In this study, the autoregressive (AR) modeling of the amplitude and frequency parameters of sinusoidal components allows us to interpolate realistically missing audio data, especially in the case of musical modulations such as vibrato or tremolo. The problem of phase discontinuity at the boundaries of the gap is also addressed using an original method that spread the phase error across the interpolating region.

Those methods are used to design an original algorithm for the interpolation of a missing region of a whole set of sinusoids. The objective and subjective tests show that the quality is significantly improved compared to common sinusoidal and temporal interpolation techniques of missing audio data.