Audio Post-Processor Transient Detection Pre-Echo Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing audio coding technologies introduce pre- and post-echo artifacts at low bitrates, particularly in block-based processing, which affect the subjective quality of audio signals, and existing post-processing solutions require guidance from the encoder or cannot be applied to previously coded materials.
Innovation Solution
A non-guided post-processor that converts audio signals to a time-frequency representation, estimates transient locations, and manipulates the representation to reduce or eliminate pre-echoes and enhance transient attacks, using prediction filters to shape the spectral frames without explicit transient detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of substance
If block-based audio processing is used for compression, then audio signal compression is achieved, but pre- and post-echo artifacts are introduced
Solution Approach 1:
The patent applies preliminary action by detecting transient locations in the audio signal before encoding and using this information to adjust the encoding parameters or apply preprocessing to prevent pre-echo artifacts from occurring in the first place. The system identifies transient onsets and modifies the block-based processing accordingly to avoid spreading quantization noise before transient events.
2Object-affected harmful factors
If transform codec block-switching is used to prevent pre- and post-echo artifacts, then artifact reduction is achieved, but the solution cannot be applied a-posteriori to archived sound material
Solution Approach 1:
The patent implements self-service by enabling the post-processor to autonomously detect transient locations and estimate pre-echo characteristics directly from the archived audio material without requiring any guidance information or metadata from the original encoder. The system independently analyzes the decoded signal to identify transient events and applies appropriate correction measures.
3Object-affected harmful factors
If state of the art post-processing techniques are used to suppress pre- and post-echo artifacts, then artifact reduction is achieved, but control information from the encoder is required
Solution Approach 1:
The patent applies the taking out principle by extracting and removing the dependency on encoder control information. Instead of relying on metadata or side information transmitted from the encoder, the post-processor extracts all necessary transient detection and pre-echo estimation information directly from the decoded audio signal itself, making the system independent of the encoding process.
4Ease of manufacture
If quantization noise is spread over the entire block duration, then frequency domain transform coding is simplified, but pre-echo artifacts are introduced
Solution Approach 1:
The patent implements local quality by applying different noise shaping or quantization strategies to different portions of the audio block based on the presence of transient events. Instead of uniform noise distribution across the entire block, the system identifies transient locations and applies localized processing to prevent pre-echo artifacts only in the regions where they occur, while maintaining simple uniform processing in non-transient regions.
Data Source
AI summary
Apparatus for post-processing an audio signal, including: a converter for converting the audio signal into a time-frequency representation; a transient location estimator for estimating a location in time of a transient portion using the audio signal or the time-frequency representation; and a signal manipulator for manipulating the time-frequency representation, wherein the signal manipulator is configured to reduce or eliminate a pre-echo in the time-frequency representation at a location in time before the transient location or to perform a shaping of the time-frequency representation at the transient location to amplify an attack of the transient portion.