NVIDIA Launches Open-Source Video Processing Framework For Python
NVIDIA's "VideoProcessingFramework" is an open-source set of C++ libraries that are wrapped around by Python bindings for interacting with their closed-source Video Codec SDK. The function of this framework is to make it easy to exploit GPU-accelerated video encode/decode from Python.
While Python isn't the most performant language, it's easy-to-use and with its mass following NVIDIA has decided to publish this Video Processing Framework that amounts to a Python wrapper around their existing Video Codec SDK C++ stack for GPU-based video encode/decode on Kepler and newer.
So you end up with easy-to-read/write code while exploiting high-performance GPU-based video acceleration.
The code is available as NVIDIA's VideoProcessingFramework on GitHub.
In announcing VPF on their developer blog they provided a brief Python code example of using the PyNvCodec module for showing video transcoding in Python.
While Python isn't the most performant language, it's easy-to-use and with its mass following NVIDIA has decided to publish this Video Processing Framework that amounts to a Python wrapper around their existing Video Codec SDK C++ stack for GPU-based video encode/decode on Kepler and newer.
So you end up with easy-to-read/write code while exploiting high-performance GPU-based video acceleration.
The code is available as NVIDIA's VideoProcessingFramework on GitHub.
In announcing VPF on their developer blog they provided a brief Python code example of using the PyNvCodec module for showing video transcoding in Python.
26 Comments