Multichannel Audio Upmixing with IPD-Based Phase Reconstruction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing signal processing methods, such as those using decorrelators, fail to precisely reproduce phase or delay differences between channel signals and cannot restore inter-channel phase differences, leading to difficulties in sound image localization and reverberation reconstruction.
Innovation Solution
A signal processing apparatus and method that shifts the phase of decoded audio or speech signals based on phase shift flags, utilizes inter-channel phase difference (IPD) coding and mode flags, and modifies IPD values to enhance sound quality, allowing for precise reproduction of phase differences and improved sound image localization and reverberation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a decorrelator is used to generate a stereo signal from a mono signal, then the signal can be processed, but the phase or delay difference between channel signals cannot be precisely reproduced
Solution Approach 1:
The patent changes the parameter representation from time-domain phase/delay differences to frequency-domain inter-channel phase difference (IPD) values. By computing IPD in the frequency domain and using it to modulate the stereo signal, the system achieves precise phase reproduction without complex time-domain processing.
Solution Approach 2:
The patent replaces the mechanical decorrelator approach with a frequency-domain IPD-based processing system. Instead of using complex time-domain filtering and delay networks, the invention uses spectral analysis and phase modulation in the frequency domain to achieve the same stereo effect with higher precision.
2Loss of information
If inter-channel level difference value and inter-channel correlation value are used for coding, then the signal can be compressed, but inter-channel phase difference cannot be restored and reflected
Solution Approach 1:
The patent performs preliminary computation of inter-channel phase difference values during the encoding phase and stores them in the bitstream. This preliminary action ensures that the phase difference information is preserved and can be directly used during decoding without requiring complex real-time computation, thus preventing information loss while maintaining coding efficiency.
Solution Approach 2:
The patent adds a new dimension to the spatial coding parameters by introducing frequency-domain IPD values alongside the traditional time-domain ILD and ICC parameters. This multi-dimensional approach allows comprehensive representation of spatial audio characteristics including phase differences that cannot be captured by level and correlation alone.
3Measurement precision
If inter-channel phase difference value is transmitted for every frame, then precise sound image localization can be achieved, but coding efficiency decreases
Solution Approach 1:
The patent applies partial action by selectively transmitting IPD values only when necessary. The system determines whether to include IPD information in each frame based on whether the phase difference exceeds a threshold or changes significantly from the previous frame, thus achieving precise localization only when needed while maintaining high coding efficiency for static or simple scenes.
Solution Approach 2:
The patent implements dynamic IPD transmission where the decision to transmit phase difference information varies over time based on signal characteristics. The system adapts the transmission strategy frame-by-frame, transmitting detailed phase information during dynamic scenes requiring precise localization while using coarser representations during static periods, optimizing the trade-off between precision and efficiency.
Data Source
AI summary
A method of processing a signal is disclosed. The present invention includes receiving (a) downmix signal being generated from plural-channel signal and (b) spatial information indicating attribute of the plural-channel signal in order to upmix the downmix signal and including phase shift flag indicating whether phase of a frame of at least one channel of the plural-channel signal is shifted; obtaining inter-channel phase difference (IPD) coding flag indicating whether IPD value is used to the spatial information from a header of the spatial information; obtaining IPD mode flag indicating whether the IPD value is used to frame of the spatial information from the frame based on the IPD coding flag; obtaining the IPD value of parameter band in the frame, based on the IPD mode flag; upmixing plural-channel signal by applying the IPD value to the downmix signal; and shifting the phase of the frame of the at least one channel of the plural-channel signal based on the phase shift flag.