Audio Gain Interpolation in Decoding for Smoother Waveforms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio coding techniques, such as MPEG AAC, face issues with discontinuous temporal waveforms in audio signals due to uniform gain application across frames, leading to degradation in sound quality when significant gain changes occur between frames.
Innovation Solution
A decoding apparatus and method that reads encoded gain values at multiple sample positions and uses interpolation information to apply either linear or non-linear interpolation between these points, allowing for more precise gain adjustments at each sample position, thereby smoothing the audio waveform and reducing code amount.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If uniform gain is applied to all samples within one frame, then the gain control is simple to implement, but the temporal waveform becomes discontinuous between frames causing degradation in sound quality
Solution Approach 1:
The frame is divided into multiple sub-sections with different gain values applied to each sub-section. This segmentation allows the gain to vary within a frame rather than being uniform, thereby maintaining temporal waveform continuity while keeping the implementation relatively simple through discrete gain zones.
Solution Approach 2:
The gain value is made dynamic by introducing a gain slope that changes over time within the frame. Instead of a static uniform gain, the gain varies dynamically across sub-sections based on the gain slope, which maintains waveform continuity while improving sound quality.
2Reliability
If gain values are varied at each sample position to maintain waveform continuity, then the sound quality improves, but the code amount increases
Solution Approach 1:
Instead of varying gain at every sample position, the gain variation is applied locally to specific sub-sections of the frame. Each sub-section has a representative gain value, and this local approach maintains waveform continuity while significantly reducing the number of gain parameters that need to be encoded.
Solution Approach 2:
The gain information is represented by changing parameters such as gain slope and sub-section boundaries rather than encoding individual gain values for each sample. This parameter-based representation reduces the code amount while still achieving continuous waveform transitions.
3Ease of operation
If linear interpolation is used between gain sample positions, then the implementation is simple, but the gain accuracy is insufficient when gain changes are significant
Solution Approach 1:
The interpolation method is made dynamic by selecting between linear and non-linear interpolation based on the gain change magnitude. When gain changes are significant, non-linear interpolation is applied to maintain accuracy, while linear interpolation is used when changes are small, thus balancing simplicity and precision.
Solution Approach 2:
The interpolation approach changes based on parameters such as the gain difference between sample positions. By monitoring these parameters, the system automatically selects the appropriate interpolation method, achieving both simplicity and accuracy adaptively.
Data Source
AI summary
The present technology relates to a decoding apparatus, a decoding method and a program which make it possible to obtain sound with higher quality.A demultiplexing circuit demultiplexes an input code string into a gain code string and a signal code string. A signal decoding circuit decodes the signal code string to output a time series signal. A gain decoding circuit decodes the gain code string. That is, the gain decoding circuit reads out gain values and gain inclination values at predetermined gain sample positions of the time series signal and interpolation mode information. An interpolation processing unit obtains a gain value at each sample position between two gain sample positions through linear interpolation or non-linear interpolation according to the interpolation mode based on the gain values and the gain inclination values. A gain applying circuit adjusts a gain of the time series signal based on the gain values. The present technology can be applied to a decoding apparatus.