Subband Harmonic Transposition With Reduced Audio Artifacts
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing high-frequency reconstruction (HFR) methods in audio source coding systems face challenges in achieving high-fidelity audio reproduction with reduced computational complexity, particularly for signals with prominent periodic structures, leading to artifacts like 'ghost' pitches and increased computational complexity due to the need for fine frequency resolution and oversampling.
Innovation Solution
A cross-product enhanced HFR system that processes audio signals using a subband block-based approach, involving an analysis filter bank, subband processing unit, and synthesis filter bank, which applies block extraction, nonlinear frame processing, and overlap-and-add techniques to generate time-stretched and frequency-transposed signals with reduced intermodulation products and computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If complex modulated filter banks with very fine frequency resolution and high degree of oversampling are employed, then audio quality is improved, but computational complexity becomes very high
Solution Approach 1:
The patent applies segmentation by dividing the audio signal into multiple subbands using a filter bank with coarser frequency resolution. Instead of processing the entire signal with fine resolution, each subband is processed independently with lower computational requirements, while collectively maintaining high-fidelity audio reproduction through the combination of subband outputs.
Solution Approach 2:
The patent changes the sampling parameters by employing a lower degree of oversampling compared to state-of-the-art methods. By carefully selecting and optimizing the oversampling factor for each subband, the system achieves acceptable audio quality with reduced computational complexity, trading off some oversampling for lower processing requirements.
2Speed
If harmonic transposition of order Qφ > 1 is applied to signals with prominent periodic structure, then frequency transposition is achieved, but ghost pitch and metallic character artifacts occur
Solution Approach 1:
The patent applies dynamics by adaptively selecting different transposition orders Qφ for different subbands based on the signal characteristics. Instead of using a fixed high transposition order throughout, the system dynamically adjusts the transposition order to minimize ghost pitch artifacts while maintaining effective frequency transposition, making the processing adaptive to local signal properties.
Solution Approach 2:
The patent applies local quality by allowing different subbands to have different processing characteristics, including different transposition orders and oversampling factors. This enables optimized processing for each subband's specific frequency content and signal characteristics, reducing artifacts like ghost pitch in regions where they would otherwise occur while maintaining high transposition effectiveness.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
The invention provides an efficient implementation of cross-product enhanced high-frequency reconstruction (HFR), wherein a new component at frequency QΩ + rΩ0 is generated on the basis of existing components at Ω and Ω + Ω0. The invention provides a block-based harmonic transposition, wherein a time block of complex subband samples is processed with a common phase modification. Superposition of several modified samples has the net effect of limiting undesirable intermodulation products, thereby enabling a coarser frequency resolution and/or lower degree of oversampling to be used. In one embodiment, the invention further includes a window function suitable for use with block-based cross-product enhanced HFR. A hardware embodiment of the invention may include an analysis filter bank (101), a subband processing unit (102) configurable by control data (104) and a synthesis filter bank (103).