Audio Signal Encoding Layers for Compression Quality
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autoencoder models for audio signal encoding and decoding suffer from signal loss at each encoding layer, making it difficult to reconstruct high-quality audio signals due to repeated encoding operations.
Innovation Solution
The method involves separate encoding and quantization processes for each layer in the autoencoder structure, where each encoding layer generates a bitstring by encoding and quantizing the signal from the previous layer, and corresponding decoding layers dequantize and decode these bitstrings to reconstruct the audio signal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If encoding operation is performed repeatedly by multiple layers, then compression efficiency is improved, but signal quality deteriorates due to loss of original audio signal
Solution Approach 1:
The encoding process is segmented into multiple encoding layers, where each layer performs encoding independently and generates its own bitstring. This segmentation allows the system to achieve high compression efficiency through multiple layers while preserving signal quality by preventing cumulative signal loss that would occur in repeated sequential encoding operations.
2Measurement precision
If encoding and decoding are performed using neural network model, then coding quality is improved, but device complexity increases
Solution Approach 1:
The neural network model is segmented into multiple independent encoding layers, each with its own encoder and decoder components. This segmentation enables the system to achieve high coding quality through sophisticated neural network processing while managing device complexity by distributing the computational load across multiple modular layers rather than requiring a single complex model.
Data Source
AI summary
Disclosed are a method of encoding and decoding an audio signal and an encoder and a decoder performing the method. The method of encoding an audio signal includes identifying an input signal, and generating a bitstring of each encoding layer by applying, to the input signal, an encoding model including a plurality of successive encoding layers that encodes the input signal, in which a current encoding layer among the encoding layers is trained to generate a bitstring of the current encoding layer by encoding an encoded signal which is a signal encoded in a previous encoding layer and quantizing an encoded signal which is a signal encoded in the current encoding layer.