Nearly Data Moshing in Jitter – Ned Rush Gets Glitchy with Max and Push

Ned Rush explores “data moshing (sort of)” using cv.jit motion tracking and feedback loops in Max for Live. No FFmpeg. No file editing. Just webcam-powered glitch madness—live and hands-on.

Motion-Driven Glitch FX Using cv.jit & Real-Time Feedback Loops

In this experimental (and chaotically fun) Max for Live exploration, Ned Rush attempts something “kind of like” data moshing—only it’s done in real-time with Jitter, cv.jit, and a webcam. There’s no FFmpeg, no Avidemux, no actual I-Frame deletion here—just clever motion tracking, displacement textures, and feedback loops that push your visuals into warped, distorted territory.

What starts as a camera feed quickly mutates into a live texture loop, with movement data captured from your webcam manipulating pixels in ghostly, reactive patterns. It’s messy. It’s brilliant. It’s classic Ned Rush.

What You’ll Need

• Max/MSP
• cv.jit package (download via Max’s Package Manager)
• A webcam or video file
• A willingness to let things glitch beautifully

Key Techniques Covered

Motion Tracking with cv.jit.lkflow

Ned shows how to extract motion direction from your webcam—up, down, left, right—and split that data into separate visual planes for further control.

Creating a Texture Feedback Loop

Using jit.gl.texture, jit.gl.pix, and zl.redge, Ned sets up a loop that constantly feeds a stored frame back into itself—ideal for that melting, warping data-mosh vibe.

Mapping Motion to Displacement

The heart of the patch uses jit.gl.pix to sample and displace pixel coordinates based on live movement data. A set of jit.3m and math objects translate gesture intensity into gain-scaled offsets that visually “push” the texture around in real-time.

Threshold Filtering for Control

To keep things readable (and less chaotic), Ned adds jit.op thresholding to isolate stronger movement data—ensuring only significant motion affects the feedback image.

Why It’s Not Technically Data Moshing

True data moshing involves manipulating encoded video data—motion vectors, I-frames, compression artifacts. What Ned’s doing here is real-time distortion and motion-based feedback, but it shares that smeared, pixel-drifting aesthetic we associate with glitched-out moshed footage. So no, it’s not purist, but yes—it’s properly cool.