iSize Technologies, the London-based deep-tech company specializing in deep learning for video delivery, are introducing BitSave v.2, the next-generation ‘precoder’ for perceptual optimization of video content prior to subsequent encoding.
BitSave v.2 is a pixel-to-pixel neural-network solution that optimizes video frames before the pixels reach the video compression software according to the latest advances in perceptual quality metrics. It enhances details of the areas of each frame that affect the perceptual quality score of the content after encoding and attenuates details that are not important.
This is designed to optimally balance perceptual quality against encoding bitrate. The solution is data-driven, trained on large image and video datasets, with no hand-crafted parameters.
It incurs a single-frame latency, can be used with any third-party video encoder with no changes in the encoding pipeline, and can run in real-time for up to 4K resolution.