Audio Signal Bandwidth Extension With Temporal Spreading And Decimation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio signal bandwidth extension methods face challenges with high computational complexity and quality loss, particularly at low bitrates, leading to artifacts like roughness and unpleasant timber due to unharmonized tonal signal portions.
Innovation Solution
A method involving temporal signal spreading and decimation, using a phase vocoder to generate a time signal spread by a factor greater than 1, followed by decimation and bandpass filtering to create a high-frequency signal portion, without requiring complex analysis/synthesis filterbanks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If harmonic transposition with filterbank patching is used for bandwidth extension, then audio quality is improved, but computational complexity increases significantly
Solution Approach 1:
The patent extracts only the essential spectral envelope information from the lower band signal and transposes it to the upper band, omitting the complex filterbank patching operations. This extraction approach maintains audio quality by preserving the spectral characteristics while significantly reducing computational complexity by eliminating the need for multiple filterbank analyses and synthesizes.
Solution Approach 2:
Instead of analyzing the signal into filterbanks, patching channels, and synthesizing (the conventional approach), the patent inverts the process by directly transposing the time-domain lower band signal to generate the upper band signal. This inversion eliminates the complex intermediate filtering and patching stages while achieving similar audio quality results.
2Device complexity
If copying function of low-frequency signal portions into high frequency range is used, then computational complexity is reduced, but audio quality deteriorates due to artifacts like roughness and unpleasant timber
Solution Approach 1:
The patent applies parameter changes by using spectral envelope information and harmonic relationships to guide the transposition process. By adjusting the spectral envelope parameters and harmonic structure during the frequency transposition, the method maintains natural timbre and avoids the roughness artifacts associated with simple copying, while keeping computational complexity low.
Solution Approach 2:
The patent introduces spectral envelope information as an intermediary between the simple copying process and the complex filterbank method. This intermediary guides the transposition to preserve harmonic relationships and spectral characteristics, improving audio quality without requiring the full complexity of filterbank patching operations.
3Device complexity
If simple copying of bandpass signals into successive filterbank channels is used, then device complexity is reduced, but manufacturing precision decreases due to coarse approximation
Solution Approach 1:
The patent changes the parameters by using spectral envelope information to guide the transposition process, transforming the simple copying operation into a more accurate spectral reconstruction. This approach improves spectral accuracy by ensuring that the transposed signal matches the original spectral characteristics, while maintaining low device complexity by avoiding complex filterbank operations.
Data Source
Figure 1
Figure 2A
Figure 2B
AI summary
For a bandwidth extension of an audio signal, in a signal spreader the audio signal is temporally spread by a spread factor greater than 1. The temporally spread audio signal is then supplied to a decimator to decimate the temporally spread version by a decimation factor matched to the spread factor. The band generated by this decimation operation is extracted and distorted, and finally combined with the audio signal to obtain a bandwidth extended audio signal. A phase vocoder in the filterbank implementation or transformation implementation may be used for signal spreading.