Time Stretching

Time stretching, the process of altering the duration of an audio or video signal without changing its pitch, is a field of active research aiming to improve the quality and robustness of time-scale modification. Current efforts focus on enhancing existing methods, particularly addressing challenges in handling noise and transient components within signals, often employing techniques like sinusoidal modeling, noise morphing, and deep neural networks (including WaveNet-like architectures and GANs) for improved synthesis and representation. These advancements have implications for various applications, including slow-motion video production, speech enhancement for individuals with dysarthria, and improved robustness in neuromorphic vision systems.