Speech Processing Device Band Group Delay Compensation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional speech processing technologies face challenges in reproducing speech waveforms due to deviations between reconstructed and analysis source phase, and difficulties in rapid waveform generation using group delay features.
Innovation Solution
A speech processing device and method that include a spectrum parameter calculation unit, phase spectrum calculation unit, group delay spectrum calculation unit, band group delay parameter calculation unit, and band group delay compensation parameter calculation unit to enhance reproducibility by calculating and compensating phase spectra and group delay parameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional speech processing methods are used, then the basic speech analysis and synthesis functions are achieved, but the reproducibility of speech waveform is poor due to phase deviation
Solution Approach 1:
The invention introduces band group delay compensation parameters to adjust and optimize the phase spectrum reconstruction process. By calculating compensation parameters for different frequency bands and applying them to correct phase deviations, the system achieves more accurate waveform reproduction while maintaining computational efficiency.
2Productivity
If group delay feature amounts are used for waveform generation, then processing speed is improved, but it becomes difficult to generate waveforms rapidly with high accuracy
Solution Approach 1:
The invention divides the frequency spectrum into multiple bands and calculates group delay compensation parameters for each band separately. This segmentation allows the system to process different frequency regions independently, improving both computational efficiency and accuracy by applying targeted compensation where needed rather than uniform processing across the entire spectrum.
Solution Approach 2:
The invention applies different compensation strategies to different frequency bands based on their specific characteristics. By calculating and applying band-specific group delay compensation parameters, the system optimizes waveform reproduction accuracy for each frequency region while maintaining overall processing efficiency, rather than applying a single uniform approach.
Data Source
AI summary
A speech processing device includes a hardware processor configured to receive input speech and extract speech frames from the input speech. The hardware processor is configured to calculate a spectrum parameter for each of the speech frames, calculate a first phase spectrum for each of the speech frames, calculate a group delay spectrum from the first phase spectrum based on a frequency component of the first phase spectrum, calculate a band group delay parameter in a predetermined frequency band from the group delay spectrum, and calculate a band group delay compensation parameter to compensate a difference between a second phase spectrum reconstructed from the band group delay parameter and the first phase spectrum. The hardware processor is configured to generate a speech waveform based on the spectrum parameter, the band group delay parameter, and the band group delay compensation parameter.