Tuesday, July 5th, 2022 Posted by Jim Thacker

Kinetix releases Kinetix 1.5

Originally posted on 16 March 2022. Scroll down for news of the Kinetix 1.5 update.

Kinetix has rolled out a major update to its browser-based service for converting video footage of an actor into 3D animations that can be used in tools like Blender, Maya, Unity or Unreal Engine.

New features in version 1.4 of the online AI motion-capture tool include support for hand capture, half-body recognition, XR export, and over 40 stock 3D characters and motions from Adobe’s Mixamo library.

In addition, all of Kinetix’s features are now available to use for free.

One of a number of new online services for extracting 3D animation from video
Launched in 2020, and still officially in advanced beta, Kinetix is a browser-based service for extracting the motion of an actor from video footage, and retargeting it to a 3D character.

It is one of a number of online AI-based motion-capture platforms to have launched in the past two years – recent alternatives include PixCap and Plask – but its founders have a pedigree in the industry.

CTO Henri Mirande worked at facial capture firm Dynamixyz, later bought by Take-Two Interactive Software.

Capture body movements and export them for use in DCC applications or game engines
Kinetix users can upload videos in MP4, MOV or AVI format, or just paste in the URL of a YouTube video.

The system extracts the motion of the actor’s body – there’s no facial capture yet – from the footage and retargets it to a 3D character: Kinetix has a library of stock characters ranging from realistic to cartoony.

There aren’t currently many options for editing the animation, but users can trim or retime the motion, assign a background for the character, and dress the scene using a library of stock 3D props and furniture.

The animation can then be rendered as a video in WEBM format, or the animated character can be exported in FBX or glTF format, with 4,096 x 4096px PBR textures following a specular-glossiness workflow.

Exported animation can be used in any DCC application that supports the FBX format – Kinetix has tutorials for Blender and Maya – or in game engines like Unity and Unreal Engine.

Kinetix 1.4 in action: hand movements extracted from the YouTube video of ‘Never Gonna Give You Up’ by Rick Astley retargeted to Mixamo’s Mousey stock character inside the online mocap tool.

New in Kinetix 1.4: hand capture and new stock characters from Adobe’s Mixamo library
Key changes in Kinetix 1.4 include half-body recognition – previously, Kinetix recommended that the actor’s entire body remained visible throughout a shot – and support for hand movements.

As a quick test, we fed in the YouTube video of Never Gonna Give You Up by Rick Astley, immortalised in a billion rickrolls, and processed a section in which only the top half of Rick’s body is visible.

The results, embedded above, are pretty good: as you might expect, there’s some foot slippage, but the hand gestures come through clearly.

The stock character we used is one of over 40 new characters and animations from online animation platform Mixamo available in Kinetix, following a recent partnership between Kinetix and Adobe.

Other changes include XR Export, with users able to generate a URL for viewing a scene in extended reality.

Kinetix has also removed Kinecoins, its internal payment system, making all of Kinetix’s features available for free: previously, paying subscribers could access more tools and process more minutes of video each month.

Updated 5 July 2022: Kinetix has released Kinetix 1.5.

The update makes it possible to import characters in FBX or glTF (.glb) format, and to mirror animation.

Workflow improvements include visualisation of animation transitions in Kinetix’s Tweener panel, and the option to resize and reposition imported videos in the Preview section.

Pricing and system requirements
Kinetix 1.5 is currently available to use for free. It’s a browser-based service, so there are few system requirements, although Kinetix recommends using it on a machine with at least 4GB of RAM.

Read more about Kinetix in the online documentation

Visit the Kinetix website
(Includes a link to register for the beta)