Subband Audio Encoding Using Gain Differences for Tone Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing signal encoding and decoding technologies face challenges in maintaining tone quality when expanding the frequency band, often resulting in auditory distortion and inefficiencies in encoding high-frequency signals, particularly when reducing bit rates.
Innovation Solution
A signal encoding device and method that partition input time-series signals into subbands, generate and encode low-frequency and high-frequency subband signals, and use gain information to control and encode differences between subbands, allowing for efficient multiplexing and decoding to minimize distortion and improve encoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If high-efficiency compression is applied to reduce bit rate, then data amount is reduced, but tone quality is drastically deteriorated
Solution Approach 1:
The audio signal is divided into low-frequency and high-frequency components. The low-frequency signal is encoded with high precision to maintain tone quality, while the high-frequency signal is encoded with reduced precision or synthesized from low-frequency information, thereby reducing overall bit rate while preserving perceptual quality.
Solution Approach 2:
Different encoding precision is applied to different frequency bands. The low-frequency band receives higher encoding precision to preserve fundamental tone quality, while the high-frequency band uses lower precision or synthesis techniques, optimizing the balance between bit rate and perceived audio quality.
2Quantity of substance
If high-frequency signal is limited to maintain tone quality and reduce bit rate, then bit rate is reduced, but high-frequency signal cannot be reproduced
Solution Approach 1:
The low-frequency signal serves as an intermediary to reconstruct the high-frequency signal. By encoding the low-frequency signal with sufficient precision and using it to synthesize or predict the high-frequency components, the system recovers high-frequency information without directly encoding it, thereby maintaining bit rate efficiency while preventing information loss.
3Device complexity
If band expansion is performed at decoding side only, then format is not changed, but dramatic effect is not shown in tone quality
Solution Approach 1:
The encoding process prepares the signal by separating frequency bands and encoding them with appropriate precision levels. This preliminary processing enables the decoder to reconstruct high-frequency components more effectively, achieving dramatic tone quality improvement without requiring format changes or complex decoding operations.
4Loss of information
If LPC filter or subband filter bank is used to expand band, then band expansion is achieved, but auditory distortion occurs
Solution Approach 1:
The system changes the encoding parameters differently for low-frequency and high-frequency bands. Low-frequency components are encoded with higher precision and preserved directly, while high-frequency components use synthesis or prediction methods with adjusted parameters. This parameter differentiation expands the effective bandwidth while minimizing auditory distortion by avoiding uniform processing of all frequency bands.
Data Source
AI summary
A signal encoding device for encoding an input time-series signal includes: partitioning means; low-frequency encoding means; high-frequency gain information generating means; low-frequency reference value information generating means; high-frequency gain difference information generating means; high-frequency gain difference information encoding means; and multiplexing means.