GMAT — NVIDIA GPU media processing toolkit

Description

NVIDIA datacenter GPU video/image processing and transcode toolkit (formerly ffmpeg-gpu-demo) with smart decode and HEIF via NVENC/NVDEC.

URL

• Visit resource

Related resources

• Evaluation of NVENC Split-Frame Encoding (SFE) for UHD Video TranscodingAcademic evaluation of NVENC's split-frame encoding feature for UHD video transcoding throughput and quality.
• Evolution of NVENC Efficiency: Pascal to BlackwellLongitudinal academic analysis of NVENC hardware-encoder efficiency across GPU generations, including new UHQ tuning mode.
• AV1 Encoding with libsvtav1 on QencodeThe libsvtav1 encoder (SVT-AV1) is an AV1 implementation tuned for both compression efficiency and parallel performance. This guide covers…
• auto-transcode-watcherPython+watchdog+FFmpeg folder watcher with file stabilization, retry queue, rename templates, and YAML config.
• Open-Source Software Encode/Decode For H.266/VVC ProgressingPhoronix status roundup tracking maturity of the open-source VVC/H.266 ecosystem, including native FFmpeg decoder progress.
• FOSDEM 2024: VVdeC<>Arm — Optimizing an Open Source VVC Decoder for Arm ArchitecturesFOSDEM Open Media devroom talk detailing Arm-specific performance optimization work on the open-source VVdeC VVC decoder.
• Differentiable VMAF Re-implementation on PyTorchDifferentiable PyTorch reimplementation of VMAF metric enabling gradient-based codec/RDO optimization, with companion code.
• Performance Comparison of AV1, JEM, VP9, and HEVC EncodersFraunhofer HHI SPIE 2017 paper comparing codec internals, BD-BR, and complexity across AV1, JEM, VP9, and HEVC.
• nano-hevcMinimal educational HEVC encoder written in pure Python, readable-over-fast, for teaching codec internals.
• Opus (xiph/opus)Reference implementation of the Opus audio codec, recently adding deep-learning-based packet loss concealment.