Multi-Overlap Window Switching for Low-Delay Audio Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio and image signal processing technologies face challenges in efficiently encoding and decoding signals with transients, leading to pre-echo noise and inefficiencies in low-delay communication applications.
Innovation Solution
The proposed solution involves an apparatus and method for generating and decoding audio or image signals using a window sequence controller to generate overlapping window functions, a preprocessor for folding-in operations, and a processor for spectral value processing, allowing for adaptive overlap widths and transform lengths based on transient locations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If asymmetric transition windows are used for block switching between long and short transforms, then adaptability to non-stationary signal characteristics is improved, but device complexity and look-ahead requirements increase
Solution Approach 1:
The patent extracts and removes the asymmetric transition windows from the encoding process. Instead of using complex start and stop windows for block switching, the invention directly switches between long and short transforms without requiring transition windows, thereby simplifying the device while maintaining adaptability to non-stationary signals
Solution Approach 2:
The patent inverts the conventional approach by eliminating the need for transition windows entirely. Rather than using asymmetric windows to facilitate block switching, the invention achieves block switching directly by selecting between long and short transforms based on signal characteristics, thus reducing complexity while preserving adaptability
2Productivity
If look-ahead is used to determine transform switching, then coding efficiency is improved, but delay increases which is unacceptable in low-delay communication applications
Solution Approach 1:
The patent enables the encoding system to determine transform length without requiring look-ahead by using the current frame's signal characteristics alone. The encoder analyzes the current frame's non-stationarity and autonomously decides whether to use long or short transforms, eliminating the need for future signal knowledge and reducing delay while maintaining coding efficiency
Solution Approach 2:
The patent performs signal analysis and transform length determination within the current frame processing without requiring preliminary analysis of future frames. By making encoding decisions based on current frame characteristics alone, the system eliminates look-ahead delay while maintaining the ability to adapt to non-stationary signals
3Device complexity
If reduced overlap is used to avoid asymmetric transition windows, then device complexity is reduced, but coding precision and audio quality deteriorate
Solution Approach 1:
The patent changes the overlap parameter from reduced overlap to full overlap (50% of frame length) when using short transforms. This full overlap ensures smooth transitions and maintains coding precision by providing sufficient overlapping region for proper signal reconstruction, while the device complexity remains low because asymmetric transition windows are eliminated
4Reliability
If Temporal Noise Shaping is used to minimize temporal spread of coding error, then audio quality is improved, but device complexity increases
Solution Approach 1:
The patent extracts and removes the Temporal Noise Shaping (TNS) component from the encoding system. By eliminating TNS, the invention reduces device complexity while maintaining acceptable audio quality through the use of full overlap and direct block switching without asymmetric transition windows
Data Source
AI summary
An apparatus for generating an encoded signal includes: a window sequence controller for generating a window sequence information for windowing an audio or image signal, the window sequence information indicating a first window for generating a first frame of spectral values, a second window function and at least one third window function for generating a second frame of spectral values, wherein the first window function, the second window function and the one or more third window functions overlap within a multi-overlap region; a preprocessor for windowing a second block of samples corresponding to the second window function and the at least one third window functions using an auxiliary window function to acquire a second block of windowed samples, a spectrum converter for applying an aliasing-introducing transform; and a processor for processing the first frame and the second frame to acquire encoded frames of the audio or image signal.


