Speech Decoder High Frequency Time Envelope Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing time envelope deformation method in speech coding struggles to accurately adjust the waveform of high frequency band components due to insufficient correlation between low and high frequency band time envelopes, leading to persistent pre-echoes and post-echoes in decoded signals.
Innovation Solution
A speech decoder that utilizes multiple low frequency band time envelopes to calculate and adjust the high frequency band time envelope, leveraging their correlation to produce a less distorted time envelope shape, thereby reducing pre-echoes and post-echoes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If time envelope deformation is performed using single time envelope information from low frequency band components, then the decoding process is simple, but the correlation between low and high frequency band time envelopes is insufficient, making it difficult to adjust the waveform accurately and resulting in pre-echoes and post-echoes
Solution Approach 1:
The patent divides the time envelope information into multiple segments corresponding to different low frequency band components. Instead of using a single time envelope, the system extracts multiple time envelope pieces from different frequency sub-bands and processes them separately before synthesizing the final high frequency band time envelope, thereby improving accuracy while maintaining operational feasibility
Solution Approach 2:
The patent transitions from a single-dimensional time envelope approach to a multi-dimensional approach by incorporating time envelope information from multiple frequency dimensions. The system processes time envelopes from different low frequency sub-bands (different frequency dimensions) and combines them to create a comprehensive high frequency time envelope, adding dimensional richness to the deformation process
2Adaptability or versatility
If frequency envelope adjustment is performed on high frequency band components, then bandwidth extension is achieved, but the time envelope of high frequency band components becomes deformed and flatter, causing pre-echoes and post-echoes
Solution Approach 1:
The patent performs preliminary extraction and processing of time envelope information from multiple low frequency band components before the frequency envelope adjustment is applied to high frequency band components. By preparing the time envelope deformation parameters in advance based on multiple frequency sources, the system compensates for the flattening effect that will occur during frequency envelope adjustment, preserving the original time envelope shape characteristics
Solution Approach 2:
The system uses the extracted time envelope information from multiple low frequency sub-bands as feedback to guide the high frequency band time envelope construction. The processed time envelope pieces provide feedback signals that inform the deformation process, allowing the system to counteract the flattening effect and maintain accurate time envelope shapes after frequency envelope adjustment
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A speech decoder (1) includes a demultiplexing unit (1a), a low frequency band decoding unit (1b), a band splitting filter bank unit (1c), a coded sequence analysis unit (Id), a coded sequence decoding / dequantization unit (1e), a high frequency band generation unit (1h), low frequency band time envelope calculation units (1f1 to 1fn) that acquire a plurality of low frequency band time envelopes, a time envelope calculation unit (1g) that calculates high frequency band time envelopes using time envelope information and the plurality of low frequency band time envelopes, a time envelope adjustment unit (1i) that adjusts the time envelope of high frequency band components using the time envelopes obtained by the time envelope calculation unit (1g), and a band synthesis filter bank unit (1j).