Digital Audio Signal Extreme Value Correction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for processing digital audio signals, such as those used in digital television broadcasting and MP3 compression, often result in significant audio quality degradation due to the cutoff of high-frequency signal components and quantization errors, leading to a noticeable difference between the corrected and original audio signals, especially in compressed or repeatedly dubbed audio.
Innovation Solution
A method that detects extreme values in digital audio signals, calculates inter-extreme differences, and generates corrective values to adjust adjacent samples, effectively compensating for audio quality degradation by adding or subtracting these values based on the detected extreme values and their differences, thereby enhancing the audio quality to a level closer to the original.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If lossy compression is applied to reduce data size, then transmission efficiency is improved, but audio quality deteriorates due to cutoff of high-frequency signal components
Solution Approach 1:
The patent applies preliminary action by pre-emphasizing the amplitude of audio signal samples before compression. Specifically, it increases the amplitude of samples corresponding to extreme values (peaks and valleys) and adjacent samples before the lossy compression process. This preliminary amplification ensures that when high-frequency components are cut off during compression, the essential waveform characteristics are preserved, allowing the audio quality to be recovered closer to the original after decompression.
2Device complexity
If only extreme value samples are corrected, then processing complexity is reduced, but audio quality improvement is insufficient
Solution Approach 1:
The patent applies local quality by differentiating the correction approach for different regions of the audio signal. It identifies extreme value samples (local maxima and minima) and applies different correction strategies to them versus their adjacent samples. The extreme value samples receive correction based on their inter-extreme relationships, while adjacent samples receive correction based on their proximity to extreme values. This localized differentiation ensures comprehensive audio quality improvement without requiring complex processing of the entire signal uniformly.
3Quantity of substance
If compression ratio is increased to reduce file size, then storage efficiency is improved, but audio fidelity deteriorates
Solution Approach 1:
The patent applies preliminary action by pre-emphasizing the amplitude of audio signal samples before compression. Specifically, it increases the amplitude of samples corresponding to extreme values (peaks and valleys) and adjacent samples before the lossy compression process. This preliminary amplification ensures that when high-frequency components are cut off during compression, the essential waveform characteristics are preserved, allowing the audio quality to be recovered closer to the original after decompression.
Data Source
AI summary
Every extreme value in an audio waveform represented by a digital audio signal having a sequence of samples is detected. A number of samples between samples corresponding to the first and second latest extreme values is detected. A corrective value is generated in response to the detected sample number and a difference between the first and second latest extreme values. Ones are designated among samples in response to the detected sample number. The designated samples include at least (1) a sample adjacently following the sample corresponding to the second latest extreme value, (2) a sample adjacently preceding the sample corresponding to the first latest extreme value, and (3) one of the sample corresponding to the first latest extreme value and the sample corresponding to the second latest extreme value. The designated samples are corrected in response to at least one of current, previous, and feature corrective values.