– player and number detection with RF-DETR – player tracking with SAM2 – team clustering with SigLIP, UMAP and K-means – number recognition with SmolVLM2
Splat settings: Splat MCMC algorithm, image size 3k, max splat 1 or 5 million splats, with Antialiasing, Stop Training after 80k. In COLMAP for SfM.. it’s basically all defaults, except same camera for all images, and set an initial parameter.
“The Lionsgate catalog is too small to create a model,” a source tells The Wrap. “In fact, the Disney catalog is too small to create a model.” … Another issue is the rights of actors and the model for remuneration if their likeness appears in an AI-generated clip. It is a legal gray area with no clear path.
𝐇𝐨𝐥𝐨𝐥𝐮𝐦𝐢𝐧𝐞𝐬𝐜𝐞𝐧𝐭™ 𝐃𝐢𝐬𝐩𝐥𝐚𝐲 (HLD for short) — razor thin (as thin as a 17mm), full 4K resolution, and capable of generating holographic presence at a magical price. A breakthrough display that creates a holographic stage for people, products, and characters, delivering the magic of spatial presence in a razor-thin form factor. Our patented hybrid technology creates embedded holographic layer creates the 3D volume, transforming standard video into dimensional, lifelike experiences.
Create impossible spatial experiences from standard video
Characters, people, and products appear physically present in the room
The embedded holographic layer creates depth and dimension that makes subjects appear to float in space
First scalable holographic illusion that doesn’t require a room-sized installation
The dimensional depth and presence of pepper’s ghost illusions, without the bulk
Built for the real world
FHD clarity (HLD 16″) and 4K clarity (HLD 27″ and HLD 86″), high brightness for any lighting environment
Thin profile fits anywhere traditional displays go
Wall-mounted, it creates the illusion of punching a hole through the wall into another dimension
Films beautifully for social sharing – the magic translates on camera
Works with what you have
Runs on your existing digital signage solution, CMS, and 4K video distribution infrastructure
Standard HDMI input or USB loading
2D video workflow with straightforward, specific requirements: full-size subjects on green/white backgrounds or created with our templates (Cinema4D, Unity, Adobe Premiere Pro)
This is Ray3. The world’s first reasoning video model, and the first to generate studio-grade HDR. Now with an all-new Draft Mode for rapid iteration in creative workflows, and state of the art physics and consistency. Available now for free in Dream Machine.
Ray3’s native HDR delivers studio-grade fidelity. It generates video in 10, 12 & 16-bit high dynamic range with details in shadows and highlights in vivid color. Convert SDR to HDR, export EXR for seamless integration and unprecedented control in post-production workflows.
Reasoning enables Ray3 to understand nuanced directions, think in visuals and language tokens, and judge its generations to give you reliably better results. With Ray3 you can create more complex scenes, intricate multi-step motion, and do it all faster.
With reasoning, Ray3 can interpret visual annotations enabling creatives to now draw or scribble on images to direct performance, blocking, and camera movement. Refine motion, objects, and composition for precise visual control, all without prompting.
Draft Mode is a new way to iterate video ideas, fast. Explore ideas in a state of flow and get to your perfect shot. With Ray3’s new Hi-Fi diffusion pass, master your best shots into production-ready high-fidelity 4K HDR footage. 5x faster. 5x cheaper. 100x more fun.
Ray3 offers production-ready fidelity, high octane motion, preserved anatomy, physics simulations, world exploration, complex crowds, interactive lighting, caustics, motion blur, photorealism, and detail nuance, delivering visuals ready for high-end creative production pipelines.
Ray3 is an intelligent video model designed to tell stories. Ray3 is capable of thinking and reasoning in visuals and offers state of the art physics and consistency. In a world’s first, Ray3 generates videos in 16bit High Dynamic Range color bringing generative video to pro studio pipelines.The all-new Draft Mode enables you to explore many more ideas, much faster and tell better stories than ever before.
On Tuesday, the three media companies filed a lawsuit against MiniMax, a Chinese AI company that is reportedly valued at $4 billion, alleging “willful and brazen” copyright infringement
“I have a beta version of it and it’s pretty amazing,” Kutcher said of the platform in a recent conversation with former Google CEO Eric Schmidt at the Berggruen Salon in Los Angeles.
“Why would you go out and shoot an establishing shot of a house in a television show when you could just create the establishing shot for $100? To go out and shoot it would cost you thousands of dollars,” Kutcher said. “Action scenes of me jumping off of this building, you don’t have to have a stunt person go do it, you could just go do it [with AI].
QuickTime (.mov) files are fundamentally time-based, not frame-based, and so don’t have a built-in, uniform “first frame/last frame” field you can set as numeric frame IDs. Instead, tools like Shotgun Create rely on the timecode track and the movie’s duration to infer frame numbers. If you want Shotgun to pick up a non-default frame range (e.g. start at 1001, end at 1064), you must bake in an SMPTE timecode that corresponds to your desired start frame, and ensure the movie’s duration matches your clip length.
How Shotgun Reads Frame Ranges
Default start frame is 1. If no timecode metadata is present, Shotgun assumes the movie begins at frame 1.
Timecode ⇒ frame number. Shotgun Create “honors the timecodes of media sources,” mapping the embedded TC to frame IDs. For example, a 24 fps QuickTime tagged with a start timecode of 00:00:41:17 will be interpreted as beginning on frame 1001 (1001 ÷ 24 fps ≈ 41.71 s).
Embedding a Start Timecode
QuickTime uses a tmcd (timecode) track. You can bake in an SMPTE track via FFmpeg’s -timecode flag or via Compressor/encoder settings:
Compute your start TC.
Desired start frame = 1001
Frame 1001 at 24 fps ⇒ 1001 ÷ 24 ≈ 41.708 s ⇒ TC 00:00:41:17
