Task 2: A possible cheating method, using100 points HRTF for higher upsampling factor

[Chin-Yun Yu (Joey)]

Hi challenge organisers,

As the title says, it's possible to hack the challenge in the levels of upsampling from 19, 5, and 3 directions by copying the 100 directions HRTF directly, as they're the same subject. 

Will there be any verification step to detect this in the final evaluation?

For example, explicitly checking whether the HRTFs at these 100 directions are precisely equal to the ground truth using np.allclose.

Just want to make sure the challenge runs fairly.

Best,

Chin-Yun

[Aidan Hogg]

Hi Chin-Yun

Thank you for highlighting this so that we can clarify.

I can reassure you that the 12 sofa files are from 12 different participants. That is to say, subject 1, for example, is a different participant for every sparsity level (3, 5, 19, 100). This makes it impossible to cheat in the way you have suggested, and if you did, you would likely massively degrade your performance.  

Thank you again for the question. Let me know if I have misunderstood anything.

Cheers,

Aidan

---------------------------------------------------------

Dr Aidan Hogg

Lecturer at Queen Mary University of London

Honorary Research Associate at Imperial College London

Centre for Digital Music 

Electronic Engineering and Computer Science

Queen Mary University of London

327 Mile End Road, London

E1 4NS, U.K

Email: a.h...@qmul.ac.uk

Personal Website: aidanhogg.uk

QMUL Group Website: c4dm.eecs.qmul.ac.uk

QMUL Website: qmul.ac.uk/eecs/people/profiles/aidanhogg

Imperial Group Website: axdesign.co.uk

Imperial Website: imperial.ac.uk/people/a.hogg

[Chin-Yun Yu (Joey)]

Hi Aidan,

I thought the "subject" in the task description meant "participant".

If it's indeed from 12 different participants then all good.

It would be better if this is clearly explained in the description.

Cheers,

Chin-Yun

Aidan Hogg 在 2024年6月27日 星期四上午9:46:04 [UTC+2] 的信中寫道：