Subband Prediction in Critically Sampled Filterbanks With Lower Aliasing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio source coding systems face challenges in efficiently encoding subband predictors at low bit rates while minimizing noise shaping artifacts and aliasing effects, particularly in critically sampled filterbanks used for high-quality audio coding.
Innovation Solution
Implementing a model-based prediction method in the subband domain using signal models to describe subband predictors, which allows for compact parameterization and reduces computational complexity, and employing cross-subband predictor terms to minimize aliasing artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If subband prediction is used in critically sampled filterbanks, then audio coding quality is improved, but aliasing artifacts appear in the output signal
Solution Approach 1:
The patent converts the harmful aliasing artifacts into beneficial information by using them for noise shaping. The prediction error signal, which contains aliasing artifacts, is deliberately shaped to place quantization noise in less perceptible frequency regions, thereby improving perceived audio quality while maintaining bit efficiency.
Solution Approach 2:
The patent changes the parameters of the prediction error signal by applying noise shaping filters that modify the spectral distribution of quantization noise. This allows control over where aliasing artifacts appear in the frequency spectrum, moving them to less perceptible regions.
2Quantity of substance
If backwards adaptive prediction scheme is used to encode subband predictors, then bit rate is reduced, but prediction accuracy degrades at low bit rates
Solution Approach 1:
The patent introduces an intermediate parametric representation that bridges the gap between encoder and decoder. Instead of directly transmitting predictor coefficients or relying solely on backwards adaptation, the system uses compact parametric models (such as sinusoidal models or autoregressive models) to describe the prediction process, allowing accurate reconstruction at low bit rates.
3Object-generated harmful factors
If long term prediction is implemented by converting time domain delay to subband predictor, then alias artifacts are avoided, but computational complexity increases significantly
Solution Approach 1:
The patent replaces the mechanical process of time-domain delay conversion with filterbank operations with a more efficient parametric approach. By using parametric models that directly operate in the subband domain, the system avoids the computationally intensive conversion process while achieving similar or better performance.
4Manufacturing precision
If subband prediction is applied, then noise shaping control is improved within subbands, but aliasing artifacts appear in the final output signal
Solution Approach 1:
The patent addresses the limitation of subband prediction by introducing cross-subband prediction terms. This adds a new dimension to the prediction process, allowing the system to control noise shaping not only within individual subbands but also across subband boundaries, thereby reducing aliasing artifacts in the reconstructed signal.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6a
AI summary
The present document relates to audio source coding systems. In particular, the present document relates to audio source coding systems which make use of linear prediction in combination with a filterbank. A method for estimating a first sample (615) of a first subband signal in a first subband of an audio signal is described. The first subband signal of the audio signal is determined using an analysis filterbank (612) comprising a plurality of analysis filters which provide a plurality of subband signals in a plurality of subbands from the audio signal, respectively. The method comprises determining a model parameter (613) of a signal model; determining a prediction coefficient to be applied to a previous sample (614) of a first decoded subband signals derived from the first subband signal, based on the signal model, based on the model parameter (613) and based on the analysis filterbank (612); wherein a time slot of the previous sample (614) is prior to a time slot of the first sample (615); and determining an estimate of the first sample (615) by applying the prediction coefficient to the previous sample (614).