Video Codec SDK 13.1.15 is released

Gaming and Visualization Technologies Video Codec, PyNv & OFA
,
1

NVIDIA Video Codec SDK 13.1.15 is now available with following enhancements:

  • NVENC features support:

    • UHQ mode support for iterative encoding
    • Support application allocated CUarray as NVENC input surface
    • AV1 Hierarchical B-Frame reference mode support

  • NVDEC features support:

    • Expose QP, MB type decode stats
    • Support application allocated CUarray as NVDEC input surface
    • Support Seek to a certain frame
    • View information using MV-HEVC decode

  • Transcode features support:

    • Modular, pipeline and queue based transcode sample apps
    • Optimization of Transcode pipeline by minimizing GPU memory copies and CUDA context switches

  • Docker:

    • Dockerfile for building a complete SDK development environment
    • Pre-built samples, FFmpeg, Vulkan SDK, and test vector generation
3

Great!

Why is the SDK (and headers) behind a login form? This makes automated/full-serviced build scripts impossible and means you always need manual intervention. Isn’t it time to put these headers on e.g. Github?

4

Great update from NVIDIA. The new Video Codec SDK 13.1.15 brings some valuable improvements, especially for developers working with GPU-accelerated video workflows.

The support for application-allocated CUarrays in both NVENC and NVDEC should help reduce unnecessary memory transfers and improve efficiency. The AV1 hierarchical B-frame support is also a nice addition for better compression performance and quality.

The new transcode pipeline optimizations, along with the Docker development environment, should make it easier for developers to build, test, and optimize video processing applications across different environments.

Looking forward to seeing how these enhancements improve real-world encoding, decoding, and streaming workloads.