--- title: NVIDIA Video Codec SDK 13.1: Zero-Copy Transcode, AV1... description: NVIDIA Video Codec SDK 13.1 introduces several major improvements for GPU-accelerated video pipelines. Key encode additions include Hierarchical Reference Mode... canonical: https://daily.dev/posts/nvidia-video-codec-sdk-13-1-zero-copy-transcode-av1-b-frames-and-frame-accurate-seek-k2lcpkzvh twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: NVIDIA Video Codec SDK 13.1: Zero-Copy Transcode, AV1 B-Frames, and Frame-Accurate Seek | daily.dev og:description: NVIDIA Video Codec SDK 13.1 introduces several major improvements for GPU-accelerated video pipelines. Key encode additions include Hierarchical Reference Mode... og:url: https://daily.dev/posts/nvidia-video-codec-sdk-13-1-zero-copy-transcode-av1-b-frames-and-frame-accurate-seek-k2lcpkzvh og:image: https://api.daily.dev/og/posts/K2lCPkzVh.png og:image:alt: NVIDIA Video Codec SDK 13.1: Zero-Copy Transcode, AV1 B-Frames, and Frame-Accurate Seek og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # NVIDIA Video Codec SDK 13.1: Zero-Copy Transcode, AV1 B-Frames, and Frame-Accurate Seek **[NVIDIA Developer](https://daily.dev/sources/nvidiadev)** · 12 min read · 1 upvotes · 0 comments ## Summary NVIDIA Video Codec SDK 13.1 introduces several major improvements for GPU-accelerated video pipelines. Key encode additions include Hierarchical Reference Mode for AV1 supporting up to 31 B-frames for better quality without performance penalties, and UHQ tuning info combined with iterative encoding. On the decode side, per-macroblock statistics (QP, coding unit type, motion vectors) are now exposed for H.264/HEVC with no CPU overhead, enabling GPU-accelerated analytics. A new frame-accurate seek API allows random access to specific frames via a GOP-aware architecture, useful for AI inference and video editing workflows. The transcoder pipeline is redesigned with a modular, queue-based architecture and a new AppTransZeroCopy application that eliminates intermediate buffer copies by sharing CUarray surfaces between NVDEC and NVENC, reducing SM utilization and memory footprint. An official Docker-based development environment is also included for reproducible setup. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What new AV1 encoding feature does NVIDIA Video Codec SDK 13.1 add for NVENC? SDK 13.1 adds Hierarchical Reference Mode for AV1, increasing NVENC's maximum B-frames from 7 to 31 by arranging them in a tree-like reference structure with non-reference leaf nodes and a middle B-frame root. It supports 1, 3, 7, 15, and 31 B-frame configurations, is most effective at seven or more, adds no performance penalty, though video memory consumption increases. H.264 and HEVC support is planned for future driver versions. _Track NVENC and AV1 encoding updates like this on daily.dev as video pipelines evolve._ ### How does zero-copy transcoding work in NVIDIA Video Codec SDK 13.1's AppTransZeroCopy sample? AppTransZeroCopy lets NVDEC and NVENC operate directly on the same GPU memory using application-allocated CUarray buffers created with cuArray3DCreate and the CUDA_ARRAY3D_VIDEO_ENCODE_DECODE flag. The same CUarray registers as an external output surface for the decoder and an input resource for the encoder, eliminating intermediate format-conversion copies, lowering SM utilization, reducing GPU memory footprint, and enabling higher concurrent-session throughput. _Developers optimizing GPU transcode pipelines can follow zero-copy techniques like this via daily.dev._ ### What versions does the official Docker environment for NVIDIA Video Codec SDK 13.1 pin? The Docker-based development environment for Video Codec SDK 13.1 pins CUDA 12.3.2, Vulkan SDK 1.4.304.1, and Ubuntu 22.04 LTS in a two-stage build (builder and runtime stages). Developers supply the SDK zip via the SDK_ZIP build argument and can override the FFmpeg build via FFMPEG_URL, with pre-built samples like AppDec and AppEncCuda available inside the container for immediate testing. _daily.dev helps engineers standardizing GPU video toolchains keep up with dev-environment changes like this._ ## Similar posts on daily.dev - [Real-Time Decoding, Algorithmic GPU Decoders, and AI Inference Enhancements in NVIDIA CUDA-Q QEC](https://daily.dev/posts/real-time-decoding-algorithmic-gpu-decoders-and-ai-inference-enhancements-in-nvidia-cuda-q-qec-kvhzxiihf) · NVIDIA Developer · 6 upvotes · 0 comments - [NVIDIA VA-API Driver 0.0.18 Ships More Fixes, Chrome/Chromium Decode On Direct Backend](https://daily.dev/posts/nvidia-va-api-driver-0-0-18-ships-more-fixes-chrome-chromium-decode-on-direct-backend-ejpz4kpdu) · Phoronix · 0 upvotes · 0 comments - [HandBrake 1.11 adds ProRes, DNxHR encoders and MOV output](https://daily.dev/posts/handbrake-1-11-adds-prores-dnxhr-encoders-and-mov-output-npu4efrh5) · omg\! ubuntu\! · 2 upvotes · 0 comments - [Video Encoding and Decoding with Vulkan Compute Shaders in FFmpeg](https://daily.dev/posts/video-encoding-and-decoding-with-vulkan-compute-shaders-in-ffmpeg-wmurcqryk) · Khronos · 0 upvotes · 0 comments --- Tags: [#gpu](https://daily.dev/tags/gpu), [#cuda](https://daily.dev/tags/cuda) [View this post on daily.dev](https://daily.dev/posts/nvidia-video-codec-sdk-13-1-zero-copy-transcode-av1-b-frames-and-frame-accurate-seek-k2lcpkzvh) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"NVIDIA Video Codec SDK 13.1: Zero-Copy Transcode, AV1 B-Frames, and Frame-Accurate Seek","url":"https://daily.dev/posts/nvidia-video-codec-sdk-13-1-zero-copy-transcode-av1-b-frames-and-frame-accurate-seek-k2lcpkzvh","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/nvidia-video-codec-sdk-13-1-zero-copy-transcode-av1-b-frames-and-frame-accurate-seek-k2lcpkzvh"},"datePublished":"2026-07-31T15:17:26.635Z","dateModified":"2026-09-13T21:32:44.026Z","description":"NVIDIA Video Codec SDK 13.1 introduces several major improvements for GPU-accelerated video pipelines. Key encode additions include Hierarchical Reference Mode...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/7ae783a1eadaa19a1ab4e33b495db8a3?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/7ae783a1eadaa19a1ab4e33b495db8a3?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"NVIDIA Developer","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"NVIDIA Developer","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/86e45aab42ba48ce83103d01b1119910","url":"https://daily.dev/sources/nvidiadev"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/nvidia-video-codec-sdk-13-1-zero-copy-transcode-av1-b-frames-and-frame-accurate-seek-k2lcpkzvh","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":1},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"gpu,cuda","timeRequired":"PT12M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"NVIDIA Developer","item":"https://daily.dev/sources/nvidiadev"},{"@type":"ListItem","position":3,"name":"NVIDIA Video Codec SDK 13.1: Zero-Copy Transcode, AV1 B-Frames, and Frame-Accurate Seek"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/nvidia-video-codec-sdk-13-1-zero-copy-transcode-av1-b-frames-and-frame-accurate-seek-k2lcpkzvh#faq","mainEntity":[{"@type":"Question","name":"What new AV1 encoding feature does NVIDIA Video Codec SDK 13.1 add for NVENC?","acceptedAnswer":{"@type":"Answer","text":"SDK 13.1 adds Hierarchical Reference Mode for AV1, increasing NVENC's maximum B-frames from 7 to 31 by arranging them in a tree-like reference structure with non-reference leaf nodes and a middle B-frame root. It supports 1, 3, 7, 15, and 31 B-frame configurations, is most effective at seven or more, adds no performance penalty, though video memory consumption increases. H.264 and HEVC support is planned for future driver versions. Track NVENC and AV1 encoding updates like this on daily.dev as video pipelines evolve."}},{"@type":"Question","name":"How does zero-copy transcoding work in NVIDIA Video Codec SDK 13.1's AppTransZeroCopy sample?","acceptedAnswer":{"@type":"Answer","text":"AppTransZeroCopy lets NVDEC and NVENC operate directly on the same GPU memory using application-allocated CUarray buffers created with cuArray3DCreate and the CUDA_ARRAY3D_VIDEO_ENCODE_DECODE flag. The same CUarray registers as an external output surface for the decoder and an input resource for the encoder, eliminating intermediate format-conversion copies, lowering SM utilization, reducing GPU memory footprint, and enabling higher concurrent-session throughput. Developers optimizing GPU transcode pipelines can follow zero-copy techniques like this via daily.dev."}},{"@type":"Question","name":"What versions does the official Docker environment for NVIDIA Video Codec SDK 13.1 pin?","acceptedAnswer":{"@type":"Answer","text":"The Docker-based development environment for Video Codec SDK 13.1 pins CUDA 12.3.2, Vulkan SDK 1.4.304.1, and Ubuntu 22.04 LTS in a two-stage build (builder and runtime stages). Developers supply the SDK zip via the SDK_ZIP build argument and can override the FFmpeg build via FFMPEG_URL, with pre-built samples like AppDec and AppEncCuda available inside the container for immediate testing. daily.dev helps engineers standardizing GPU video toolchains keep up with dev-environment changes like this."}}]} ```