Hyperbolic Audio Source Separation

12/09/2022
by   Darius Petermann, et al.
0

We introduce a framework for audio source separation using embeddings on a hyperbolic manifold that compactly represent the hierarchical relationship between sound sources and time-frequency features. Inspired by recent successes modeling hierarchical relationships in text and images with hyperbolic embeddings, our algorithm obtains a hyperbolic embedding for each time-frequency bin of a mixture signal and estimates masks using hyperbolic softmax layers. On a synthetic dataset containing mixtures of multiple people talking and musical instruments playing, our hyperbolic model performed comparably to a Euclidean baseline in terms of source to distortion ratio, with stronger performance at low embedding dimensions. Furthermore, we find that time-frequency regions containing multiple overlapping sources are embedded towards the center (i.e., the most uncertain region) of the hyperbolic space, and we can use this certainty estimate to efficiently trade-off between artifact introduction and interference reduction when isolating individual sounds.

READ FULL TEXT

page 1

page 4

research
11/07/2018

Class-conditional embeddings for music source separation

Isolating individual instruments in a musical mixture has a myriad of po...
research
12/09/2021

Music demixing with the sliCQ transform

Music source separation is the task of extracting an estimate of one or ...
research
02/01/2018

Approximate Message Passing for Underdetermined Audio Source Separation

Approximate message passing (AMP) algorithms have shown great promise in...
research
07/29/2023

Moisesdb: A dataset for source separation beyond 4-stems

In this paper, we introduce the MoisesDB dataset for musical source sepa...
research
10/04/2019

Modeling the Comb Filter Effect and Interaural Coherence for Binaural Source Separation

Typical methods for binaural source separation consider only the direct ...
research
11/14/2022

MedleyVox: An Evaluation Dataset for Multiple Singing Voices Separation

Separation of multiple singing voices into each voice is a rarely studie...
research
11/06/2019

Finding Strength in Weakness: Learning to Separate Sounds with Weak Supervision

While there has been much recent progress using deep learning techniques...

Please sign up or login with your details

Forgot password? Click here to reset