6th Community-Based Signal Separation Evaluation Campaign (SiSEC 2016)
Deadline for submission of separation results for audio tasks 

July, 22nd 2016

Following the success of the five previous editions, we are glad to announce the launch of the SiSEC 2016 campaign for the evaluation of source separation algorithms. This campaign is open to anyone, whether you are an expert or a newcomer, you are interested in e.g. audio, biomedical signal or image processing, and your algorithm relies on prior information or not.

This is a community-based endeavor: all aspects of evaluation, including datasets and evaluation criteria, are agreed between the participants.

This year, SiSEC features four audio tasks and one biomedical task:
- Underdetermined speech and music mixtures (UND)
- Two-channel mixtures of speech and real-world background noise (BGN)
- Professionally produced music recordings (MUS) *** Updated 2016! ***
- Asynchronous recordings of speech mixtures (ASY) 
- Biomedical signal separation (BIO) ***New 2016!!***

UND, BGN and ASY are run with the same datasets and metrics as SiSEC 2015. The music separation task MUS features an updated dataset and convenient scripting tools.
If you wish to participate or comment upon the proposed tasks, you can post your comments on https://sisec.inria.fr/

  • 22nd Jul, 2016:      Deadline for submission of separation results for UND, BGN, ASY, MUS
  • 14th Aug, 2016:     Deadline for submission of separation results for BIO
  • 2nd Sep., 2015:     Paper submission deadline of LVA/ICA 2017
  • Nov., 2016:           Publication of SiSEC 2016 results (on the web)
  • 2nd. Dec, 2016:     Deadline for LVA/ICA camera-ready papers
  • 21-23 Feb., 2017:  LVA/ICA conference

Please register to the mailing list This email address is being protected from spambots. You need JavaScript enabled to view it. at

SiSEC 2016 Evaluation Chair
Antoine Liutkus


Springer Proceedings

Springer Proceedings are available for download here [Free access during 4 weeks]