Adaptive ADPCM Quantization for Sound Compression and Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional ADPCM techniques face limitations in further improving compressibility while maintaining sound quality, as they struggle to effectively adapt quantization characteristics based on short- and long-period changes in sound signals.
Innovation Solution
The proposed ADPCM encoding and decoding apparatus utilize high-frequency and low-frequency measuring sections to detect short- and long-period changes in sound signals, allowing for adaptive quantization and inverse-quantization, which optimizes quantization rates and bit lengths, thereby enhancing compressibility and sound quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional ADPCM techniques are used to compress data, then data compressibility is improved, but sound quality deteriorates
Solution Approach 1:
The patent implements dynamic adaptation of quantization characteristics by detecting short-period and long-period signal changes. The system switches between different quantization characteristics (first and second characteristics) based on the detected signal properties, allowing the quantization width to adapt dynamically to varying signal conditions. This dynamic approach maintains sound quality by using appropriate quantization levels for different signal types while achieving compression.
Solution Approach 2:
The patent changes the quantization parameter (quantization width) based on detected signal characteristics. By measuring short-period changes (high-frequency components) and long-period changes (low-frequency components), the system selects appropriate quantization characteristics from multiple predefined sets. This parameter adaptation allows the system to maintain high sound quality for complex signals while achieving efficient compression for simpler signals.
2Quantity of substance
If quantization width is reduced to improve compression, then data compressibility is improved, but measurement precision of signal changes deteriorates
Solution Approach 1:
The patent employs multiple quantization characteristics with different quantization widths suited for different signal conditions. For signals with large short-period or long-period changes, the system selects quantization characteristics with larger widths to maintain measurement precision. For signals with small changes, narrower quantization widths are used to achieve better compression. This adaptive parameter selection resolves the contradiction between compression efficiency and measurement precision.
Solution Approach 2:
The patent applies different quantization characteristics locally based on the specific properties of the signal being encoded. By detecting whether the signal exhibits large or small short-period changes and large or small long-period changes, the system selects the appropriate quantization characteristic for that local signal segment. This localized adaptation ensures that each part of the signal is quantized with the appropriate precision, maintaining overall measurement accuracy while optimizing compression.
Data Source
AI summary
A signal corresponding to a short-period change and a signal corresponding to a long-period change of a sound signal are detected, and optimal quantization is performed based on the combination of the two signals. In an ADPCM encoding apparatus (100), a differential value dn between a 16-bit input signal Xn and a decoded signal Yn-1 of one sample ago is calculated by a subtractor (102). Thereafter, the 16-bit differential value dn is adaptively quantized by an adaptive quantizing section (103), so as to be converted to a (1 to 8)-bit length-variable ADPCM value Dn. Thereafter, the ADPCM value Dn is compression-encoded by a compression-encoding section (108) to generate a signal D′n, and the signal D′n is framed by a framing section (130) and outputted. Further, in an ADPCM decoding apparatus, a framed input signal is subjected to a reverse of the aforesaid process so as to be decoded.