eduzhai > Applied Sciences > Engineering >

Neural Granular Sound Synthesis

  • king
  • (0) Download
  • 20210507
  • Save

... pages left unread,continue reading

Document pages: 6 pages

Abstract: Granular sound synthesis is a popular audio generation technique based onrearranging sequences of small waveform windows. In order to control thesynthesis, all grains in a given corpus are analyzed through a set of acousticdescriptors. This provides a representation reflecting some form of localsimilarities across the grains. However, the quality of this grain space isbound by that of the descriptors. Its traversal is not continuously invertibleto signal and does not render any structured temporality. We demonstrate thatgenerative neural networks can implement granular synthesis while alleviatingmost of its shortcomings. We efficiently replace its audio descriptor basis bya probabilistic latent space learned with a Variational Auto-Encoder. A majoradvantage of our proposal is that the resulting grain space is invertible,meaning that we can continuously synthesize sound when traversing itsdimensions. It also implies that original grains are not stored for synthesis.To learn structured paths inside this latent space, we add a higher-leveltemporal embedding trained on arranged grain sequences. The model can beapplied to many types of libraries, including pitched notes or unpitched drumsand environmental noises. We experiment with the common granular synthesisprocesses and enable new ones.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×