Digital Audio Signal Synthesis with Pitch Period Correction
Solution Overview
Problem
Existing digital audio signal decoding techniques struggle to effectively correct lost blocks, particularly in predictive coders, leading to synchronization issues and unpleasant artifacts when transitory sounds are repeated, as they rely on the stationarity of the signal, which is not valid for all speech sounds.
Innovation Solution
A method is introduced to synthesize digital audio signals by determining a repetition period in valid blocks, correcting samples to limit amplitude of transitory signals, and copying these corrected samples into replacement blocks, which are then used to replace invalid blocks, thereby avoiding the repetition of plosives and adapting energy control through variable attenuation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If pitch period repetition is used to replace lost blocks, then signal continuity is maintained, but transitory sounds are incorrectly repeated causing unpleasant artifacts
Solution Approach 1:
The patent changes the parameter of signal stationarity detection by analyzing zero-crossing rates and signal variance to identify transitory sounds. When transitory characteristics are detected, the system switches from pitch period repetition to alternative reconstruction methods, thereby avoiding the repetition of unwanted transitory sounds while maintaining signal continuity for stationary segments.
Solution Approach 2:
The system dynamically adapts the block replacement strategy based on real-time analysis of signal characteristics. By continuously monitoring zero-crossing rates and variance, the system transitions between different reconstruction approaches (pitch repetition vs. alternative methods) depending on whether the current block contains transitory or stationary signals, optimizing quality while maintaining continuity.
2Ease of operation
If pitch period repetition is applied to all lost blocks, then decoding simplicity is maintained, but synchronization issues occur in predictive coders
Solution Approach 1:
The patent introduces feedback mechanisms where the analysis of zero-crossing rates and signal variance from received blocks informs the replacement strategy for lost blocks. This feedback loop allows the system to adjust its reconstruction approach based on actual signal characteristics, maintaining synchronization by adapting to the true signal structure rather than blindly repeating pitch periods.
Solution Approach 2:
The system performs preliminary analysis of signal characteristics (zero-crossing rates, variance) in valid blocks before attempting to reconstruct lost blocks. This preliminary action enables the system to pre-determine the appropriate replacement strategy, avoiding synchronization issues by preparing the correct reconstruction approach in advance based on observed signal patterns.
3Device complexity
If stationarity assumption is made for block replacement, then processing complexity is reduced, but accuracy deteriorates for non-stationary speech sounds
Solution Approach 1:
The patent introduces additional parameters (zero-crossing rate, signal variance) to detect transitory sounds and dynamically changes the processing approach based on these parameters. When transitory characteristics are detected, the system switches from simple pitch repetition to more accurate reconstruction methods, thereby maintaining low complexity for stationary signals while improving accuracy for non-stationary segments.
Solution Approach 2:
The system dynamically adjusts its processing complexity based on signal characteristics. For stationary signals, it uses simple pitch period repetition; for transitory signals detected through zero-crossing and variance analysis, it employs more sophisticated reconstruction methods. This dynamic adaptation optimizes the balance between processing complexity and replacement accuracy.
Data Source
AI summary
The present invention relates to signal modification before pitch period repetition for the synthesis of blocks lost on decoding digital audio signals. The effects of repetition of transitories, such as the plosives of a speech signal, are avoided by comparing the samples of a pitch period with those of the previous pitch period. The signal is modified preferentially by taking the minimum between a current sample (e(3)) of the last pitch period (Tj) and at least one sample (e(2−T0) of approximately the same position in the previous pitch period (Tj−1).