Pitch Cycle Energy Parameter Determination for Speech Encoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for encoding speech signals are inefficient in terms of bandwidth and storage resources, often degrading signal quality due to incomplete representation of speech components, particularly in low bit rate encoding scenarios.
Innovation Solution
A method and apparatus for determining pitch cycle energy parameters in electronic devices, which involves obtaining a frame of a speech signal, calculating filter coefficients, and segmenting a residual signal to map energy regions between peaks, allowing for improved scaling and synthesis of an excitation signal to match the energy contour of the speech signal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If speech signals are encoded using compression schemes to reduce bandwidth and storage requirements, then bandwidth and storage efficiency is improved, but speech signal quality deteriorates due to incomplete representation of speech components
Solution Approach 1:
The speech signal is divided into pitch cycles, and each pitch cycle is further segmented into multiple segments. This segmentation allows the encoding system to capture important speech characteristics at different granularities, maintaining speech quality while reducing the overall data rate by not encoding every sample at full resolution.
Solution Approach 2:
Different encoding precision is applied to different parts of the speech signal. The excitation signal is encoded with higher precision at pitch peak locations where it contains most of the speech energy, and with lower precision in between peaks. This local quality approach maintains overall speech quality while reducing total bandwidth requirements.
2Quantity of substance
If low bit rate encoding is used to reduce bandwidth consumption, then bandwidth efficiency is improved, but speech representation completeness deteriorates
Solution Approach 1:
Each pitch cycle is divided into multiple segments, allowing the system to capture essential speech characteristics distributed across segments. This segmentation enables low bit rate encoding by selecting only the most important segment parameters to encode, rather than encoding the entire pitch cycle at full resolution.
Solution Approach 2:
The system encodes only the most critical portions of the speech signal in detail (pitch peaks and adjacent regions) while using coarser representation for other parts. This partial action approach ensures that the essential speech information is preserved at low bit rates without attempting to encode every detail.
3Manufacturing precision
If pitch cycle energy parameters are determined and excitation signal scaling is applied to match energy contours, then speech quality is improved, but computational complexity increases
Solution Approach 1:
The pitch cycle is divided into segments for energy parameter determination. This segmentation reduces computational complexity by limiting the calculation scope to smaller segments rather than processing the entire pitch cycle at once, while still capturing the essential energy variations needed for quality speech reconstruction.
Solution Approach 2:
Energy parameters are determined locally for each segment of the pitch cycle, particularly focusing on regions around pitch peaks where energy variations are most significant. This local approach improves speech quality by capturing critical energy information while reducing overall computational complexity compared to analyzing the entire signal globally.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An electronic device for determining a set of pitch cycle energy parameters is described. The electronic device includes a processor and executable instructions stored in memory. The electronic device obtains a frame, a set of filter coefficients and a residual signal based on the frame and the set of filter coefficients. The electronic device determines a set of peak locations based on the residual signal and segments the residual signal such that each segment includes one peak. The electronic device determines a first set of pitch cycle energy parameters based on a frame region between two consecutive peak locations and maps regions between peaks in the residual signal to regions between peaks in a synthesized excitation signal to produce a mapping. The electronic device determines a second set of pitch cycle energy parameters based on the first set of pitch cycle energy parameters and the mapping.