@torres_be_ Profile picture

Bernardo Torres

@torres_be_

PhD Student @TelecomParis, @tp_adasp. Previously intern at @SonyCSLMusic. AI/ML for audio and music signal processing and synthesis.

Similar User
فَرَحْ photo

@farah_1711

Nayera wael photo

@NayraWael3

Marwa🇵🇸 photo

@Marwa13617055

Festival Vézère photo

@festivalVezere

Anne Laure Simonian photo

@AnLorSimonian

Lancelot Da Costa photo

@lancelotdacosta

R 💕 photo

@Raghadoood

joao pedro satuf photo

@jpsatuf

أروه 🦉 photo

@arwaelkady10

منّه photo

@mennazzam

Mylène Schweitz-Grauss photo

@MyleneSchweitz

مَنار photo

@mnrnrz

Louison | NostalgicKid photo

@_nostalgickid

Serge Logereau photo

@sergelogereau

鹿🦌 photo

@NoraTah176

Bernardo Torres Reposted

Interested in ill-posed learning tasks, uncertainty prediction, conditional density estimation or multi-head deep neural networks ? In our new paper, accepted at #ICML24, we tackle these challenges by exploring the Winner-Takes-All (WTA) training scheme. [1/n]

Tweet Image 1

Signals with a high degree of autocorrelation (such as pitched signals) make the training of convnets on raw audio unstable. For Gaussian initialization, the greater the input’s autocorrelation, the greater the variance of the output. Huge relief seeing these kinds of papers <3

Training convnets on waveforms is hard—far harder than on magnitude spectrograms. "Instabilities in Convnets for Raw Audio" approaches this phenomenon from the perspective of sensitivity to initialization. IEEE Signal Processing Letters vol. 31 preprint: hal.science/hal-04528116

Tweet Image 1


So many amazing people in this photo!

@adasp group from ⁦⁦@telecomparis@ieeeICASSP⁩ 2014, Seoul, Korea

Tweet Image 1


Great stuff!

arxiv.org/abs/2404.07970 The paper preprint is out! We'll make an official announcement soon but feel free to have a look first! :)



Bernardo Torres Reposted

🥳 We present our #ICASSP2024 paper: A diffusion model that generates production-quality (bass) audio stems to any audio input. 😎🎸 According to our experience, that's more useful to artists than generating full mixes. 🙃 📜Paper: arxiv.org/abs/2402.01412 🎶Demo:…


Bernardo Torres Reposted

the math in most diffusion papers


Bernardo Torres Reposted

🎹🧠super happy to share that our DDSP review has been published in @FrontiersIn Signal Processing! 🧠🎹 frontiersin.org/articles/10.33… thanks to my excellent co-authors @jordieshier, @noindent, George Fazekas, and @InstrumentsLab; to @cerkut1 for editing; and to all the reviewers!


Congratz @howariou , @deeplearnmusic, @gaetan_hadjeres , @GeoffroyPeeters on the best paper award with PESTO at @ISMIRConf !!!!


Happy to be participating in this year’s #ISMIR2023!

Tweet Image 1

Bernardo Torres Reposted

Huan had her visa to travel to Italy for @ISMIRConf rejected - for the second time in a row, despite having papers to present both years, which is incredibly unfair! If anyone can help, please get in touch 🙏

Does any one know lawyer in Italy and could reference me to? For visa refusals, the embassy told me to find lawyers in Italy. I would like to pay whatever I can to get it sorted…



Bernardo Torres Reposted

Oldies but goldies: H. Kuhn, The Hungarian Method for the assignment problem, 1955. The most celebrated algorithm to solve efficiently (almost cubic time) the assignment problem. en.wikipedia.org/wiki/Hungarian…


PESTO is a pitch estimator that uses a smart architecture design to naturally get pitch-shift invariance. This + SSL training results in an extremely small model with a great performance. Cool work!

New paper: PESTO pitch estimation is on par with CREPE, but ~800x smaller. Due to its self-supervised objective, it can be easily trained on custom data. 🪩Site: tinyurl.com/pesto-pitch 📜Paper: arxiv.org/pdf/2309.02265… 🧾Code: github.com/SonyCSLParis/p… @SonyCSLMusic @howariou

Tweet Image 1


Bernardo Torres Reposted

New paper: PESTO pitch estimation is on par with CREPE, but ~800x smaller. Due to its self-supervised objective, it can be easily trained on custom data. 🪩Site: tinyurl.com/pesto-pitch 📜Paper: arxiv.org/pdf/2309.02265… 🧾Code: github.com/SonyCSLParis/p… @SonyCSLMusic @howariou

Tweet Image 1

Bernardo Torres Reposted

We've improved the SOTA of singer representation learning (comparing different SSL techniques). We provide paper, code, and trained models: 🪩Blog: tinyurl.com/ssl-singer 📜Paper: tinyurl.com/ssl-sing-paper 🧾Code: tinyurl.com/ssl-sing-github Great work @torres_be_! 😃 @ISMIRConf

Tweet Image 1

Bernardo Torres Reposted

Álbuns brasileiros além do limite da foto, gerado por IA: Chega de Saudade, João Gilberto

Tweet Image 1

Bernardo Torres Reposted

updated the classic music production meme "I thought that using loops was cheating"

Tweet Image 1

Loading...

Something went wrong.


Something went wrong.