Audio Signal Processing Shape Prediction Spectral Hole Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
When both audio and speech properties coexist in a signal, existing coding schemes lead to degraded audio coding efficiency and sound quality, particularly due to increased quantization errors and spectral holes at low bit rates.
Innovation Solution
An apparatus and method that apply one of two coding schemes to a frame, using shape prediction and psychoacoustic models to compensate for spectral holes by determining the most suitable prediction mode and generating perceptual gain values based on correlation, thereby substituting spectral holes with previous or current frame data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If spectral coefficients are quantized at low bit rate, then data transmission efficiency is improved, but quantization error increases causing spectral holes and degraded sound quality
Solution Approach 1:
The patent creates a copy of the spectral shape from adjacent frequency regions or temporal neighborhoods to fill spectral holes. Instead of transmitting all spectral coefficients, the system identifies spectral holes (where coefficients are zero or near-zero after quantization) and synthesizes replacement values by copying shape characteristics from other parts of the signal, thereby maintaining sound quality without increasing bit rate.
Solution Approach 2:
The patent introduces an intermediary process between quantization and reconstruction that detects spectral holes and applies shape prediction algorithms. This intermediary step analyzes the spectral shape characteristics and uses predictive models to generate plausible spectral values, acting as a mediator that bridges the gap caused by aggressive quantization.
2Device complexity
If one coding scheme is applied to signals with both audio and speech properties, then coding complexity is reduced, but audio coding efficiency and sound quality are degraded
Solution Approach 1:
The patent implements a dynamic coding approach where the system adapts its processing based on the detected signal characteristics. By analyzing whether the current frame contains audio or speech properties, the system dynamically selects appropriate processing strategies, including whether to apply shape prediction and spectral hole filling, thereby optimizing coding efficiency without excessive complexity.
Solution Approach 2:
The patent applies different processing strategies to different parts of the signal based on local characteristics. Instead of uniformly processing all spectral coefficients, the system identifies specific regions with audio properties that would benefit from shape prediction and spectral hole filling, applying these techniques locally only where needed rather than globally to the entire signal.
3Device complexity
If spectral holes are not compensated, then decoding complexity is reduced, but sound quality is significantly degraded
Solution Approach 1:
The patent extracts and identifies spectral holes from the quantized spectral coefficients by detecting regions where coefficients are zero or near-zero. This extraction process separates the problematic spectral holes from the rest of the spectral data, allowing targeted compensation only in the identified hole regions rather than processing the entire spectrum.
Solution Approach 2:
The patent performs preliminary detection and identification of spectral holes before the final spectral reconstruction. By pre-identifying which spectral coefficients need compensation and preparing shape prediction models in advance, the system streamlines the decoding process and reduces overall complexity despite adding the spectral hole filling functionality.
Data Source
AI summary
A method of processing an audio signal is disclosed. The present invention includes a method for processing an audio signal, comprising: receiving, by an audio processing apparatus, the spectral data including a current block, and substitution type information indicating whether to apply a shape prediction scheme to a current block; when the substitution type information indicates that the shape prediction scheme is applied to the current block, receiving lag information indicating an interval between spectral coefficients of the current block and the predictive shape vector of a current frame or a previous frame; obtaining spectral coefficients by substituting for spectral hole included in the current block using the predictive shape vector.