Audio Encoding Gain Correction via Differential Prediction

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional audio encoding techniques are insufficient in reducing the code rate of gain information and fail to provide fine control over each frequency band, resulting in unsatisfactory sound quality.

Innovation Solution

The method corrects gain information from past and initial gains to suppress the gain code rate, allowing for control of gain per band as a minimum unit, reducing the code rate of gain information, and improving sound quality with reduced calculation and bit rates.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the initial gain A is always encoded for each band, then the gain information can be accurately represented, but the code rate of gain information increases

Engineering Contradiction:
Improvegain information accuracyVSAvoidgain code rate
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent applies preliminary action by predicting the gain value for the current frame based on gain values from previous frames before actual encoding. The prediction unit calculates predicted gain information using past frame data, and this predicted value serves as a baseline that requires minimal encoding. Only the difference between actual and predicted gain needs to be encoded, significantly reducing the code rate while maintaining accuracy.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent changes the parameter representation from encoding absolute gain values to encoding differential values (d_scf). By transforming the gain parameter from its original form to a difference form relative to predicted values, the encoding efficiency improves. The differential values have smaller magnitudes and can be encoded with fewer bits, thus reducing the gain code rate while preserving the necessary information.

Inventive Principle:
Principle #35Parameter changes

2Loss of information

If the same gain is applied to multiple frequency bands, then the gain code rate is reduced, but fine control over each band is lost resulting in unsatisfactory sound quality

Engineering Contradiction:
Improvegain code rateVSAvoidsound quality
Core Design Contradiction:
Loss of informationVSManufacturing precision

Solution Approach 1:

The patent applies segmentation by dividing the frequency spectrum into multiple bands and applying independent gain control to each band. Instead of using a single common gain for all bands, the system calculates and encodes separate gain information for each frequency band. This segmentation enables fine-grained control over each band's amplitude characteristics, improving sound quality while the prediction mechanism keeps the overall code rate manageable.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements local quality by allowing different encoding precision and prediction strategies for different frequency bands. Each band can have its own predicted gain value and differential encoding parameters tailored to its specific characteristics. This localized approach optimizes the balance between code rate and sound quality for each band individually, rather than applying a uniform strategy across all frequencies.

Inventive Principle:
Principle #3Local quality

3Measurement precision

If Huffman code length increases for large gain differences, then large gain variations are accurately encoded, but the overall code rate increases

Engineering Contradiction:
Improvegain difference encoding accuracyVSAvoidbit rate
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent uses feedback by continuously using previously decoded gain information to predict current frame gain values. The prediction unit incorporates past frame gain data and uses it to generate predictions for the current frame. This feedback mechanism ensures that gain differences to be encoded are typically small, allowing the Huffman coding scheme to use shorter code lengths for the majority of cases while still accurately representing larger variations when they occur.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP2159790B1Audio encoding method, audio decoding method, audio encoding device, audio decoding device, program, and audio encoding/decoding system
Publication Date: 2019.11.13 NEC CORP
  • EP2159790B1 patent drawingFigure 1
  • EP2159790B1 patent drawingFigure 2
  • EP2159790B1 patent drawingFigure 3

AI summary

An audio encoding device (1A) corrects initial gain information calculated for an arbitrary frame, based on gain information of a stored past frame, thereby calculating gain information to be used in the frame. The audio encoding device (1A) encodes the calculated gain information as a difference from the gain information of the past frame. An audio decoding deice (3A) receives the differential gain, and calculates the gain of the arbitrary frame based on the gain used in the past frame, thereby generating a decoded audio signal.