Apparatus and Method for Noise Shaping Using Subspace Projections for Low-Rate Coding of Speech and Audio

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

This invention provides a method and apparatus for encoding and decoding speech and audio signals at low bit-rates, using a noise shaping technique based on subspace projections. The core idea is to transform the audio signal to a new domain where quantization noise can be precisely controlled and shaped to match perceptual models, thereby maintaining higher audio quality even when using fewer bits. The approach works by applying mathematically designed transforms (using covariance matrices) before quantization, allowing efficient compression without the muffled sound typical of other low-bitrate codecs.

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

• Mobile voice communication systems (e.g., voice over LTE, 5G, or VoIP) where low bitrate is required.
• Audio streaming services where bandwidth is limited, such as online radio or music streaming to remote areas.
• Archival of large audio datasets at low storage cost while preserving perceptual quality.
• Hearing aids and cochlear implants requiring low-power, low-data-rate audio transmission and processing.
• Teleconferencing systems aiming to deliver clear speech at reduced network load.
• Embedded devices (like smart speakers or IoT audio sensors) with limited computational and transmission resources.

BenefitsContent extracted from patent full text and abstract with AI.

• Implements low-bitrate audio encoding without introducing perceptible quality loss or muffled sound, outperforming traditional codecs in quality at low rates.
• Reduces computational complexity compared to state-of-the-art techniques, making it suitable for mobile and resource-limited hardware.
• Enables efficient and flexible shaping of quantization noise according to perceptual (psychoacoustic) models, improving subjective listening experience.
• Allows both speech and general audio to be handled efficiently, uniting two traditionally separate audio coding paradigms.
• Hybrid approach offers scalable performance, converging to traditional high-efficiency coding at higher bitrates.
• Improved signal-to-noise ratio (SNR) and perceptual quality (as confirmed by listening tests and objective measures) at low bitrates.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

An apparatus for encoding an audio input signal to obtain an encoded audio signal is provided. The apparatus comprises a transformation module (110) configured to transform the audio input signal from an original domain to a transform domain to obtain a transformed audio signal. Moreover, the apparatus comprises an encoding module (120), configured to quantize the transformed audio signal to obtain a quantized signal, and configured to encode the quantized signal to obtain the encoded audio signal. The transformation module (110) is configured to transform the audio input signal depending on a plurality of predefined power values of quantization noise in the original domain.