Subband Harmonic Transposition for Low-Artifact Audio Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing harmonic transposition methods in audio coding suffer from high computational complexity and generate unwanted artifacts such as ghost pitches and metallic sounds, particularly for signals with prominent periodic structures, while also requiring complex filter banks and oversampling.
Innovation Solution
Implementing a cross-product enhanced harmonic transposition method using subband block processing with coarse frequency resolution and reduced oversampling, applying frame-wise phase offsets and magnitude adjustments to suppress intermodulation products, allowing for efficient generation of time-stretched and frequency-transposed signals.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If complex modulated filter banks with fine frequency resolution and high oversampling are used, then audio quality is improved, but computational complexity becomes very high
Solution Approach 1:
The audio signal is divided into multiple subbands using a filter bank with coarser frequency resolution. Each subband is processed independently with block-based harmonic transposition, allowing the system to achieve high audio quality through distributed processing rather than requiring a single complex high-resolution filter bank.
Solution Approach 2:
The invention changes the oversampling parameter from high to reduced levels, and compensates for the resulting quality loss through block-based processing with frame-wise phase offsets and magnitude adjustments. This parameter change reduces computational complexity while maintaining audio quality through alternative processing mechanisms.
2Device complexity
If traditional harmonic transposition is applied to signals with prominent periodic structure, then processing is simplified, but ghost pitch and metallic artifacts are generated
Solution Approach 1:
Frame-wise phase offsets are applied to subband blocks before transposition to pre-compensate for the ghost pitch effect. By adjusting the phase of each block independently, the system counteracts the periodic artifacts that would otherwise be generated by traditional harmonic transposition of periodic signals.
Solution Approach 2:
The system applies dynamic phase offsets and magnitude adjustments to each subband block individually, allowing the processing characteristics to adapt to the local signal properties. This dynamic approach enables suppression of artifacts in periodic regions while maintaining natural sound in other regions.
3Object-generated harmful factors
If subband block processing with frame-wise phase offsets is applied, then intermodulation products are suppressed, but processing steps increase
Solution Approach 1:
Multiple processing operations (phase offset application, magnitude adjustment, block transposition) are merged into a unified block-based processing framework. By combining these operations at the block level rather than applying them separately to individual samples, the system suppresses intermodulation products while minimizing the increase in processing complexity.
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 S2 + 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).