SBR Temporal Grid Encoding for Transient Pre-Echo Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio encoding methods using SBR technique for low bit rates result in audible artefacts, particularly at transients, due to coarse temporal grids leading to pre-echo issues, while fine grids increase bit consumption.
Innovation Solution
Introduce a new SBR frame class (LD_TRAN) where grid boundaries align with frame boundaries, using a transient position indication to define a short grid area around transients, reducing the need for variable frame classes and minimizing bit transmission.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If a coarse temporal grid is used for spectral envelope encoding, then bit consumption is reduced, but audible artefacts and pre-echo occur at transients
Solution Approach 1:
The temporal grid is segmented into two types: regular grid areas and transient grid areas. Transient grid areas are identified by detecting transient positions and creating shorter grid segments around them, while regular areas use the standard coarse grid. This segmentation allows fine resolution only where needed (at transients) while maintaining coarse resolution elsewhere, thus reducing overall bit consumption while avoiding pre-echo artefacts.
Solution Approach 2:
Different temporal resolutions are applied locally depending on the signal characteristics. In transient grid areas, a fine temporal grid is used to accurately capture transient events, while in regular grid areas, a coarse temporal grid is used to minimize bit consumption. This local adaptation of grid fineness resolves the contradiction by applying fine resolution only where it is necessary to avoid artefacts.
2Measurement precision
If a fine temporal grid is used for spectral envelope encoding, then transient representation is improved, but bit consumption increases
Solution Approach 1:
The temporal grid is segmented into two types: regular grid areas and transient grid areas. Transient grid areas are identified by detecting transient positions and creating shorter grid segments around them, while regular areas use the standard coarse grid. This segmentation allows fine resolution only where needed (at transients) while maintaining coarse resolution elsewhere, thus reducing overall bit consumption while avoiding pre-echo artefacts.
Solution Approach 2:
Different temporal resolutions are applied locally depending on the signal characteristics. In transient grid areas, a fine temporal grid is used to accurately capture transient events, while in regular grid areas, a coarse temporal grid is used to minimize bit consumption. This local adaptation of grid fineness resolves the contradiction by applying fine resolution only where it is necessary to avoid artefacts.
3Measurement precision
If variable frame classes are used to handle transients, then transient processing is improved, but decoding delay increases
Solution Approach 1:
The grid structure is made dynamic by allowing transient grid areas to have different boundaries than regular grid areas. The transient grid areas are defined relative to the detected transient position, allowing the grid to adapt its structure dynamically based on signal content. This dynamic approach improves transient processing while maintaining compatibility with the standard frame structure, thus reducing decoding delay.
Solution Approach 2:
The temporal resolution parameter is changed locally around transient positions. By detecting transient positions and creating shorter grid segments around them, the effective temporal resolution is increased at critical moments without changing the overall frame structure or requiring variable frame classes. This parameter adaptation resolves the contradiction by improving transient accuracy without introducing additional decoding delay.
Data Source
AI summary
The transient problem may be sufficiently addressed, and for this purpose, a further delay on the side of the decoding may be reduced if a new SBR frame class is used wherein the frame boundaries are not shifted, i.e. the grid boundaries are still synchronized with the frame boundaries, but wherein a transient position indication is additionally used as a syntax element so as to be used, on the encoder and/or decoder sides, within the frames of these new frame class for determining the grid boundaries within these frames.


