Design And Implementation Of Audio Resampling Algorithms

With the development of theory and algorithms of digital signal processing, multi-rate signal processing becomes more and more important in multimedia signal processing. In order to reduce both computational and storage complexities, sampling rates conversion techniques are definitely required, audio resampling algorithm can be used to convert audio signal between arbitrary sampling rates.Firstly, the basic concepts and the principle of multi-rate signal processing is introduces briefly in the paper; then three audio resampling algorithms, i.e., direct interpolation, Lagrange interpolation and sine interpolation algorithms, are designed and implemented. Direct interpolation method calculates the magnitude of the interpolation points linearly based on the two adjacent points, while Lagrange interpolation algorithm calculates the magnitude of the interpolation points using the Lagrange polynomial with a predefined window length, the sine interpolation method calculates the magnitude of the interpolation points based on the trigonometric polynomial with the dynamically chosed window length. Arbitrary sampling rates conversions have been implemented with the designed algorithms supporting both up-sampling and down-sampling.Results of the evaluations, including time consumption, SNR and SegSNR, have shown that the three proposed algorithms can assure both high quality of the reconstruction signals and high conversion efficiency.