Audio Encoding Using Adaptive Noise Shaping Paths
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Transform-based audio coding techniques face issues with pre-echo and time domain aliasing due to quantization noise, particularly when using modulated complex lapped transform (MCLT) and time noise shaping (TNS), which can lead to artifacts in audio quality.
Innovation Solution
The method involves performing frequency domain noise shaping (FDNS) on MCLT-generated signals and selectively applying complex temporal noise shaping (CTNS) based on prediction gain, with signal processing paths that include or exclude noise shaping operations, to generate and decode audio signals effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Object-affected harmful factors
If time noise shaping (TNS) is performed on MCLT coefficients to address pre-echo, then pre-echo is reduced, but time domain aliasing (TDA) is caused
Solution Approach 1:
The patent changes the domain parameter from time domain noise shaping to frequency domain noise shaping. By performing noise shaping in the frequency domain (FDNS) rather than the time domain (TNS), the patent eliminates time domain aliasing while still addressing pre-echo through frequency-based noise redistribution.
Solution Approach 2:
The patent substitutes the time-domain noise shaping mechanism with a frequency-domain noise shaping mechanism. This replacement allows noise to be shaped and redistributed in the frequency spectrum rather than in the time domain, avoiding the aliasing artifacts that occur with time-domain processing.
2Object-affected harmful factors
If frequency domain noise shaping (FDNS) is performed on MCLT signals, then pre-echo is reduced, but complex temporal noise shaping (CTNS) may be required to fully address artifacts
Solution Approach 1:
The patent implements dynamic selection between different noise shaping paths based on prediction gain thresholds. The system adapts its complexity by choosing between a simple FDNS path and a more complex CTNS path depending on the signal characteristics, thereby optimizing the balance between artifact reduction and processing complexity.
Solution Approach 2:
The patent segments the signal processing into distinct paths: a first path for signals with prediction gain below a threshold (using only FDNS) and a second path for signals with prediction gain above the threshold (using both FDNS and CTNS). This segmentation allows the system to apply complex processing only when necessary.
3Reliability
If multiple signal processing paths are implemented for adaptive noise shaping, then audio quality is improved, but encoding and decoding complexity increases
Solution Approach 1:
The patent uses prediction gain as a feedback parameter to dynamically select between different noise shaping paths. The prediction gain calculation provides feedback about the signal characteristics, enabling the system to adaptively choose the most appropriate processing path and achieve better audio quality while managing complexity through intelligent decision-making.
Data Source
AI summary
A method of encoding/decoding an audio signal and a device for performing the same are provided. The method of encoding an audio signal includes generating, based on the audio signal, a linear prediction coding (LPC) bitstream and a frequency-domain signal of the audio signal, generating, based on the LPC bitstream and the frequency-domain signal, a first residual signal including information on a frequency envelope of the frequency-domain signal, and outputting a second residual signal by processing a first residual signal through one of a plurality of signal processing paths.