Coding and Decocidng of Pulse and Residual Parts of an Audio Signal

Simple SummaryContent extracted from patent full text and abstract with AI.

This invention introduces a new method for encoding and decoding audio signals by splitting the signal into 'pulse' (transient) portions and 'stationary' (steady/background) portions. It includes a specialized pulse extractor that identifies sharp transients in the audio using a high-resolution spectrogram, encodes these pulses separately, and then processes the remaining signal (residual) with another encoder. On decoding, the two parts are combined to reconstruct high-quality audio. This approach significantly improves the audio quality especially for signals with strong transients, such as percussive sounds or voiced speech, by handling different components with tailored strategies.

Use CasesContent extracted from patent full text and abstract with AI.

• Audio compression for music streaming services, especially for genres with strong beats or transients (e.g., EDM, classical, jazz).
• Speech codecs used in telecommunication (e.g., VoIP, mobile voice calls) to improve speech clarity and naturalness.
• Professional audio editing and mastering tools where efficient storage and high-quality transient reproduction are required.
• Hearing aids or assistive listening devices to enhance speech intelligibility and sound quality, especially in noisy environments.
• Portable audio recorders or microphones where high-fidelity yet efficient storage is needed.
• Game audio engines or virtual reality sound systems for lifelike reproduction of environmental and impact sounds.

BenefitsContent extracted from patent full text and abstract with AI.

• Improves coding efficiency and audio quality, especially for signals with strong transients or pulses.
• Enables more accurate reproduction of percussive and voiced speech sounds, reducing typical compression artifacts.
• Allows specialized treatment of different signal components (transient vs. stationary), optimizing overall fidelity and bit rate usage.
• Reduces distortion and smearing of transients, often a problem in traditional transform codecs like MP3 or AAC.
• Provides a flexible framework that can be adapted for different audio coding standards or devices.
• Potentially enables lower bitrates without sacrificing perceived audio quality, which is crucial for streaming and storage.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

The present invention relates to an audio encoder (100, 101) for encoding an audio signal (PCMi) comprising an pulse portion (P) and a stationary portion, comprising: a pulse extractor (11,110) configured for extracting the pulse portion (P) from the audio signal (PCMi), further comprising a pulse coder (132) for encoding the extracted pulse portion (P) to acquire an encoded pulse portion (CP); wherein the pulse extractor (110) is configured to determine a spectrogram of the audio signal (PCMi) to extract the pulse portion (P), wherein the spectrogram having higher time resolution than the signal encoder (152, 156'); a signal encoder (152, 156') configured for encoding a residual (R) signal derived from the audio signal (PCMi) to acquire an encoded residual (CR) signal, the residual (R) signal being derived from the audio signal (PCMi) so that the pulse portion (P) is reduced or eliminated from the audio signal (PCMi); and an output interface (170) configured for outputting the encoded pulse portion (CP) and the encoded residual (CR) signal to provide an encoded signal.