Restoration_of_signals_with_interpolation
Restoration_of_signals_with_interpolation_techniques
Music recordings often suffer from noise. An approach to deal with this issue is to treat noisy segments as missing data. To restore them, one may employ interpolation techniques. Here, music signals are modeled as an autoregressive (AR) process, and three interpolation methods are developed. Maximum likelihood, Gibbs sampling, and Expectation Maximization.