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

Juan Alonso Moreno; Francesco Bigoni; George Palamas

doi:10.5281/zenodo.3898855

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 proceeding › Article in proceeding › Research › peer-review

2 Citations (Scopus)

101 Downloads (Pure)

Abstract

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 language	English
Title of host publication	Proceedings of the 17th Sound and Music Computing Conference
Editors	Simone Spagnol, Andrea Valle
Number of pages	7
Publisher	Axea sas/SMC Network
Publication date	2020
Pages	364-370
ISBN (Electronic)	978-88-945415-0-2
DOIs	https://doi.org/10.5281/zenodo.3898855
Publication status	Published - 2020
Event	17th Sound and Music Computing Conference - Torino, Italy Duration: 24 Jun 2020 → 26 Jun 2020 Conference number: 17 https://smc2020torino.it/uk/

Conference

Conference	17th Sound and Music Computing Conference
Number	17
Country/Territory	Italy
City	Torino
Period	24/06/2020 → 26/06/2020
Internet address	https://smc2020torino.it/uk/

Series	Proceedings of the Sound and Music Computing Conference
ISSN	2518-3672

Access to Document

10.5281/zenodo.3898855Licence: CC BY 3.0

Open Access articleFinal published version, 588 KBLicence: CC BY 3.0

AUB Link

Search for the material in Aalborg University Library's search engine

Cite this

@inproceedings{ed80dfc64a2f4ca1af663f73f6851d3b,

title = "Latent Birds: A Bird's-Eye View Exploration of the Latent Space",

abstract = "The use of a generative approach for sound synthesis breaksthrough 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 onmel-spectrograms of bird song and call recordings. Thevocalizations are reconstructed by sampling the latent spaceand decompressing the resulting mel-spectrogram. Theresults are quite promising, in that our system is able togenerate a variety of bird vocalizations depicting plausiblesongs and calls, by interpolating between existing vocalizations or sampling the latent space. A Twitter bot thatpublishes a unique daily bird vocalization is also implemented.",

author = "Moreno, {Juan Alonso} and Francesco Bigoni and George Palamas",

year = "2020",

doi = "10.5281/zenodo.3898855",

language = "English",

series = "Proceedings of the Sound and Music Computing Conference",

publisher = "Axea sas/SMC Network",

pages = "364--370",

editor = "Simone Spagnol and Andrea Valle",

booktitle = "Proceedings of the 17th Sound and Music Computing Conference",

note = "17th Sound and Music Computing Conference, SMC 2020 ; Conference date: 24-06-2020 Through 26-06-2020",

url = "https://smc2020torino.it/uk/",

}

Moreno, JA, Bigoni, F & Palamas, G 2020, Latent Birds: A Bird's-Eye View Exploration of the Latent Space. in S Spagnol & A Valle (eds), Proceedings of the 17th Sound and Music Computing Conference. Axea sas/SMC Network, Proceedings of the Sound and Music Computing Conference, pp. 364-370, 17th Sound and Music Computing Conference, Torino, Italy, 24/06/2020. https://doi.org/10.5281/zenodo.3898855

Latent Birds: A Bird's-Eye View Exploration of the Latent Space. / Moreno, Juan Alonso; Bigoni, Francesco; Palamas, George.
Proceedings of the 17th Sound and Music Computing Conference. ed. / Simone Spagnol; Andrea Valle. Axea sas/SMC Network, 2020. p. 364-370 (Proceedings of the Sound and Music Computing Conference).

Research output: Contribution to book/anthology/report/conference proceeding › Article in proceeding › Research › peer-review

TY - GEN

T1 - Latent Birds

T2 - 17th Sound and Music Computing Conference

AU - Moreno, Juan Alonso

AU - Bigoni, Francesco

AU - Palamas, George

N1 - Conference code: 17

PY - 2020

Y1 - 2020

N2 - The use of a generative approach for sound synthesis breaksthrough 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 onmel-spectrograms of bird song and call recordings. Thevocalizations are reconstructed by sampling the latent spaceand decompressing the resulting mel-spectrogram. Theresults are quite promising, in that our system is able togenerate a variety of bird vocalizations depicting plausiblesongs and calls, by interpolating between existing vocalizations or sampling the latent space. A Twitter bot thatpublishes a unique daily bird vocalization is also implemented.

AB - The use of a generative approach for sound synthesis breaksthrough 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 onmel-spectrograms of bird song and call recordings. Thevocalizations are reconstructed by sampling the latent spaceand decompressing the resulting mel-spectrogram. Theresults are quite promising, in that our system is able togenerate a variety of bird vocalizations depicting plausiblesongs and calls, by interpolating between existing vocalizations or sampling the latent space. A Twitter bot thatpublishes a unique daily bird vocalization is also implemented.

U2 - 10.5281/zenodo.3898855

DO - 10.5281/zenodo.3898855

M3 - Article in proceeding

T3 - Proceedings of the Sound and Music Computing Conference

SP - 364

EP - 370

BT - Proceedings of the 17th Sound and Music Computing Conference

A2 - Spagnol, Simone

A2 - Valle, Andrea

PB - Axea sas/SMC Network

Y2 - 24 June 2020 through 26 June 2020

ER -

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

Abstract

Conference

Access to Document

AUB Link

Fingerprint

Cite this