Subband Harmonic Transposition for Low-Artifact Audio HFR
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 overhead.
Innovation Solution
The implementation of cross-product enhanced HFR using subband block-based processing, which involves an analysis filter bank, subband processing unit, and synthesis filter bank, allows for reduced computational effort by applying frame-wise phase offsets and magnitude adjustments, enabling efficient time stretching and frequency transposition with coarser frequency resolution and lower oversampling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If complex modulated filter banks with fine frequency resolution and high oversampling are used for high quality harmonic HFR, then audio quality is improved, but computational complexity becomes very high
Solution Approach 1:
The patent divides the audio signal into multiple subbands using a filter bank, processing each subband separately. This segmentation allows coarser frequency resolution within each subband while maintaining overall high audio quality, significantly reducing computational complexity compared to processing the entire spectrum with fine resolution.
Solution Approach 2:
The patent applies different processing strategies to different subbands. By identifying periodic structures in specific subbands and applying cross-product enhancement selectively, the system maintains high audio quality where needed while using simpler processing in other regions, optimizing the balance between quality and computational load.
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 introduces cross-product terms as intermediary elements that mediate between the original harmonic components and the transposed output. By computing products of subband signals at different frequency offsets and adding these cross-products to the direct transposition path, the system regresses missing harmonic partials and eliminates ghost pitch artifacts while maintaining the desired frequency transposition effect.
Solution Approach 2:
The patent combines multiple processing paths (direct transposition and cross-product enhancement) to create a composite transposition system. This composite approach integrates the simplicity of direct transposition with the artifact-reduction capabilities of cross-product methods, achieving high-quality frequency transposition without ghost pitch.
3Loss of substance
If single sideband modulation based HFR is used with low bandwidth core signal, then bitrate is reduced, but dissonant ringing artifact results
Solution Approach 1:
The patent modifies the transposition approach by using subband-specific frequency offsets instead of a fixed global offset. By adapting the frequency shift parameter to each subband's characteristics and using cross-product enhancement, the system maintains efficient bitrate usage while eliminating the dissonant ringing artifacts that result from uniform frequency shifting across all frequencies.
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).