Spectral Domain Bass Filtering for Low Delay Audio Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The low pass filter in the bass post-processing of decoded audio signals introduces significant delay in audio encoding and decoding schemes, particularly in low delay applications, which is problematic for maintaining signal timing alignment across branches.
Innovation Solution
The proposed solution involves filtering the decoded audio signal in the spectral domain, such as the QMF or MDCT domain, rather than the time domain, to reduce delay. This approach uses frequency-selective weighting of subband signals without incurring systematic delay, and subsequent subband-wise subtraction is performed in the spectral domain, followed by conversion back to the time domain only when necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If low pass filtering is performed in the time domain for bass post-processing, then frequency selectivity is achieved, but significant delay is introduced
Solution Approach 1:
The patent replaces time-domain mechanical filtering with spectral-domain processing. Instead of using a time-domain low pass filter that introduces delay, the invention transforms the signal to the spectral domain (MDCT or QMF), applies frequency-selective weighting to subband signals, and transforms back. This substitution eliminates systematic filter delay while achieving the same frequency selectivity effect.
Solution Approach 2:
The patent changes the domain parameter from time domain to spectral domain. By performing filtering operations in the spectral domain rather than the time domain, the system achieves frequency selectivity through weighting coefficients applied to spectral subbands, avoiding the inherent delay of time-domain filtering while maintaining the desired frequency response characteristics.
2Ease of manufacture
If time-domain filtering is used for bass post-processing, then implementation is straightforward, but delay accumulates in the encoding/decoding chain
Solution Approach 1:
The patent merges the bass post-processing filtering operation with the existing time-frequency transformation steps already present in the audio decoding chain. By combining the filter implementation with the MDCT/QMF transform that is already required for other processing, the system avoids adding separate filtering stages and their associated delays, achieving frequency selectivity without increasing overall system delay.
3Measurement precision
If frequency-selective filtering is applied in the time domain, then spectral shaping is achieved, but signal timing alignment across branches is compromised
Solution Approach 1:
The patent substitutes time-domain filtering with spectral-domain weighting to maintain signal timing alignment. By applying frequency-selective weighting in the spectral domain rather than through time-domain filtering, the invention achieves spectral shaping without introducing the phase distortion and timing misalignment that characterize time-domain filter operations, thereby maintaining reliable synchronization across parallel signal branches.
Data Source
AI summary
An apparatus for processing a decoded audio signal including a filter for filtering the decoded audio signal to obtain a filtered audio signal, a time-spectral converter stage for converting the decoded audio signal and the filtered audio signal into corresponding spectral representations, each spectral representation having a plurality of subband signals, a weighter for performing a frequency selective weighting of the filtered audio signal by a multiplying subband signals by respective weighting coefficients to obtain a weighted filtered audio signal, a subtractor for performing a subband-wise subtraction between the weighted filtered audio signal and the spectral representation of the decoded audio signal, and a spectral-time converter for converting the result audio signal or a signal derived from the result audio signal into a time domain representation to obtain a processed decoded audio signal.