Overview

NVIDIA’s Video Codec SDK offers hardware-accelerated video encoding and decoding through highly optimized C/C++ APIs. Such encoding and decoding of videos is also useful for a wide range of users, including computer vision experts, researchers and Deep Learning (DL) developers. The objective of PyNvVideoCodec is to provide simple APIs for harnessing such video encoding and decoding capabilities when working with videos in Python.

PyNvVideoCodec is a library that provides Python bindings over C++ APIs for hardware-accelerated video encoding and decoding. Internally, it utilizes core APIs of NVIDIA Video Codec SDK and provides the ease-of-use inherent to Python. It relies on an external FFmpeg library for demuxing and muxing media files.

PyNvVideoCodec gives encode and decode performance (FPS) close to Video Codec SDK.

Here is a high level block diagram showing client application, PyNvVideoCodec library and related components.

Figure 1 High Level Architecture Diagram