Localized Video Encoding Using Reusable Original Frames
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video encoding methods redundantly re-encode original video data when creating localized videos, leading to disproportionately large file sizes and reduced cache efficiency in content delivery networks (CDNs), which affects the quality of experience for end-users.
Innovation Solution
A method that computes a prediction frame based on a target frame of a localized video and a reference frame of a decoded original video, generates a residual frame, and performs encoding operations on the residual frame to create an encoded localization layer, allowing reuse of original video frames where possible, thereby reducing redundant encoding.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional encoders are used to encode chunks of localized videos, then localized video data can be customized for target audiences, but original video data is redundantly re-encoded leading to disproportionately large file sizes
Solution Approach 1:
The video data is segmented into localized portions (requiring encoding) and original portions (to be reused). The patent divides the localized video into chunks, identifies which portions differ from the original video, and only encodes those differences while reusing the original portions, thereby resolving the contradiction between localization adaptability and data size.
Solution Approach 2:
Instead of re-encoding entire localized video chunks, the patent creates selective copies of only the modified portions while referencing the original encoded video data for unchanged portions. This copying approach allows localization customization without the storage overhead of complete re-encoding.
2Adaptability or versatility
If conventional encoders re-encode all localized video chunks, then complete localization customization is achieved, but cache efficiency in CDNs is disproportionately reduced
Solution Approach 1:
The patent segments the video into original and localized portions, allowing CDN caches to store the original encoded video data once while serving multiple localized versions by applying different localization layers to the same cached original data, thereby maintaining high cache efficiency.
Solution Approach 2:
The original encoded video data serves multiple functions: it is reused across different localized versions for unchanged portions. This multi-functionality allows a single cached copy to support multiple localization variants, improving cache efficiency while maintaining localization capability.
3Quantity of substance
If localized video chunks with minor differences are not encoded but use metadata instead, then storage space is reduced, but a large amount of original video data still needs to be re-encoded
Solution Approach 1:
The patent applies local quality by encoding only the specific portions of video data that actually differ between original and localized versions, rather than applying uniform re-encoding to entire chunks. This selective encoding based on local differences optimizes both storage efficiency and encoding productivity.
Solution Approach 2:
Instead of encoding all or nothing, the patent applies partial encoding action by identifying and encoding only the minimal necessary portions that contain localization changes, avoiding excessive re-encoding of unchanged data while maintaining adequate representation of modified content.
Data Source
AI summary
In various embodiments, a localized video encoding application encodes localized videos. In operation, the localized video encoding application computes a prediction frame based on a target frame of a localized video and at least a portion of a reference frame of a decoded original video. The localized video encoding application computes a residual frame based on the prediction frame and the target frame of the localized video. The localized video encoding application performs one or more encoding operations on the residual frame to generate a frame of an encoded localization layer. The frame of the encoded localization layer and at least one frame of an encoded original video are subsequently transmitted to another device for decoding.