Thursday, June 04, 2026

Thoughts on Generative A.I., Gutenberg, Cameras and What Comes Next

This is a concept trailer for A.I. filmmaker Rodrigo Soares' Remains: The Only Survivors. He describes it as film about "an empty world destroyed by machines...a dystopian cyberpunk future where only machines remain, and young hybrid/android girl survives by repairing broken machines. Whatever your feels about generative A.I., it's a very impressive piece of work.

And it raises the question, what happens when A.I. films become just as impressive - or even more impressive - than traditionally made ones? That's a question that Rodrigo himself seems to be wrestling with:

"Do you think a creator who uses A.I. can also be considered a real artist? 

For me, the real challenge in A.I. filmmaking is not generating the most impressive frame. It is shaping those frames in to something meaningful. A.I. can open the door to incredible worlds, but the filmmaker still needs to decide what the audience should feel, where they should look, and why each image matters."

-   Rodrigo Soares, Fableforge.ai
When Johannes Gutenberg introduced movable type and the printing press to Europe in the 1400s, the immediate reaction from academics and professional scribes was a mix of awe and absolute horror, not unlike the reaction to Generative A.I. today. Before Gutenberg, books were copied by hand. It was a sacred, slow, highly artistic process. Critics of the printing press argued that mass production would ruin the human mind and that the flood of new books flooded the world with "foolish, ignorant, malignant... and impious" texts.

They were complaining about, essentially, 15th-century print slop.

This happened again in the 19th century with the invention of the camera. Painters were horrified by cameras. The critic Charles Baudelaire famously called photography "art's most deadly enemy," fearing that a machine requiring "only a chemical process and a button push" would destroy the livelihood of true artists. But the camera didn't kill painting, just as the printing press didn't kill scholarship. Neither form was destroyed, but the goalposts shifted, from preservation to critical thinking and original synthesis.

In fine art, painters no longer had to act as human photocopiers documenting reality, they were forced to ask themselves what can a brush do that a lens cannot? The resulting existential crisis gave birth to Impressionism, Cubism, and virtually all modern art.

This is, of course, well worn ground in the ongoing debate over generative A.I. Likewise, The observation that A.I. doesn't automatically make you a great filmmaker — any more than a printing press automatically made you a great author — isn't a particularly new one. However, Rodrigo is suggesting that the goalposts in filmmaking are rapidly shifting from execution to curation.

When generative A.I. models produce work this impressive, we have to directly engage with the question: why do we need traditional forms like puppetry?

Rodrigo's argument illuminates something important: notwithstanding the complex legal and ethical issues raised by generative A.I., it is a technology that ultimately removes barriers and rewards people who actually know how to perform and direct. When he says 'the filmmaker still needs to decide what the audience should feel, where they should look, and why each image matters' - that's a description of what puppeteers and other artists already do. Generative A.I. creates the possibility of doing it with fewer restrictions.

Personally, I think that just like filmmaking, generative A.I. is going to ultimately push puppetry - both traditional and digital - to exciting new places. The transition won't be smooth, and will likely be more than a little terrifying...but it could also be terribly exciting.

Thursday, May 28, 2026

How OK Go Used Blender As a Puppet Stage

One of my favourite music videos of the past year is OK Go's Impulse Purchase, but it's not just an animated music video. It's a small, but important demonstration of how Blender can be used as a live puppet stage.

OK Go frontman Damian Kulash described the origin of his collaboration on this with directors Lucas Zanotto and Will Anderson at Blender Studio to Live Music Blog:

“It began with my love of Lucas Zanotto’s short animated loops — they’re so inventive, so full of joy, always delivering little doses of the kind of wonder we’re always searching for in our own videos. So I thought the universe flowing from his brain might be the perfect setting for a lyric video, but when I reached out to him, he had a more ambitious project to propose: a ‘live performance’ with me AS one of his characters. Suddenly it was a much weirder, more wonderful project than what I’d envisioned,”  

To make the video, Damian performed the character via a live facial motion-capture app on an iPhone. The app used OSC (the Open Sound Control protocol) to send his real-time eye, lip, and head movements into Blender, where they drove the character’s erratic facial expressions on the fly.

What makes this true digital puppetry, not just digital animation, is that Damian isn't simply supplying reference material for an animator to interpret; he is performing the character through a real-time control system. His face is the input device. Blender is the puppet stage.

Best of all, in keeping with Blender Studio's open source filmmaking ethos, a demo version of this set-up has been released for anyone to try:

To run the free demo you will need to download the following:

The full .blend production file used to make Impulse Purchase is also available for Blender Studio subscribers. Complete step-by-step instructions can be found on the Blender Studio Blog

Tuesday, May 26, 2026

Gadzooks Studio Shows Why Performance Beats Prompting

Gadzooks Studio, a Manchester-based stop motion animation studio founded by animation directors Steve Boot and Haydn Secker, has been sharing some really delightful digital puppetry experiments on Instagram. They’re building a Blender-based pipeline that utilizes TouchDesigner, which lets artists and technologists build interactive animation systems without writing traditional code.

What they've shared so far is pretty impressive:

What Gadzooks is doing here is using TouchDesigner to feed their live performance in to Blender, which acts as the character/animation environment. This allows them to create digital puppets that respond to a performer in real time: facial expressions, head turns, timing, reactions, and little bits of character business.

TouchDesigner, created by Toronto-based Derivative, is a big part of Gadzooks’ not-so-secret sauce (although, to their credit, they’ve been sharing much of their workflow on Instagram). TouchDesigner is able to take a variety of live inputs - everything from camera footage to audio, MIDI, sensors, controllers, or just about anything else that outputs a signal - and use that input to control a character in real time without traditional hand-keyed animation. It’s a more sophisticated version of what I was trying to build with Panda Puppet many years ago - a way to make it possible to use almost anything to perform a digital character.

It’s a really great example of the advantage digital performance has over AI-based prompting: the immediacy of puppetry, with the flexibility and control of a digital animation pipeline. The best of both worlds.

I'm excited to see what others might be able to do with TouchDesigner, as well as what else Gadzooks is working on. To keep up-to-date with them, follow Gadzooks.Studio on Instagram!

Friday, May 22, 2026

Inside the Digital Puppetry of Jim Henson’s Puppet Up!

The Henson Digital Puppetry Studio is a digital puppetry platform the Jim Henson Company has been building for decades, with roots that go back to Waldo C. Graphic in the late 1980s. The technology has come a long way since then, and in recent years has become good enough that the Jim Henson Company regularly uses it in Puppet Up!, their touring live improvised puppet comedy show.

In this video from Tested, Henson puppeteers Sarah Oh and Dan Garza show host Adam Savage how they use their puppeteering and improv skills to blend a digital character seamlessly - and hilariously - with a group of practical puppets in a single performance before a live audience. 

What makes this so impressive, of course, isn’t just the technology. It’s that Oh and Garza are skilled puppeteers doing what skilled puppeteers have always done: listening, reacting, finding the joke, and creating the illusion of life where it doesn't actually exist at all.

The octopus may be digital, but the performance is remarkable because it's unmistakably human.

If you want to see the Henson Digital Puppetry Studio in action yourself, Puppet Up! will be playing the Montalbán Theatre in Los Angeles this summer. Visit PuppetUp.com for upcoming tour dates and full details.

Friday, November 14, 2025

The Magic Behind ILM's Real-Time Rocket Is the Performer

ILM's Christopher Jones (left) and creative director Michael Koperwas with Real-Time Rocket in the background (source: Industrial Light and Magic).

Industrial Light and Magic made a splash last year with Real-Time Rocket, an immersive digital puppetry experience for fans of Marvel Studios that was featured at Disney's D23 Expo.

Marvel fans at D23 were able to directly interact with Rocket Raccoon from Marvel's Guardians of the Galaxy, thanks to a collaboration between Marvel Studios, Disney’s StudioLab, and ILM, that enabled improv actors to perform Rocket using markerless motion capture technology.

ILM showcased an updated version of the activation at the Disney Accelerator Demo Day earlier this month:

Real-Time Rocket won a VES Award for Outstanding Visual Effects in a Special Venue Project, which is impressive, although having live audiences interact with a digital character is far from a new idea. Real-time digital characters have been featured in attractions like Turtle Talk With Crush at the Disney theme parks for over twenty years.

The specific challenge for ILM here was that due to the inherent limitations of markerless motion capture, they had to engineer a way to take a relatively limited amount of performance data from a performer and translate that into natural, biologically accurate poses appropriate for Rocket's raccoon anatomy.

@ign Rocket doesn’t know what IGN is, but you can interact with Rocket at the TVA at D23! #disney #d23 #marvel #marvelstudios #guardians #rocketraccoon #gotg #guardiansofthegalaxy #tva #loki #ign #mcu #movie ♬ original sound - IGN Entertainment

Amidst all the hype about employing Unreal Engine, markerless motion capture and the use of a LIMINAL 3D display, it's easy to forget that while the technology behind Real-Time Rocket is remarkable, what audiences responded to so strongly has very little to do with a technology pipeline, and everything to do with the performers using it.

Real-Time Rocket really works because of the classic Disney character performer training behind the scenes. Actors were trained extensively on Marvel lore, Rocket's speech patterns, and improvising with a live audience. The performers had a two-way video feed, which enabled them to spot specific people in the crowd to react to. That is just the latest iteration of what puppeteers have been doing digitally in various forms going all the way back to Jim Henson and Waldo C. Graphic in the 1980s.

So the actual magic at work here isn't digital - it's human.   

Sergei Obraztsov famously said that his hand was the soul of his puppet. Here, the performer provides the character's soul. ILM's digital pipeline is, ultimately, just the puppet. 

Monday, April 08, 2024

Experimental Digital Shadow Puppets in Blender

Lyle at LJ Puppetry recently shared this digital puppetry experiment. Lyle was inspired by the Henson Digital Puppetry Studio and its use of a Waldo, but wanted to see if he could create digital shadow puppets that were controlled in real-time via motion capture inside Blender. Lyle himself admits that his results were mixed, but I think his project may indirectly suggest some really interesting technological possibilities.

Of course, experimenting with digital shadow puppets is a well-trodden path. I began my own journey in to real-time animation experimenting with digital shadow puppets when I first started writing Machin-X, although back then we had to rely on colour blob tracking and custom scripting in Adobe Director because there were no off the shelf tools. 

Today, it's much, much easier to do this with off-shelf open source software like Blender, which is what Lyle used to perform this multi-scene shadow story - a magi, a tiger that transforms into a woman, a figure that coalesces from stars - entirely through hand and finger movements captured by a webcam:

The shadow characters in the video are flat black figures built using Blender's grease pencil tool. Lyle controlled them in real time by mapping specific finger positions to specific puppet movements — thumb open stands the magi up, a closed fist brings him to his knees. What we're seeing in the video isn't playback; it's a live performance into Blender, with the puppeteer off-camera and the output on screen.

Although his experiment wasn't entirely successful, Lyle's documentation is an excellent starting point for someone else who wants to try experimenting with this on their own. You can read an in-depth write-up of his workflow here

This might be the right approach with the wrong tool

I've been following motion capture developments in Blender for a long time. I was very excited by the BlendArMocap plugin when I first heard about it, but I've since discovered that it has some real limitations. Dropped frames (the "jerky mocap" that Lyle mentions) and the ability to only perform one puppet at a time are two of the known issues.

Personally, I suspect that BlendArMocap might be the wrong implementation of the right technology. It's notoriously difficult to manage motion tracking inside Blender itself (which is what BlendArMocap does), but the underlying technology that makes that markerless motion tracking on a standard webcam possible, Google's open source Mediapipe framework, is brilliant.

TouchDesigner's node-based interface can look intimidating, but it's a popular tool used by artists all over the world.

Mediapipe offers real possibilities for digital puppetry. The ability to use markerless motion capture - especially in a tool like Blender's innovative Grease Pencil - is tantalizing, but a better way of utilizing it might be with a program like TouchDesigner. TouchDesigner is used by artists and performers to process, and react to live data — from motion capture to audio to video — in real time. If you've seen a live concert where the visuals pulse and shift in sync with the music, that was likely done with TouchDesigner, and it can also send clean motion data directly into Blender in real time, via OSC. 

Frankly, it's much easier to fix problems like "jerky mocap" in TouchDesigner before the data ever reaches Blender, which in turn creates a much more streamlined workflow.

We need more experiments like this 

I love what Lyle has shared here, and that he is encouraging others to learn from his project. He's given me some things to think about, and I'm very excited to see what others do with Blender and tools like TouchDesigner and Mediapipe.

More than anything though, this experiment serves to remind us that creating great puppetry is hard, creating digital puppetry is especially hard...and recreating shadow puppetry in particular is deceptively hard. 

Monday, April 13, 2015

Digital Wayang Kulit



Here's a nine minute demo of a Digital Wayang Kulit (Indonesian shadow puppetry) program developed at the MSc in Digital Education program at University of Edinburgh. The 2D figure is controlled by the digital Dalang (puppeteer) using a Gametrak controller and a Wiimote.

There have been a lot of shadow puppet inspired digital puppetry demos created over the years (this one is from 2013), but I love how fluid the movement of this one is.

Special thanks to Jane for submitting this!