Follow developments in AV1, AV2, H.264/AVC, HEVC/H.265, VVC/H.266, FFmpeg, cloud transcoding, hardware acceleration, bitrate optimization, and video quality measurement. StreamingMeme curates and summarizes the updates that matter to streaming engineers, product teams, and media technology leaders.
The Alliance for Open Media is developing the AV2 video codec, which aims to provide 30% better compression efficiency than AV1. Google plans to begin deploying AV2 on YouTube in early 2027, with support for multi-view streaming, low-latency communications, and improved HDR rendering.
China's Supreme People's Court has rejected InterDigital's appeal regarding an AV1 and VP9 codec patent, marking the company's ninth consecutive appellate defeat in the country. The ruling upholds a previous decision to invalidate several patent claims for lack of inventive step, impacting InterDigital's intellectual property portfolio in the video codec space.
OBS Studio 33.0 has been released, featuring an upgrade to the Chromium Embedded Framework version 150 and a new directory structure for third-party plugins. The update also introduces NVENC support for Arm64 systems, improved source snapping, and various stability fixes for recording and streaming workflows.
YouTube has announced plans to implement the AV2 video codec for web-based streams by early 2027, aiming to leverage its 30% improvement in compression efficiency over AV1. While the format has broad industry support, widespread adoption is currently hindered by the high computational requirements for hardware decoding, prompting a focus on software-based implementations like the dav2d decoder.
Qualcomm has filed a patent for a system that offloads complex video encoding tasks from XR headsets to smartphones. By using the phone to predict and calculate compression parameters, the technology aims to reduce heat, weight, and battery consumption on head-mounted devices.
Scalstrm is promoting its ability to deliver 110 broadcast-grade HD channels from a single 1U server by integrating NETINT video processing units with its own cloud-native software. The solution is designed to help pay-TV operators and broadcasters reduce infrastructure costs and increase channel density.
FFmpeg 9.1 has been released, introducing a new Vulkan-accelerated APV encoder and NVIDIA optical flow interpolation for frame rate up-conversion. The update also includes various codec improvements, filter enhancements, and support for new formats like LCEVC payload merging.
The Alliance for Open Media reports that its AV2 codec is seeing rapid industry adoption four months after its release, with 30% greater compression efficiency than AV1. Major members including Google, YouTube, and Intel are actively developing support, with YouTube planning to begin streaming in AV2 by early 2027.
Boris FX has released Vegas Pro 2027, an update to its video editing platform that introduces GPU-accelerated encoding for NVIDIA and Intel hardware, a rebuilt color management pipeline, and new AI-driven masking and transcription tools. The release also includes expanded support for professional color spaces and high-resolution audio exports.
MainConcept has released Codec SDK 17.0, which extends its Easy Video API (EVA) to support end-to-end transcoding pipelines on NVIDIA Blackwell GPUs for Linux Arm and NETINT Quadra VPUs for Linux x86. This update aims to unify hardware-accelerated workflows under a single API to reduce integration costs and maintenance for developers.
Euroconsumers is demanding €62.8 million in compensation from Disney+ for subscribers in 11 EU countries, citing the loss of 4K, HDR, and 3D features due to an ongoing patent dispute with InterDigital. The organization alleges that Disney+ violated EU consumer laws by failing to deliver promised service quality and providing inadequate notice regarding the degradation of streaming features.
Researchers from the University of Hong Kong and Kling AI have introduced AVE, a unified, mask-free framework for human-centric audio-visual editing. The project includes a 100K-pair dataset and a multi-stage training curriculum designed to improve lip-speech synchronization and temporal consistency in video production workflows.
Bridge Technologies has containerized its monitoring tools to integrate with Grass Valley’s MXL-native AMPP ecosystem. This integration allows broadcasters to deploy monitoring microservices on-demand within software-defined media workflows.
Lipsie Languages has launched LipsieSync, a hybrid dubbing workflow that combines professional human voice acting with speech-to-speech AI to maintain vocal identity across multiple languages. The service is currently being utilized for documentary and creator-led content, with pricing starting at 70 euros per minute.
Researchers at Tsinghua University have introduced Persistent Planar Memory (PPMem), a method for generative AI video models that uses compact geometric plane primitives to maintain visual consistency. By replacing dense 3D point clouds with lightweight surface representations, the system reduces inference requirements and improves scene memory compared to existing approaches like Gen3C and SPMem.
Researchers from KlingAI have proposed Mise-en-Video (MeV), a framework for multi-shot video generation that utilizes 3D-aware layout guidance and hierarchical context attention. The model, trained on a 160K-video dataset, enables precise spatial and narrative control by allowing creators to stage shots in 3D space.
Bitmovin has launched Encoder Version 3, which introduces updated HEVC and AV1 codecs designed to reduce bitrates and cloud compute usage. The release also adds an MCP server for AI agent integration and updates pricing structures for AV1 and HDR formats.
VAST has open-sourced EdenLive, a generative AI model capable of real-time video re-rendering of coarse game engine output. The release includes 5B and 14B parameter models, which can achieve over 24 FPS at 720p on an Nvidia RTX 5090.
MKVToolNix 103 has been released, adding support for setting stereo mode from frame packing arrangement SEI messages for AVC and HEVC streams. The update also includes a fix for an MP4 reader bug that caused frame loss and various improvements to the GUI multiplexer and chapter editor.
Qligent is introducing a tiered software model for its Vision monitoring platform, which includes a free version for foundational broadcast compliance. The update adds support for hardware ingest, ST-2110, SDI, and ASI, while offering paid tiers for AI-driven operational intelligence and infrastructure observability.
Apple has announced the fourth-generation Apple TV 4K, powered by the new A19 chip which includes a dedicated AV1 hardware decoder. The device also features far-field microphones for hands-free Siri control and supports a new four-speaker surround sound configuration with HomePod speakers.
Bunny.net has launched an official Stream SDK for React Native and Expo, providing developers with native video playback, resumable uploads, and live streaming capabilities. The SDK utilizes TypeScript APIs and supports both iOS and Android platforms to streamline mobile video integration.
Researchers from Fudan University and OPPO have introduced EVA-Cap, an audiovisual captioning framework that utilizes event-centric alignment and reinforcement learning to improve temporal synchronization and character attribution. The model, trained on the new 63K-pair EVA-Data dataset, demonstrates significant performance gains over existing open-source models on both general and event-centric audiovisual benchmarks.
Researchers from Wuhan University and Alibaba Group have developed StreamErase, a framework for real-time, sub-second latency video object removal. The model utilizes a three-stage distillation pipeline to achieve 26 FPS performance on a single H200 GPU, enabling stable, long-horizon video editing.
Apple has released the 4th generation Apple TV 4K, which includes a new A19 chip featuring a dedicated AV1 hardware decoder and support for Dolby Vision Profile 8.1. These updates enable more efficient video compression and broader format compatibility for streaming services that adopt these standards.
Apple has announced a new Apple TV 4K model featuring the A19 processor, native AV1 codec support, and integrated Siri AI capabilities. The device also functions as a smart home hub, supporting a new line of security and climate accessories developed in partnership with LG Electronics and Schneider Electric.
Spanish broadcaster RTVE has deployed Ateme’s Titan Edge platform across three new transportable production units to handle satellite contribution. The software-based solution integrates encoding, modulation, and decoding into 1RU appliances to reduce hardware complexity and streamline field operations.
A report from InterDigital and Parks Associates indicates that HDR is transitioning from a premium differentiator to a standard expectation across streaming, live sports, and gaming. The study highlights that AI-assisted enhancement and cloud production are key drivers in reducing the technical barriers to widespread HDR implementation.
PlayBox Neo has updated its broadcast automation platform to include role-based security, LDAP support, and improved BXF integration for traffic systems. The update also introduces a new integration with EditShare, enabling editors to access media during the ingest process.
Apple has announced the fourth-generation Apple TV 4K, featuring the new A19 processor, hardware-based AV1 video decoding, and AI-powered Siri voice controls. These updates impact how broadcasters and streaming services optimize their distribution workflows and content discovery strategies for connected television platforms.
Advanced Coding Technologies (ACT) has filed a lawsuit against Amazon in the Eastern District of Texas, alleging patent infringement related to the AV1 video codec. This litigation is part of a broader trend of patent holders challenging the royalty-free status of the AV1 standard, following similar legal actions against other major technology companies.
Nvidia has integrated its RTX Video Super Resolution, TrueHDR, and frame generation tools into the open-source ComfyUI node package. This update allows creators to perform AI-driven video processing and HDR conversion locally on RTX 30, 40, and 50 series GPUs while preserving audio tracks.
Spanish broadcaster RTVE has deployed Ateme's TITAN Edge software platform within three new mobile production units to manage satellite contribution workflows. The integration consolidates encoding, modulation, and decoding into a single 1RU appliance to reduce hardware complexity and improve operational efficiency in live field environments.
RTVE has deployed Ateme's TITAN Edge platform to support satellite contribution for three new transportable production units. The software-based solution integrates encoding, modulation, demodulation, and decoding into a single 1RU appliance to streamline field operations.
VVC adoption is accelerating as major smartphone manufacturers integrate hardware support and join the VVC Advance patent pool. This shift addresses the codec's deployment barrier by providing the necessary device scale for streaming providers to adopt the standard.
Microsoft is expanding its Advanced Shader Delivery (ASD) technology to Windows 11, allowing precompiled shaders to be distributed alongside games for AMD, Qualcomm, Intel, and Nvidia hardware. By shifting shader compilation from local user devices to cloud-based infrastructure, the system significantly reduces initial game launch times and runtime stutter.
The JPEG Committee has finalized the JPEG XE standard (ISO/IEC 26112-1) as the first international standard for lossless coding of visual events from event-based vision sensors. The committee also advanced several other standards, including JPEG Trust, JPEG DNA, JPEG Pleno, JPEG XS, and JPEG AI, during its 112th meeting in Leiria, Portugal.
A developer has identified a persistent chroma drift issue in NVIDIA's Vulkan Video H.265 encoder when processing 3840x2160 content. The bug, which manifests as color degradation in the top-left corner of static scenes, appears to be an encoder-side reference picture mismatch that does not occur with other hardware encoders or alternative Vulkan implementations.
ReflowPDF has released its CSS conformance test results, reporting a 94.62% pass rate across the Web Platform Tests suite for its browser-less HTML-to-PDF engine. The company published detailed breakdowns by CSS specification and property to provide transparency into its rendering capabilities.
Crusoe engineers provide a technical breakdown of GPU execution, comparing traditional CUDA C++ programming with the Triton domain-specific language. The article details how PyTorch kernel launches function and how Triton simplifies block-level GPU programming for developers.
A 2026 industry survey conducted by NETINT indicates that 57% of organizations are currently using or planning to deploy AV1, significantly outpacing the 33% adoption rate for H.266/VVC. The report highlights that while AV1 is gaining traction for web delivery, H.266/VVC adoption is hindered by licensing costs and limited hardware support.
Prodys has launched QuantumX, a software-defined audio and video codec that runs as a containerized application in Linux, Docker, or Kubernetes environments. The solution is designed to replace dedicated physical hardware with a scalable, subscription-based virtualized model for broadcast contribution and distribution.
Cobalt Digital is introducing an expanded ST 2110 and IPMX ecosystem at NAB Show New York 2026, featuring new JPEG XS support and hybrid satellite-internet transport capabilities. The updates aim to assist broadcasters in managing SDI-to-IP transitions and maintaining resilient content distribution during C-band spectrum reconfigurations.
LucidLink has updated its Connect platform with a new management interface that allows IT administrators to monitor and manage distributed storage sources directly. The update enables teams to access files across various storage systems through a single LucidLink Filespace without requiring data migration.
A Video.js developer explored the implementation of custom web-based video player layouts for the iPhone Duo using the Device Posture and Viewport Segments APIs. The experiment demonstrates how developers can leverage these APIs to create dual-screen UI experiences, while noting current limitations in browser fullscreen behavior on folding devices.
NVIDIA has expanded its Green Contexts feature to the CUDA 13.1 Runtime API, enabling developers to explicitly partition GPU execution resources like SMs and workqueues. This update allows latency-sensitive kernels to run concurrently with bulk workloads, improving performance for multi-component streaming and AI applications.
Wohler outlines the technical requirements for confidence monitoring in immersive audio workflows, emphasizing the need to validate metadata and rendering behavior alongside signal integrity. The company highlights its iVAM2-CMD and MAVRIC solutions as tools for end-to-end monitoring of object-based audio formats like Dolby Atmos and MPEG-H.
Executives from Qualcomm and VITEC discuss the strategic importance of VVC compression in managing infrastructure costs and energy consumption amid rising AI-generated video traffic. The discussion emphasizes that codec innovation is becoming a critical business metric for sustainable video delivery across the entire ecosystem.
Researchers have introduced a unified compression and editing framework for AI-generated video that utilizes a frozen generative model to preserve semantic consistency. By transmitting compact side information, the codec enables both high-quality reconstruction and prompt-based editing without requiring separate models or additional training.
PyTorch has launched TorchCodec, a new library designed to centralize all media decoding and encoding capabilities for the PyTorch ecosystem. This architectural change narrows the scope of TorchVision and TorchAudio to focus exclusively on media transformation, aiming to improve performance and simplify maintenance for developers building generative AI and vision-language models.
