Extraweg (AKA Oliver Latta: Twitter, Youtube) is a superb and surreal animator whose computer rendered grotesques of human faces and forms writhing, merging, and doing the impossible are spellbindingly weird. Read the rest
Brilliant Game Studios follows up on their 2016 video that showed off a crowd-renderer by pitting 11,000 penguins against 4,000 Santas with a new video demoing "our new GPU accelerated dismemberment and procedural wounding system. Limbs can be cut off on a massive scale. Wound from getting hit appear as deep gouges which warp their mesh." I'm really glad that we're using GPUs to do sensible things again, rather than computing virtual Beanie Babies. (via JWZ) Read the rest
In 1991, Jean Henri Gaston Giraud (aka Moebius) and a team of animators created this gorgeous short pilot for a film called Starwatcher. According to this Wired feature that Mark wrote in 1994, "Starwatcher was slated to be the first feature-length animated movie to be made with 3-D computer graphics. But the film's producer died in a car accident, and shortly thereafter it was discovered that the French production company bankrolling the film was FF85 million (US$15 million) in debt. (Many suspected the car accident was no accident.)"
This whimsical series of images by Filtre Studio imagines Queen Elizabeth straightening paintings and vacuuming up after her dogs. What's most interesting is that the entire room was created digitally. Read the rest
Stranger Things special effects producers Paul and Christina Graff explain the season two VFX of the Upside Down, the demodog Dart, and that shadowy motherfucker from the finale.
Researchers from graphics company NVIDIA developed a deep learning system that automatically reconstructs corrupted images or fills in missing parts:
Read the rest
The method, which performs a process called “image inpainting”, could be implemented in photo editing software to remove unwanted content, while filling it with a realistic computer-generated alternative.
“Our model can robustly handle holes of any shape, size location, or distance from the image borders. Previous deep learning approaches have focused on rectangular regions located around the center of the image, and often rely on expensive post-processing,” the NVIDIA researchers stated in their research paper. “Further, our model gracefully handles holes of increasing size.”
Physicists at BYU have demonstrated a volumetric projection system that works by using a laser to unevenly heat single cellulose molecules in order to shove them around in 3D space, then painting the positioned molecules with lasers that cause them to glow; by choreographic both sets of lasers, extremely high-resolution moving images can be attained. Read the rest
Argentinian motion graphics designer Esteban Diacono produces the most amazing, surreal little animations for his social media, especially the weird, motion-deforming humaniform figures. Read the rest
An "adversarial perturbation" is a change to a physical object that is deliberately designed to fool a machine-learning system into mistaking it for something else. Read the rest