At the IBC 2026 conference, NVIDIA announced a major expansion to its NVIDIA AI for Media collection, a suite of GPU-accelerated software development kits (SDKs) and microservices designed to enhance audio, video, and augmented-reality effects for media workflows.
Product LaunchesNvidiaNVIDIA AI for Media
NVIDIA Expands AI for Media Collection with New Tools for Broadcast and Streaming
New tools in the collection include NVIDIA 3D Body Pose, which estimates 2D and 3D human joint locations and angles from video to enable motion tracking without marker-based systems. This technology is designed to support sports replay enhancement, biomechanics analysis, and animation workflows.
NVIDIA also introduced Video Frame Generation (VFG), which uses generative AI to insert new frames between existing ones to smooth motion. VFG can increase frame rates by 2x or 4x, with ongoing development aiming for 8x slow-motion support. Additionally, NVIDIA Video Super Resolution (VSR) uses AI to upscale video while reducing noise, blur, and compression artifacts, now supporting 10-bit video formats.
Other updates include NVIDIA TrueHDR, which enables real-time conversion of standard-dynamic-range (SDR) video to high-dynamic-range (HDR) output with up to 2,000 nits of brightness. For localization, NVIDIA enhanced its LipSync and Active Speaker Detection services, which help developers create systems that match mouth movements to target audio and identify multiple speakers in a scene.
Sources
- NVIDIAが映像フレーム補間や生中継翻訳などの放送業界向けAIコレクション「NVIDIA AI for Media」の大幅拡充を発表 （GIGAZINE, 2026-09-15）
- NVIDIA Blog