Frequency-Domain Audio Coding With Interleaved Transform Length Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing frequency-domain audio coding systems face challenges in achieving satisfactory quality for audio signals with dense transients, such as rain or applause, at low bitrates, due to issues like time-smearing and increased data overhead when using either long or short transform blocks.
Innovation Solution
A frequency-domain audio codec is enhanced to support additional transform lengths while maintaining backward compatibility, allowing for interleaved transmission of frequency-domain coefficients and independent operation of coefficient and scale factor extraction, enabling responsive switching between transform lengths.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If long transform blocks are used for coding audio frames, then spectral resolution is improved, but time-smearing of coding error (pre-echo) occurs frequently for signals with dense transients
Solution Approach 1:
The system dynamically switches between long and short transform blocks based on the temporal stationarity of the signal. For signals with dense transients, short blocks are used to reduce pre-echo, while for stationary signals, long blocks are used to improve spectral resolution. This dynamic adaptation resolves the contradiction by adjusting the transform block length according to signal characteristics.
Solution Approach 2:
The transform block length parameter is changed based on signal analysis. The system monitors signal properties and adjusts the block length parameter to optimize performance for different signal types, thereby resolving the trade-off between spectral resolution and pre-echo artifacts.
2Object-generated harmful factors
If short transform blocks are used for coding audio frames, then time-smearing is reduced, but data overhead increases leading to spectral holes
Solution Approach 1:
The system uses short transform blocks only when necessary (for transient signals) and switches to long blocks for stationary signals, thereby minimizing data overhead while still reducing pre-echo where needed. This dynamic approach avoids the continuous overhead penalty of using short blocks for all signals.
Solution Approach 2:
The audio signal is segmented into frames, and each frame is further divided into sub-frames for transform coding. This segmentation allows the system to apply short transforms only to specific portions of the signal where transients occur, rather than applying them uniformly, thus reducing overall data overhead.
3Measurement precision
If a new frequency-domain audio codec supporting additional transform lengths is developed, then coding quality for specific signal types is improved, but backward compatibility with existing coders and decoders is lost
Solution Approach 1:
The system is designed to support multiple transform block lengths (long, short, and intermediate) within a single unified framework. This multi-functionality allows the codec to adapt to different signal types while maintaining compatibility with existing decoders that expect standard long or short blocks, as the intermediate blocks can be processed using existing decoding paths.
Solution Approach 2:
The new transform length capability is nested within the existing codec framework. The system extends the existing transform block length options rather than creating a completely new system, allowing old decoders to function with the extended capabilities while new decoders can fully exploit the additional transform length options for improved quality.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
A frequency-domain audio codec is provided with the ability to additionally support a certain transform length in a backward-compatible manner, by the following: the frequency-domain coefficients of a respective frame are transmitted in an interleaved manner irrespective of the signalization signaling for the frames as to which transform length actually applies, and additionally the frequency-domain coefficient extraction and the scale factor extraction operate independent from the signalization. By this measure, old-fashioned frequency-domain audio coders/decoders, insensitive for the signalization, would be able to nevertheless operate without faults and with reproducing a reasonable quality. Concurrently, frequency-domain audio coders/decoders able to support the additional transform length would offer even better quality despite the backward compatibility. As far as coding efficiency penalties due to the coding of the frequency domain coefficients in a manner transparent for older decoders are concerned, same are of comparatively minor nature due to the interleaving.