Low Bitrate

Low-bitrate audio coding aims to compress audio signals—speech, music, and general sounds—into extremely small data sizes while preserving high perceptual quality. Current research focuses on developing neural codecs employing architectures like transformers, convolutional neural networks, and vector quantization, often incorporating self-supervised learning and hybrid approaches combining neural networks with traditional digital signal processing techniques to improve efficiency and robustness. These advancements are significant for applications requiring efficient audio transmission and storage, such as mobile communication, streaming services, and assistive technologies, by enabling high-quality audio at drastically reduced bandwidth requirements.