less data than traditional tracking method.
Seamless docking with Unreal Engine 4 and Unity
Tenoomi will be released together with specific UE4/Unity SDK. No matter which engine or version the developers are using, Tenoomi provides seamless docking with them. In order to achieve full-body motion capture in VR, developers no longer have to edit their own IK blueprint or C++ coding. The SDK is provided freely to developers, simply open the software and multiple virtual characters can simulate full body interaction just like in the real world.
RTS-Rovonga
Rovonga is a marker based (from head to foot) motion capture software product. It can achieve high accuracy, neatly collected data and low loss rate. This filmmaking leveled product will be mainly used for high fluency animation production, scientific research, ergonomic analysis, sports and many other sectors. It will make its debut this August in SIGGRAPH 2018 in Vancouver, Canada.
Jenn Xu, the CEO of Realis, says that Realis has being working hard on optical motion capture and human-computer interaction. Currently, the three motion capture systems focus on different specific markets and they will be helpful in next generation of industrial application.
Media Contact
Company Name: Shenzhen Realis Multimedia Technology Co.,Ltd
Contact Person: Nancy HUANG
Email: Send Email
Country: China
Website: www.realis-e.com