Video Encoding Quantization Scale Factor Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing moving-picture encoding technologies, such as those described in Japanese patent applications 10-164577/1998 and 2001-238215, face challenges in maintaining picture quality during scene changes, as they often underestimate the required encoded data, leading to a drop in quality immediately after a scene change, especially for B pictures.
Innovation Solution
The proposed apparatus and method adjust the quantization scale factor based on estimated complexities calculated from actual complexities of past encoded pictures of the same or different types, ensuring adequate data encoding for scene-change-corresponding pictures by using the actual complexity of I or P pictures for estimating the complexity of subsequent P or B pictures, respectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the quantization scale factor is adjusted based on the complexity ratio between current frame and past frames, then the encoding efficiency is improved, but the picture quality drops immediately after scene change
Solution Approach 1:
The patent applies preliminary action by detecting scene changes before encoding the subsequent frame. When a scene change is detected in frame N, the system proactively increases the quantization scale factor for frame N+1 based on the complexity of frame N-1 (which is of a different picture type), rather than waiting to calculate complexity ratios using frame N. This preliminary adjustment ensures adequate picture quality is maintained immediately after scene changes while still preserving encoding efficiency through adaptive quantization.
2Quantity of substance
If the desired amount of encoded data is decided on the basis of current-frame complexity, then the data transmission rate is controlled, but the picture quality drops after scene change
Solution Approach 1:
The patent introduces an intermediary mechanism by using the complexity of a past frame of a different picture type (I or P frame) as a mediator to estimate the complexity of the next frame when a scene change is detected. Instead of directly using the current frame's complexity (which may be artificially low right after a scene change), the system references the complexity of a frame from a different picture type that better represents the actual scene complexity. This intermediary approach allows the system to decide the desired amount of encoded data more accurately, preventing picture quality drops while maintaining data rate control.
3Device complexity
If the complexity estimation uses only past frames of the same picture type, then the encoding process is simplified, but the accuracy of complexity estimation decreases during scene changes
Solution Approach 1:
The patent applies universality by making the complexity estimation process multi-functional. The system can estimate complexity using frames of the same picture type under normal conditions, but automatically switches to using frames of different picture types (I frames for estimating P frame complexity, and P frames for estimating B frame complexity) when scene changes are detected. This universal approach allows a single complexity estimation mechanism to handle both normal encoding and scene change scenarios accurately, improving measurement precision without significantly increasing encoding process complexity through conditional logic.
Data Source
AI summary
A signal related to input moving pictures of plural different types is quantized in response to a variable quantization scale factor before being encoded. An actual complexity of each past encoded picture originating from one of the input moving pictures is calculated. A complexity of a next encoded picture is estimated in response to the actual complexity of a past encoded picture same in type as the next encoded picture in cases where the next encoded picture does not correspond to a scene change. A complexity of the next encoded picture is estimated in response to the actual complexity of a past encoded picture different in type from the next encoded picture in cases where the next encoded picture corresponds to a scene change. A value of the quantization scale factor used for the quantization is decided in response to the estimated complexity.