Procedure

Registration

Please register for the challenge using the following google form.

Mailing list

To follow everything about the challenge, please subscribe to the SPEAR Challenge google group. Information about SPEAR (milestone, major updates…) will be communicated through this mailing list. All information will also be made available on the website.

Submission

Required files

Processed Audio

Submissions must consist of processed audio from the Evaluation set sampled at 48kHz. If the processing was done at a lower sampling rate it is expected that the submitted audio is upsampled at the required sampling rate.

Model Description

In addition to the processed audio, a short description of the submitted model is required. While it is not expected to be the length of a conference article, it should allow other participants to understand the method used.

It should be a minimum of 2 pages long. A LateX template can be downloaded here.

Upload

A OneDrive Business Shared folder invite will be send to registered participants before the end date of the challenge. No registration will be needed. Please be aware that the upload may take some time. If you encounter any issue with the submission process, please contact the organising committee.

In order for your submission to be validated and to help the organising committee, please follow the structure below:

┬── Team_## (directory): Contains all relevant information for SPEAR
│   ├── model-description.pdf (file): PDF file describing the model
│   │      used in this submission. Based on the available template.
│   └── Submission_# (directory): Subfolder if multiple submisions
│      │   are made from the same team
│      │  
│      └── D#_S##_M##_ID#.wav (files): Binaural enhanced audio
│         Sampled 48kHz, 32 bits float
│
└── README.md (file): Any additional information you want to communicate to the organising team