upsampling

Upsampling is the process of increasing the sampling rate or resolution of a discrete signal or image. In signal processing, upsampling expands a sequence by a factor L by inserting L-1 zeros between samples (zero-stuffing). The resulting sequence is then passed through a low-pass (anti-imaging) filter to reconstruct a continuously valued signal and suppress imaging artifacts created by the upsampling.

In audio and communications, upsampling is used to prepare a signal for digital-to-analog converters and to match sampling rates between systems. The reconstruction filter is designed to preserve the original spectrum up to the new Nyquist frequency.

In image and video processing, upsampling enlarges the spatial resolution by interpolating new pixel values. Common interpolation methods include nearest neighbor, bilinear, bicubic, and Lanczos; higher-quality results can be obtained with spline interpolation or super-resolution techniques that use multiple frames or learned models.

Upsampling does not create new information; it estimates intermediate values based on existing samples. The choice of method involves trade-offs among quality, sharpness, artifacts, and computational cost. In practice, upsampling is used in digital media playback, display scaling, and preprocessing for further analysis, while advanced approaches like super-resolution seek to recover finer details using additional information or learned models.