LSP Coding with Predictive Correction for Error-Robust Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing linear prediction coefficient decoding devices struggle with accurate decoding of LSP parameters when transmission errors occur, leading to reduced sound quality and incorrect variable-length decoding due to prioritizing code amount over decoding accuracy.
Innovation Solution
A combined predictive and non-predictive coding/decoding method that uses differential vectors and correction vectors to ensure accurate decoding of linear prediction coefficients, even with transmission errors, by incorporating predictive and non-predictive coding/decoding units.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If vector coding using moving average prediction is used to reduce LSP code amount, then code amount is reduced, but decoding accuracy deteriorates when transmission errors occur
Solution Approach 1:
The LSP parameter vector is segmented into multiple sub-vectors (first sub-vector and second sub-vector), each coded independently with different coding methods. This segmentation allows the system to apply predictive coding to reduce code amount while maintaining accuracy through separate correction coding of specific sub-vectors, resolving the contradiction between code compression and decoding reliability.
Solution Approach 2:
Different coding strategies are applied to different parts of the LSP parameter vector based on their local characteristics. The first sub-vector uses predictive coding with moving average prediction for compression, while the second sub-vector receives correction coding to ensure accuracy. This local differentiation resolves the contradiction by optimizing each segment's coding approach according to its specific requirements.
2Quantity of substance
If predictive coding is used to reduce code amount, then code amount is reduced, but error propagation increases when transmission errors occur
Solution Approach 1:
The harmful predictive component is extracted and applied selectively only to the first sub-vector, while the second sub-vector is corrected independently without relying on predictive coding. This extraction approach allows the system to benefit from predictive coding's compression where safe, while isolating and correcting error-prone sections, thereby reducing error propagation while maintaining code efficiency.
Solution Approach 2:
The LSP parameter vector serves as an intermediary between the compressed predictive coding representation and the final accurate decoded parameters. By introducing this intermediate correction stage that processes sub-vectors independently, the system mediates between the compressed form and the error-free output, preventing error propagation while maintaining compression benefits.
3Productivity
If code amount is prioritized over decoding accuracy, then code efficiency is improved, but sound quality deteriorates
Solution Approach 1:
The coding system dynamically adjusts the balance between compression and accuracy by applying predictive coding to reduce code amount while incorporating correction coding for critical sub-vectors to maintain sound quality. This dynamic approach resolves the contradiction by adaptively managing the trade-off between code efficiency and decoding precision based on the specific characteristics of different sub-vectors.
Data Source
AI summary
A coding method and a decoding method are provided which can use in combination a predictive coding and decoding method which is a coding and decoding method that can accurately express coefficients which are convertible into linear prediction coefficients with a small code amount and a coding and decoding method that can obtain correctly, by decoding, coefficients which are convertible into linear prediction coefficients of the present frame if a linear prediction coefficient code of the present frame is correctly input to a decoding device. A coding device includes: a predictive coding unit that obtains a first code by coding a differential vector formed of differentials between a vector of coefficients which are convertible into linear prediction coefficients of more than one order of the present frame and a prediction vector containing at least a predicted vector from a past frame, and obtains a quantization differential vector corresponding to the first code; and a non-predictive coding unit that generates a second code by coding a correction vector which is formed of differentials between the vector of the coefficients which are convertible into the linear prediction coefficients of more than one order of the present frame and the quantization differential vector or formed of some of elements of the differentials.