Video Encoding, Codec & Transcoding News

Sigmedia Group, based at Trinity College Dublin, hosted Charles Poynton, an image and color scientist known for his work in digital video systems. Poynton visited after giving a masterclass on color science and discussed Sigmedia's research in color and HDR, virtual production, denoising, AV2, learned-image compressions, and future video/image coding.

MForja has launched a demo player showcasing its proprietary entropy-aware Video Distillation technology, claiming to achieve over 50% signal waste removal without perceptible quality loss. The demo allows users to view real video samples processed by MForja's technology across 4K, 1080p, and various content types, comparing file sizes before and after processing.

A GitHub project, 'Real_Time_Video_Processing_Pipeline,' was released, offering a production-grade system that simulates two camera streams from a single video file. The pipeline incorporates real-time pacing, quality filtering (blur, exposure, stuck frames), preprocessing for AI models, and live health metrics reporting. It is designed for continuous operation with stable memory usage and clean shutdown.

Dalet released its Flex Media Asset Management (MAM) 25.12 LTS update, showcasing three specific workflows. These workflows include using spectral analysis for automated content moderation, AI-driven metadata extraction to power search and recommendations as well as creating dynamic highlights for streaming platforms and social media.

The article discusses the common tool stack used by video engineers for codec comparison, highlighting the limitations of current tools for analyzing codecs like H.264, HEVC, VP9, and AV1. It specifically mentions MediaInfo, Bitrate Viewer, and Moscow State University VQMT, noting their critical gaps in providing comprehensive data for RD curves and BD-Rate.

Watch Brasil, a streaming platform operating in Brazil and Europe, has adopted Bitmovin's Player, VOD Encoder, and Observability solutions. This implementation replaces their previous legacy systems. The platform has been operating since 2018.

Chunkify published benchmarks comparing its video transcoding performance, cost, and quality against AWS Elemental MediaConvert for H.264, HEVC, and AV1 codecs across various resolutions. The results indicate Chunkify completed jobs 2.49x to 21.69x faster and with 8% to 76% lower costs, while maintaining comparable or higher VMAF quality scores.

Ateme and AMD EPYC™ are collaborating to enhance OTT efficiency through solutions that reduce server count and total cost of ownership. Their joint efforts focus on AV1 optimization and ensuring premium Quality of Experience (QoE) for sustainable streaming.

This article details technical lessons learned from building a video transcoding pipeline, focusing on two specific bugs related to PHP memory management during cloud storage downloads and filesystem I/O performance in containerized environments. It also highlights an optimization using FFmpeg's `filter_complex` for single-pass multi-resolution encoding.

A technical analysis details the Kling 3.0 API's unified multimodal architecture, which enables native 4K rendering and improved subject consistency logic for scalable video production. The article provides insight into the API's capabilities in handling high-resolution video workflows.

Joyn Switzerland, a part of ProSiebenSat.1, has implemented Zattoo Ingest for live TV processing. This deployment marks Joyn Switzerland as the first customer to fully utilize Zattoo Ingest in a production environment, processing 75 linear live signals.

This article discusses the fundamental role of data compression in modern civilization, detailing how lossless and lossy algorithms (including H.264, HEVC, AV1) reduce data size while preserving information. It highlights compression's impact on global streaming, cloud computing, and mobile internet, noting the increasing reliance on AI and machine learning for future codecs.

This article discusses the fundamental role of data compression in modern technology, particularly highlighting its importance for streaming video, internet functionality, and data storage. It explains the evolution from lossless to lossy compression, detailing various algorithms and their significant impact on reducing bandwidth and storage requirements for high-resolution content like 4K video. The article also touches on the increasing integration of AI and machine learning in developing future compression technologies.

Scalstrm and NETINT are co-hosting a full-day technical session titled "Streaming Architecture: Encoding Efficiency at Scale" in Stockholm on May 22, 2026. The event, aimed at streaming engineers and platform teams, will focus on strategies for optimizing encoding efficiency at scale, including the role of video processing units (VPUs) and dynamic orchestration workflows. Speakers from both companies will address the challenges and solutions for reducing costs, energy consumption, and operational complexity in streaming infrastructure.

Google Project Zero identified a critical zero-click exploit chain affecting Pixel 10 devices, stemming from vulnerabilities in the Dolby UDC exploit and a new driver for the Chips&Media Wave677DV video processing unit (VPU). The VPU vulnerability, which allows mapping significant portions of physical memory into userspace and potentially overwriting kernel functions, was patched in the February Pixel security bulletin, 71 days after discovery.

Chips&Media announced the completion of its hardware IP development for the AOMedia AV2 next-generation video codec. This development provides a hardware decoder for AV2, which is designed to offer 20-30% better compression efficiency than AV1, targeting North American OTT market adoption by platforms like YouTube and Netflix.

Akamai has launched a new 8-card VPU plan for Accelerated Compute Instances, expanding its cloud-based video processing offerings developed in partnership with NETINT. This new plan, powered by NETINT Quadra T1U VPUs, is designed to provide high-density transcoding capacity, enabling over 320 parallel streams for video on demand, live events, and user-generated content platforms, while also offering general-purpose compute for tasks like deinterlacing and audio processing.

StreamKernel-io is a newly released source-available JVM pipeline kernel designed for policy-aware, benchmarkable operational data movement, offering in-process AI enrichment, policy enforcement, transformation, and multi-destination delivery. This GitHub repository provides the runtime boundary, SDK contracts, and benchmark evidence, explicitly withholding private AI model artifacts and commercial implementation details for proprietary licensing.

Adobe published security advisories on May 12, 2026, to address critical vulnerabilities in several of its software products, including Adobe Premiere and Adobe Premiere Pro, Media Encoder, and After Effects. The Canadian Centre for Cyber Security recommends users and administrators review the advisories and apply available updates.

Softvelum's Nimble Streamer now supports an expanded set of audio and video codecs over Enhanced RTMP, including VP8, VP9, Opus, FLAC, AC3, and E-AC3. This update, based on the Veovera Enhanced RTMP specification, allows these codecs to be received, pulled, republished, and played back directly within Nimble Streamer's RTMP modes, reducing the need for transcoding in certain streaming workflows.

Tencent MPS offers a suite of media processing and live streaming solutions, including encoding, enhancement, streaming, and playback products. The platform claims to reduce bitrates by over 50% without quality compromise, support 8K/144FPS transcoding, and achieve cross-continental transmission latency under 400ms. Tencent MPS highlights its applicability across various industries such as pan-entertainment, online education, live sports, broadcast/OTT, e-commerce, and gaming.

Solveig Multimedia will unveil advancements in video analysis and cloud-based video solutions at Asia Tech x Singapore 2026. The company plans to introduce a new version of its Zond 265 HEVC Analyzer with VVC Codec support and an enhanced HTML5 Cloud Video Editor featuring an integrated AI Highlight Detection Module for sports events.

imperishableSecret has developed a prototype VA-API driver for NVIDIA GPUs, integrating Vulkan Video API for decoding H.264, HEVC, VP9, and AV1 codecs. The project aims to enable browser decode by translating VA decode requests into Vulkan Video and exporting dma-bufs, supporting NV12 and P010 export for various profiles.

Chips&Media has announced the completion of its hardware IP development for AV2, the next-generation high-efficiency video codec developed by AOMedia. The new AV2 decoder, which promises 20-30% improved compression efficiency over AV1, will be integrated into their 'WAVE' series and offered in customizable configurations for resolutions up to 8K@60fps. Chips&Media is targeting the North American OTT market, including YouTube and Netflix, and plans to release the hardware RTL in May.

Eyevinn has developed 'hi264', a pure Go H.264/AVC IDR decoder and bitstream generator. This toolkit enables seamless extension of fMP4 fragments with additional frames, creation of test content with grid patterns and overlays, and extraction of IDR frames as thumbnails, all without external dependencies like FFmpeg or cgo.

Bamboo Cloud published a blog post advising OTT operators on implementing 4K HDR streaming, outlining its technical aspects, including codecs like HEVC and AV1, bitrate ladders, and DRM requirements. The article describes 4K HDR as an expected baseline quality for modern smart TVs and details Bamboo Cloud's approach to handling these video formats.

V-Nova leverages hierarchical, compute-aware data formats to enhance video compression efficiency and quality across broadcast and streaming platforms. Their technologies include MPEG-5 LCEVC for video coding, SMPTE VC-6 for visual computing, and PresenZ for cinematic XR experiences. The company emphasizes enabling AI-ready features by structuring visual data for greater efficiency in AI processing workflows.

A technical analysis demonstrates that Apache Kafka's new Queues feature, specifically share groups, can significantly increase consumer throughput by mitigating head-of-line blocking when message processing involves delays, showing linear scaling up to at least an 8x increase. This scaling comes with the trade-off of losing ordering guarantees for messages. The benchmark results indicate that share groups can be used to scale processing beyond the number of partitions.

The JPEG XS standard (ISO/IEC 21122) specifies a compression technology designed for visually lossless compression with low latency and hardware implementation complexity. It supports various pixel formats and offers compression ratios up to 10:1, making it suitable for applications that previously transported uncompressed image data, such as professional video links and IP transport. The specification is multi-part, covering core coding, profiles, transport, conformance testing, and reference software.

V-Nova promotes its MPEG-5 LCEVC technology, highlighting its ability to optimize video quality, reduce bandwidth, and lower encoding costs across various applications, including streaming platforms, broadcast TV, social media, and gaming services. The technology is presented as offering up to 40% bitrate savings, 2-3x faster encoding, and improved VMAF performance, with NVIDIA supporting its use for efficiency gains in challenging streaming scenarios and integration into Next-Gen TV 3.0 in Brazil.

Colorfront is hosting a webinar on May 28, 2026, to showcase the new features of its Transkoder 2026 software, which were unveiled at NAB 2026. The updates include improved usability, enhanced audio and timeline tools, advanced content analysis, QC and media validation capabilities, AI-powered integrations like Super Resolution, and Dolby Vision 2 workflow advancements.

NETINT Technologies, a pioneer in Video Processing Units (VPUs), reflects on its journey to category creation, marked by receiving a 75th Technology and Engineering Emmy Award alongside AMD, Google, and Meta for efficient hardware video accelerators. The article highlights the growing industry shift toward purpose-built silicon for video encoding, driven by escalating video demand and the limitations of general-purpose CPUs and GPUs. NETINT details its product development, ecosystem integration efforts, and customer success stories demonstrating the power efficiency and density benefits of VPUs.

Chips&Media announced the completion of its hardware IP development supporting AV2, the next-generation high-efficiency video codec from AOMedia. The new AV2 decoder targets a 20-30% compression efficiency improvement over AV1 and will be integrated into their 'WAVE' series of multi-codec hardware IPs, offering customizable configurations up to 8K@60fps. The company is actively pursuing commercial licensing with major North American clients, particularly OTT platforms like YouTube and Netflix.

GStreamer has released version 1.28.3 of its multimedia framework, which includes security fixes and improvements to playback stability. Updates target components for NVIDIA, OpenGL, V4L2, RTSP, and Apple media.

HISPlayer has released an OpenXR-optimized SDK for its Unity and Unreal Engine video players, enabling cross-platform XR video streaming development. This SDK supports high-quality HLS & DASH streaming of 180°, 360°, 8K video with Widevine DRM across various AR/VR devices through a single integration, simplifying development previously fragmented by platform-specific requirements.

Limecraft has updated its AI-powered production platform to version 2026.3, introducing file verification with support for ASC Media Hash List technology. The update aims to enhance the security and integrity of media files within production workflows.

PlayBox Neo will introduce new upgrades to its PlayBox Neo Suite at Broadcast Asia 2026. These enhancements focus on security and streamlining hybrid broadcast operations.

A Lex Fridman Podcast transcript features Jean-Baptiste Kempf (VideoLAN) and Kieran Kunhya (FFmpeg) discussing the history, technical complexities, and community surrounding FFmpeg and VLC. The conversation spans topics including video codecs, open-source development models, the challenges of low-level assembly optimization, ultra-low latency streaming, and the future of multimedia archiving. Notable discussions include the historical significance of x264, the development of AV1/AV2 codecs, and the practical challenges of open-source maintenance.

Lex Fridman's podcast episode #496 features Jean-Baptiste Kempf (VLC lead developer) and Kieran Kunhya (FFmpeg contributor) discussing the technical details of FFmpeg, including video playback, codecs, containers, open-source development, and the history of FFmpeg and VLC. The conversation also touches on industry-specific topics such as the AV2 codec, video patents, and low-latency streaming.

Plex has released Plex Media Server version 1.43.2.10687 to all users, following its beta availability for Plex Pass users. This update introduces new filter capabilities for Atmos and video/audio/subtitle codecs, enhanced metadata support for RottenTomatoes ratings and Dolby Atmos detection, and includes fixes for several issues across hubs, metadata parsing, network TLS verification, and transcoding. Notably, it also generates unsigned QNAP packages for NAS devices, addressing previous availability issues.

NetActuate has partnered with NETINT Technologies to offer global, on-demand infrastructure for hardware-accelerated video processing. The service utilizes NETINT's Smart VPUs deployed across NetActuate's network, allowing customers to access video encoding, decoding, and AI-native processing as-a-service. Customers can deploy the VPUs on pre-installed virtual machines or request custom builds.

Net Insight has integrated InSync Technology’s FrameFormer software into its Nimbra Edge platform, adding motion-compensated frame rate and format conversion as a value-added service. Orchestrated through Nimbra Live Intelligence, the integration allows broadcasters and rights holders to deploy conversion capabilities flexibly within live workflows. The partnership aims to simplify global content distribution by adapting live feeds to different regional standards without requiring dedicated conversion infrastructure at every interconnect point.

DIRECTV has chosen Harmonic's VOS Media Software to modernize its U.S. direct-to-home (DTH) video platform. The new unified, software-based architecture, running in DIRECTV's private data center, will replace legacy systems for video ingest, playout, branding, ad insertion, and statistical multiplexing. The goal of the transformation is to reduce operational costs and improve scalability for delivering linear and pay-per-view content via satellite.

According to Netint's 2026 State of Video Encoding Survey of 286 professionals, 77% of teams report being constrained by budget or capacity rather than technology limitations. The survey also found that 30% of respondents do not track total cost of ownership, and another 30% operate without formal quality metrics. The article argues the industry's primary bottleneck has shifted from technical capability to organizational execution, with the key challenge now being the ability to operationalize existing technologies.

Colorfront has launched Colorfront Immersive Utility, a new macOS app for creating Apple Immersive Video (.aivu) files for playback on Apple Vision Pro. The application converts source video like ProRes into the MV-HEVC format, packages it with spatial audio and camera metadata, and outputs a single file ready for the device. Key features include GPU-accelerated encoding, foveation, color space conversion, multi-shot timeline support, and pre-flight validation tools.

At NAB Show 2026, Avid announced the debut of Avid Content Core, a cloud-native foundation for connected media workflows built on AWS, positioned for use by news organizations, broadcasters, studios, and streaming services. The company also highlighted continued deployment of Avid NEXIS and Avid Media Composer on AWS, including work with Amazon MGM Studios to launch cloud editorial environments globally and support distributed production teams.

Glookast has launched the Cinnafilm Tachyon plugin for its Media Producer and Media Services platforms, enabling GPU-accelerated video processing. The plugin utilizes the NVIDIA GeForce RTX 5070 GPU for standards conversion, frame rate conversion, and de-interlacing in broadcast and post-production workflows. This integration aims to provide automated conversion tasks with modes optimized for live broadcast and file-based post-production.