Luca Comanducci

Glad to have been presented the FakeMusicCaps dataset and the risks and issues behind deepfake music detection!

Link to the paper --> arxiv.org/pdf/2409.10684
Link to the dataset --> zenodo.org/records/1506...

#AIMusic #GenerativeAI #Deepfake #AI #GenerativeMusic

Riccardo Passoni, Francesca Ronchini, Luca Comanducci, Romain Serizel, Fabio Antonacci: Diffused Responsibility: Analyzing the Energy Consumption of Generative Text-to-Audio Diffusion Models https://arxiv.org/abs/2505.07615 https://arxiv.org/pdf/2505.07615 https://arxiv.org/html/2505.07615

Energy consumption of 7 text-to-audio diffusion models was analyzed, identifying the trade-off between audio quality and energy usage at inference time by finding Pareto-optimal solutions.

DDPMs, conditioned on anthropometric measurements, generated personalized HRIRs, demonstrating feasibility for HRTF personalization and comparable performance to existing methods.

thanks to all co-authors!!!

In Towards HRTF Personalization using Denoising Diffusion Models we use #DDPMS to generated personalized Head-related Transfer functions #HRTFs by conditioning the model using anthropometric measurements and working directly in the raw audio domain.

In MambaFoley we combine #DDPMS and the #Mamba selective state-space model to generate foley sounds directly in the waveform domain, while controlling the timing of the generative sounds.

2 Papers accepted at #ICASSP25 🇮🇳 hope to see you in india!

- MambaFoley: Foley Sound Generation using Selective State-Space Models (arxiv.org/pdf/2409.09162)

- Towards HRTF Personalization using Denoising Diffusion Models (arxiv soon)
#ICASSP25 #ICASSP #generativeai

new paper! 🗣️Sketch2Sound💥

Sketch2Sound can create sounds from sonic imitations (i.e., a vocal imitation or a reference sound) via interpretable, time-varying control signals.

paper: arxiv.org/abs/2412.08550
web: hugofloresgarcia.art/sketch2sound

Challenge tasks for DCASE2025 - DCASE The DCASE Steering Group has reviewed the task proposals...

The tasks for DCASE challenge 2025 have been announced.

dcase.community/articles/cha...

Stay tuned for more details.

While the technology behind these experiments may now be a little outdated I still believe that the chapter is a nice snapshot of experimental research in NMP from a few years ago!

Here we consider the case of chamber music and experiment with adaptive metronomes, to enhance the musicians synchronicity, and with binaural audio, to make the playing experience more immersive!

Synchronous Online Ensemble Performance | 3 | Music, Technology, Innov Networked Music Performances (NMPs) overcome the barriers of distance through the mediation of the Internet. In this chapter, we provide an overview of the

I am glad that after a few years this chapter came out, describing some Networked Music Performance research done during the INTERMUSIC project (intermusicproject.eu).

nice idea! would love to be added if you think it might be useful!