Audio Coding Noise Filling Using Multi-Band Quantization Error
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional audio coding techniques face challenges in effectively managing spectral bins quantized to zero, leading to audible artifacts and inefficient bit rate usage due to poor noise filling resolution and excessive side information.
Innovation Solution
An encoder and decoder system that calculates a multi-band quantization error and uses a common noise intensity value based on band gain information to introduce noise into spectral components, allowing for efficient noise filling while maintaining a good hearing impression and low bit rate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of energy
If spectral bins are quantized to zero to save bit rate, then bit rate is reduced, but audible artifacts are introduced
Solution Approach 1:
The patent converts the harmful quantization error (which causes audible artifacts) into a beneficial noise filling mechanism. By calculating the quantization error and using it to generate noise that fills spectral bins quantized to zero, the system transforms the harmful artifact into a perceptually acceptable noise that masks the quantization distortion while maintaining bit rate efficiency.
Solution Approach 2:
The patent changes the parameter of noise intensity by calculating it based on the quantization error magnitude. Instead of using fixed noise levels, the system dynamically adjusts the noise intensity parameter according to the actual quantization error in each spectral bin, allowing optimal noise filling that balances artifact reduction with bit rate conservation.
2Object-affected harmful factors
If noise filling is applied to all frequency bands, then hearing impression is improved, but bit rate increases due to excessive side information
Solution Approach 1:
The patent applies local quality by calculating noise filling parameters individually for each spectral bin based on its specific quantization error. Instead of using a uniform noise filling approach, the system adapts the noise intensity and characteristics to match the local quantization characteristics of each bin, providing optimal hearing impression while minimizing unnecessary side information transmission.
Solution Approach 2:
The system uses the quantization error itself as the basis for generating the noise filling, making the error signal serve a dual purpose: both as a measure of quantization distortion and as the source material for the noise filling. This self-service approach eliminates the need for separate noise generation mechanisms and reduces the amount of side information required.
3Measurement precision
If frequency-band-specific noise information is transmitted, then noise filling precision is improved, but device complexity and bit rate increase
Solution Approach 1:
The patent extracts the essential noise filling information directly from the quantization error calculation itself. By deriving the noise filling parameters from the existing quantization error data, the system eliminates the need for separate frequency-band-specific noise information transmission, reducing device complexity while maintaining precision through the extracted error-based noise generation.
Data Source
AI summary
An encoder for providing an audio stream on the basis of a transform-domain representation of an input audio signal includes a quantization error calculator configured to determine a multi-band quantization error over a plurality of frequency bands of the input audio signal for which separate band gain information is available. The encoder also includes an audio stream provider for providing the audio stream such that the audio stream includes information describing an audio content of the frequency bands and information describing the multi-band quantization error.A decoder for providing a decoded representation of an audio signal on the basis of an encoded audio stream representing spectral components of frequency bands of the audio signal includes a noise filler for introducing noise into spectral components of a plurality of frequency bands to which separate frequency band gain information is associated on the basis of a common multi-band noise intensity value.