Scalable Speech Encoding Using Characteristic Correction Filters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing scalable speech encoding schemes, such as those described in Patent Document 1, do not effectively specify the role of the enhancement layer, leading to complex configurations and inefficient compensation for components with poor coding quality in the speech signal decoded by the core layer.
Innovation Solution
A speech encoding apparatus with a first layer encoding section and a second layer encoding section that performs compensating processing on specific components of the encoded excitation signal, using a characteristic compensating filter and inverse filter to enhance the quality of the speech signal, particularly in the enhancement layer, to compensate for poor coding quality components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the enhancement layer is designed to compensate for quality in the band of 3.4 kHz or higher, then the coding performance is improved, but the configuration of the encoder becomes complicated because the role of the enhancement layer is not specified
Solution Approach 1:
The patent segments the speech signal into multiple layers (core layer and enhancement layer) with distinct frequency ranges and functional roles. The core layer handles lower frequencies while the enhancement layer specifically compensates for quality in higher frequencies (3.4 kHz or higher), allowing each layer to be optimized independently without increasing overall complexity
Solution Approach 2:
The enhancement layer is designed to provide localized quality improvement in specific frequency bands (3.4 kHz or higher) rather than uniformly improving all frequencies. This targeted approach maintains coding performance in critical bands while keeping the encoder configuration manageable through role specification
2Manufacturing precision
If the core layer is designed for speech of a band lower than 3.4 kHz, then the enhancement layer can compensate for quality in higher bands, but the overall system complexity increases due to unspecified enhancement layer roles
Solution Approach 1:
The speech encoding system is divided into core layer (handling frequencies below 3.4 kHz) and enhancement layer (handling frequencies of 3.4 kHz or higher). This segmentation allows each layer to have clearly defined roles and optimized processing, improving speech quality without unnecessarily increasing system complexity
Solution Approach 2:
The enhancement layer is designed with dynamic characteristics to adaptively compensate for quality issues in specific frequency bands. The layer can selectively enhance frequencies of 3.4 kHz or higher based on the input signal characteristics, providing flexible quality improvement while maintaining a manageable encoder structure through clear role definition
Data Source
AI summary
An audio encoding device for correcting a component having insufficient encoding capability in a core layer by an extended layer. A core layer encoder encodes an audio signal. An extended layer encoder encodes an encoding residual of the core layer encoder. A characteristic correction inverse filter arranged at a pre-stage of an LPC synthesis filter subjects the component having insufficient encoding capability in the core layer to an inverse characteristic correction process, and a characteristic correction filter arranged at a post-stage of the LPC synthesis filter performs a process for characteristic correction of the synthesis signal inputted from the LPC synthesis filter.