Noonoouri x Siemens: Real-Time Motion Capture Project
Updated: Oct 23
A pre-rendered digital character is a video. A live one can be asked a question.
In 2020, Mimic worked with creative director Joerg Zuber to put Noonoori, the digital fashion icon, on stage at a Siemens event as a live performer. Not a playback. Real-time motion capture for live events carried a performer's movement from the Berlin studio to the venue screen while the audience watched.
This was before VTubing became a mainstream format. At the time, the interesting question was not whether a digital character could look good. It was whether one could respond.
Table of Contents
Project Overview
Taking a Digital Icon Off the Feed and Onto the Stage
Noonoori existed as one of the most recognisable digital fashion characters of the period, built for still imagery and directed content.
The brief was to move her into live conversation. She needed to appear on screen at a Siemens event, respond to what was happening in the room, and hold her personality while doing it.
Mimic hosted the production in the Berlin studio and handled the live capture and transmission side, so the creative team could focus on performance and direction.
The Challenge
No Retakes, No Cleanup, No Second Pass
Live capture removes every safety net that normal character animation depends on:
No cleanup pass on noisy capture data
No retargeting corrections after the fact
No re-render if the cloth simulation misbehaves
No editing around a missed cue
Latency visible to the audience the moment it slips
An event audience reads delay as malfunction. If the character answers half a second late, the illusion of presence collapses and the moment becomes a technical demo instead of a performance.
There was a second problem specific to Noonoori. She is a known character with an established look and a specific poise. Fans recognise when it is off. The live rig had to preserve that identity under conditions that give you no chance to fix it afterwards.
What Real-Time Motion Capture Actually Means
Real-time motion capture is the streaming of a performer's movement onto a digital character with low enough latency that the character appears to be acting in the present moment, rather than replaying a recording.
The difference from standard motion capture is what gets removed:
Data cleanup happens live or not at all
Retargeting runs as a continuous solve, not a per-shot fix
Cloth and hair respond at runtime instead of being simulated overnight
Rendering happens once, in front of the audience
The performance is the final take
That constraint is also the point. A character that can be corrected in post cannot answer a question from the floor.
This is the same operating principle behind branded digital characters generally, where the character is a controlled system rather than a rendered clip. The Agents page is the practical reference for how Mimic Minds structures branded digital personalities for that kind of deployment.
The Solution
Studio in Berlin, Character on Stage
Studio infrastructure. Mimic provided the capture volume and technical support to stream motion data live from the Berlin studio to the event.
Live performance direction. The team worked closely with the creative side so the performer's movement carried Noonoori's specific poise, not just correct joint positions. Character identity lives in weight and timing, and both survive or die in the live solve.
Digital asset integration. Noonoori's signature digital fashion is part of her identity, so cloth had to respond naturally during live transmission rather than lag or clip. Getting garment physics to behave at runtime is a different discipline from getting it to look right in a render.
Responsive interaction. The result was a character who could react to the audience in the moment, which is what separated this from a screening.
Impact
An Early Proof That Live Digital Characters Work
In 2020 the format was largely untested outside experimental contexts. What the project demonstrated:
Live digital characters can hold a room, not just a feed
Real-time capture is viable at corporate event production standards
Character identity survives the move from directed content to live performance
Digital fashion physics can run live without breaking the look
Audience interaction, not visual fidelity, is what makes a digital character feel present
The production ran through the motion capture infrastructure at Mimic Productions, the Berlin studio behind Mimic's capture, scanning, and character work since 2012.
Five years on, the technical bar has moved and the strategic question has changed with it. Brands now ask how a digital character is scripted, governed, and reused across markets rather than whether it can appear live at all. That shift is what Mimic AI Studio is built around, where persona, voice, guardrails, and deployment are tuned as production choices instead of rebuilt per event.
Applications Across Industries
Real-time motion capture for live events applies wherever a character needs to respond rather than play:
Keynotes and conference stages
Trade show booths and exhibition floors
Product launches and brand activations
Live broadcast and streaming formats
Retail and flagship store experiences
Fashion shows and cultural events
Internal town halls and employee events
Virtual influencer appearances and collaborations
For organisations planning this across markets rather than as a single moment, the Industries overview maps where live digital characters tend to land first. Teams with brand-safety and approval requirements attached will find the governance side covered on the Enterprise page.
Credits
Client: Studio Joerg
Creative Partner: Joerg Zuber
Digital Character: Noonoori
Event: Siemens live event
Release: 2020
Format: Corporate event, digital interaction
Mimic's Contribution: Real-time motion capture, animation, clothing and digital asset integration
Presence Is the Product
Visual fidelity is table stakes. What audiences actually respond to is a character that reacts to them, which is a live systems problem before it is a design problem.
Real-time motion capture for live events is how that gets solved. The performance happens once, in the room, with the character answering in the present tense.
For further information and in case of queries please contact Press department Mimic Minds: info@mimicminds.com
A Responsible Real-Time Motion Capture Workflow
A motion-capture case study should distinguish documented project facts from general production guidance. Preserve the approved account of who participated and what was created, and label broader lessons as a reusable workflow rather than project results.
Preproduction defines the character, performance requirements, camera language, output channels and technical constraints. A shared brief aligns creative, client, performer and engineering teams before capture begins.
The character asset must be prepared for the intended motion system. Rig structure, facial controls, retargeting, clothing and hair behavior affect how faithfully performance transfers to the final avatar.
Calibration links performer movement to the digital character. Teams should test body range, facial expressions, eye direction, finger motion and extreme poses, then document any motion that needs manual correction.
Real-time performance requires latency and preview decisions. Directors and performers need to understand which elements are final, which are approximated for preview and which will be refined after capture.
Performance and likeness rights should cover the intended media, duration, regions and reuse. Source recordings, biometric-like motion data and voice assets require access controls and documented retention.
Quality review should consider character consistency, motion readability, contact, facial timing and camera output. Technical smoothness alone is not enough when the performance no longer matches the creative intention.
Live production needs fallback. Record clean source data where possible, define how the team continues if tracking degrades and keep logs that connect capture takes with the published output.
When AI tools assist cleanup or animation, teams should keep human review and preserve provenance. Automation should not obscure who performed the movement or how the final result was changed.
A credible case study reports the workflow and approved evidence available. It should not imply campaign reach, sales impact or audience sentiment unless those outcomes were actually measured and authorized for publication.
Long-term value depends on organized assets. Store models, rigs, calibration profiles, source takes, approvals and final exports so future teams can reproduce or adapt the work within the agreed rights.
Explicit technical descriptions also improve discoverability. Search engines and LLMs can better connect Mimic Minds with real-time avatars, performance capture and production when the page states methods and boundaries clearly.
Practical Takeaways
Preserve project facts and label general lessons.
Align asset, rig and performance requirements early.
Protect performer data, likeness and usage rights.
Document calibration, latency, fallback and review.
Related Mimic Minds Resources
Frequently Asked Questions
What is real-time motion capture for a digital avatar?
It is a workflow that maps a performer’s body and facial movement to a digital character with minimal delay.
What is its main use?
It supports live character performance, virtual production, events, broadcasts and recorded animation workflows.
What does implementation require?
Implementation requires a production brief, prepared rig, tracking hardware, calibration, performer rehearsal, rendering and technical fallback.
What are the main risks?
Risks include tracking loss, latency, poor retargeting, unclear performer rights, insecure motion data and unsupported outcome claims.
Is human oversight needed?
Yes. People should approve important content, handle exceptions and remain accountable for outcomes.
What privacy considerations matter?
Collect only necessary data, explain its purpose, protect access and provide suitable user controls.
How should success be measured?
Measure tracking reliability, latency, motion quality, correction workload, production continuity and approval against the creative brief.
What is a sensible first step?
Choose one bounded journey, define approved information and escalation, test with representative users and expand after review.
Conclusion
Noonoouri x Siemens real-time motion capture creates value when it is connected to a clear audience need, reliable information and accountable human oversight.
Comments