Sounds carry a large amount of information about our everyday environment and physical events that take place in it. We can perceive the sound scene we are within (busy street, office, etc.), and recognize individual sound sources (car passing by, footsteps, etc.). Developing signal processing methods to automatically extract this information has huge potential in several applications, for example searching for multimedia based on its audio content, making context-aware mobile devices, robots, cars etc., and intelligent monitoring systems to recognize activities in their environments using acoustic information. However, a significant amount of research is still needed to reliably recognize sound scenes and individual sound sources in realistic soundscapes, where multiple sounds are present, often simultaneously, and distorted by the environment.
Manipulation of provided training and development data is allowed. Task specific modifications of this rule are indicated in the individual task pages. Participants are not allowed to make subjective judgments of the evaluation data, nor to annotate it. The evaluation dataset cannot be used to train the submitted system. The use of statistics about the evaluation data in the decision making is also forbidden. Organizers of a challenge task should not participate in the task. Participation of other researchers from the institute that is (co-)organizing the task is allowed, provided that there is a clear arrangement that prevents them from receiving unfair advantage in the task. The arrangement should be documented in the technical report. By submitting their results, the participants give the rights to have their system output be evaluated and the results be published on the challenge web page. In order to support open science and make the challenge fair, withdrawing the submission is allowed only before the results are published. Once the evaluation results are published, withdrawal is not allowed. Other task specific rules may apply. These are indicated and highlighted in the individual task pages. Task organizers can exclude a submission if it violates the task rules. The system outputs that do not respect the challenge rules will be evaluated on request, but they will not be officially included in the challenge rankings. Organizers reserve the right to make changes to the rules and schedule of the challenge.
Information on use of external data is available in each task page.
Where allowed, use of external data must obey the following conditions:
The used external resource is clearly referenced and freely accessible to any other research group in the world. External data refers to public datasets or trained models. The dataset/models must be public and freely available before 1st of April 2022. Participants inform the organizers in advance about such data sources, so that all competitors know about them and have equal opportunity to use them; please send an email to the task coordinators; we will update the list of external datasets on the webpage accordingly. Once the evaluation set is published, the list of allowed external data resources is locked (no further external sources allowed).
Participants can choose to participate in only one task or multiple.
Official challenge submission consists of:
System output file (*.csv) Technical report explaining in sufficient detail the method (*.pdf) Metadata file (*.yaml)
Participants are requested to submit results in the required submission package. For submission, participants are allowed to train their system using any subset or complete set of the available development dataset.
Multiple system outputs can be submitted (maximum 4 per participant). If submitting multiple systems, the individual system output files should be packaged into same submission package. Please follow instructions carefully when forming the submission package.
The evaluation of the submitted results will be done by the organizers. Each task has its own metrics; for details please refer to the individual task pages.
The evaluation results will be announced online, and discussed at DCASE 2022 Workshop.
（https://dcase.community/workshop2022/index）