Video Decoder Film Grain Reconstruction for Efficient Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video codecs often lose or degrade film grain during encoding and decoding processes, making it difficult to compress videos with film grain effectively.
Innovation Solution
A decoder reconstructs film grain or noise dynamically by generating representative luma values and chroma film grain noise pixels, which are then applied to the reconstructed video frame, allowing for efficient compression and retention of aesthetic noise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If film grain is retained through encoding and decoding processes, then video quality and aesthetic effect are improved, but compression efficiency deteriorates due to the randomness of film grain
Solution Approach 1:
The patent extracts film grain as a separate component from the main video content during encoding. The encoder identifies and separates grain pixels from video pixels, storing grain data independently. This allows the main video to be compressed more efficiently while grain is preserved and recombined during decoding, resolving the contradiction between quality and compression efficiency.
Solution Approach 2:
The patent segments the video signal into distinct components: video pixels and grain pixels. By dividing the processing into separate streams where video data and grain data are handled independently, the system achieves better compression ratios for the video portion while maintaining grain integrity, thus improving both compression efficiency and video quality.
2Productivity
If film grain is removed during encoding to improve compression, then compression efficiency is improved, but video quality and aesthetic effect deteriorate
Solution Approach 1:
The patent performs preliminary separation of grain from video content during the encoding phase. By identifying and extracting grain pixels before the main compression process, the system prepares the data in advance so that grain can be preserved while the video portion undergoes aggressive compression, thus achieving both high compression efficiency and quality retention.
Solution Approach 2:
The patent temporarily discards grain information from the main video stream during encoding, compressing only the video portion for efficient storage and transmission. During decoding, the grain information is recovered and recombined with the decompressed video, thus achieving efficient compression while recovering the aesthetic quality.
3Speed
If traditional video coding techniques are used, then decoding speed is maintained, but film grain is lost or degraded in the encoding and decoding processes
Solution Approach 1:
The patent segments the decoding process into separate streams for video data and grain data. By processing these independently and recombining them at the end, the system maintains efficient decoding speeds for the video portion while preserving grain information that would otherwise be lost in traditional integrated decoding processes.
Solution Approach 2:
The patent implements a dynamic decoding approach where grain parameters are reconstructed based on the decoded video content. This allows the system to adaptively regenerate grain characteristics during decoding, maintaining both speed and grain retention by performing grain synthesis parallel to or after main video decoding.
Data Source
AI summary
Systems and methods for decoding are described herein. For example, a process can include obtaining luma values associated with a reconstructed video frame. The process can further include generating representative luma values based on the luma values associated with the reconstructed video frame. The process can include storing the representative luma values associated with the reconstructed video frame in a memory. The process can further include obtaining noise pixel data associated with the reconstructed video frame. The process can further include generating, using the representative luma values and the noise pixel data, chroma film grain noise pixels associated with the reconstructed video frame. The process can further include applying the chroma film grain noise pixels to the reconstructed video frame.