Harmonic Signal Coding in Transform-Based Audio Codecs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Transform-based audio codecs introduce artifacts when coding strongly harmonic signals, such as music notes or sustained vowel sounds, due to poor quantization of weak harmonics, leading to audible distortion and fluctuation.
Innovation Solution
Applying spectral weights to frequency coefficients before vector quantization to redistribute noise and ensure that weak harmonics are not completely missed, by de-emphasizing strong peaks and emphasizing weak peaks, thereby improving the preservation of perceptually important frequency components in low bitrate systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If vector quantization is used to code frequency coefficients, then encoding efficiency is improved, but weak harmonics are poorly quantized and artifacts are introduced
Solution Approach 1:
The patent applies different weighting factors to different frequency coefficients based on their harmonic strength. Strong harmonics receive lower weighting (de-emphasis) while weak harmonics receive higher weighting (emphasis), allowing the vector quantizer to allocate bits more effectively and preserve weak harmonics that would otherwise be lost in the quantization process
Solution Approach 2:
The patent modifies the frequency coefficients by multiplying them with spectral weighting factors before quantization. This parameter transformation changes the relative amplitude distribution of coefficients, transforming the quantization problem into one where weak harmonics become more prominent and can be captured with available bits
2Reliability
If spectral weights are applied to emphasize weak harmonics, then perceptual audio quality is improved, but processing complexity increases
Solution Approach 1:
The spectral weighting operation is performed in advance during the encoding process, before the vector quantization step. By pre-processing the frequency coefficients with the appropriate weighting factors, the system prepares the data in advance to ensure weak harmonics are preserved, eliminating the need for complex post-processing or additional decoding operations
Data Source
AI summary
Systems and methods include audio encoders having improved coding of harmonic signals. The audio encoders can be implemented as transform-based codecs with frequency coefficients quantized using spectral weights. The frequency coefficients can be quantized by use of the generated spectral weights applied to the frequency coefficients prior to the quantization or by use of the generated spectral weights in computation of error within a vector quantization that performs the quantization. Additional apparatus, systems, and methods are disclosed.