Latent Birds: A Bird's-Eye View Exploration of the Latent Space

Juan Alonso Moreno, Francesco Bigoni, George Palamas

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

2 Citations (Scopus)
101 Downloads (Pure)


The use of a generative approach for sound synthesis breaks
through the limitations of traditional approaches, proposing novel ways to explore creative ideas. This paper demonstrates a method to generate original bird vocalizations using a Variational Convolutional Autoencoder trained on
mel-spectrograms of bird song and call recordings. The
vocalizations are reconstructed by sampling the latent space
and decompressing the resulting mel-spectrogram. The
results are quite promising, in that our system is able to
generate a variety of bird vocalizations depicting plausible
songs and calls, by interpolating between existing vocalizations or sampling the latent space. A Twitter bot that
publishes a unique daily bird vocalization is also implemented.
Original languageEnglish
Title of host publicationProceedings of the 17th Sound and Music Computing Conference
EditorsSimone Spagnol, Andrea Valle
Number of pages7
PublisherAxea sas/SMC Network
Publication date2020
ISBN (Electronic)978-88-945415-0-2
Publication statusPublished - 2020
Event17th Sound and Music Computing Conference - Torino, Italy
Duration: 24 Jun 202026 Jun 2020
Conference number: 17


Conference17th Sound and Music Computing Conference
Internet address
SeriesProceedings of the Sound and Music Computing Conference


Dive into the research topics of 'Latent Birds: A Bird's-Eye View Exploration of the Latent Space'. Together they form a unique fingerprint.

Cite this