Encoder, decoder and method for encoding and decoding

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

The invention describes an improved method and system for encoding and decoding audio signals (such as speech or music) using signal prediction, matrix factorization, and a transform-based approach. By applying matrix factorization (like Vandermonde or eigenvalue decomposition) to the filter characteristics of the audio signal, the system transforms the audio residuals into a decorrelated frequency-like domain where they can be more efficiently quantized and encoded. This approach reduces computational complexity and improves coding efficiency and audio quality compared to conventional methods such as ACELP, especially at high bit rates.

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

• Speech codecs for mobile phones, VoIP, and communication devices
• Audio compression in music streaming, digital audio broadcasting, and file storage
• Hearing aids and assistive listening devices with low-delay audio processing
• Real-time speech or audio transmission in conferencing or gaming applications
• Integration into standards like AMR-WB, G.718, or MPEG audio codecs to enhance efficiency
• Edge devices or IoT systems requiring efficient, high-quality audio compression

BenefitsContent extracted from patent full text and abstract with AI.

• Improved computational efficiency, especially at high bit rates, enabling faster encoding and decoding with less hardware or battery drain
• Maintains or improves perceived audio quality (signal-to-noise ratio) compared to conventional codecs
• Allows for simpler and more optimal quantization in a decorrelated transform domain
• Enables efficient processing of short audio frames, which is crucial for real-time and low-latency applications
• Can be implemented as software or hardware, and is readily adaptable to various digital platforms
• Supports integration with existing audio codec standards, enhancing them without major redesign
• Reduces complexity growth with increasing bit rates, making the solution scalable and future-proof

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

An encoder for encoding an audio signal into a data stream comprises a predictor, a factorizer, a transformer and a quantize and encode stage. The predictor is configured to analyze the audio signal in order to obtain prediction coefficients describing a spectral analog of the audio signal or a fundamental frequency of the audio signal and subject the audio signal to an analysis filter function dependent on the prediction coefficients in order to output a residual signal of the audio signal. The factorizer is configured to apply a matrix factorization onto an audiocorrelation or covariance matrix of synthesis filter function defined by the prediction coefficients to obtain factorized matrices. The transformer is configured to transform the residual signal based on the factorized matrices to obtain a transformed residual signal. The quantize and decode stage is configured to quantize the transformed residual signal to obtain a quantized transformed residual signal or an encoded quantized transformed residual signal.