Question about ITDs on datasets used in Task1

[Yoshiki Masuyama]

Hi LAP challenge organizer and Task 1 participants,


I'm working on training an HRTF modeling system on multiple HRTF datasets inspired by Task 1 towards submission for the special issue on IEEE-OJSP.

When I compute ITDs for the datasets used in Task1, WiDESPREaD shows different behaviour compared with the other datasets.
The attached figure shows ITDs on the horizontal plane, where ITDs for 10 subjects are averaged for each dataset.

This could be because WiDESPREaD focuses on "Pinna"-related transfer functions, and their simulation does not consider the head shadowing effect.

Did participants of Task 1 use any techniques to harmonize this difference on ITDs? Is it suitable to include WiDESPREaD in the training of a system for "Head"-related transfer functions?

Any comments or suggestions are welcome.

Kind regards,

Yoshiki Masuyama

[Chin-Yun Yu (Joey)]

Hi Yoshiki,

I will probably get the minimum-phase widespread PRTFs and convolve them with a simple rigid sphere model so the ITDs can be at least within a normal range.
A better approach is to merge the pinna STL files with some dummy head STLs and rerun the numerical simulation.
Just my 2 cents. Might not make sense at all. Good luck with your submission.

Cheers,

Chin-Yun

Yoshiki Masuyama 在 2025年1月17日 星期五晚上9:13:49 [UTC] 的信中寫道：

[Yoshiki Masuyama]

Hi Chin-Yun,

Thank you for your advice!

Both of your suggestions sound good to me while I couldn't get details of the simulation process of WiDESPREaD yet.

Kind regards,

Yoshiki Masuyama

2025年2月1日(土) 14:53 Chin-Yun Yu (Joey) <ya7...@gmail.com>:

--
The IEEE Signal Processing Society is sponsoring the Listener Acoustic Personalisation Challenge.
---
You received this message because you are subscribed to the Google Groups "SONICOM LAP Challenge" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sonicom-lap-chal...@googlegroups.com.
To view this discussion, visit https://groups.google.com/d/msgid/sonicom-lap-challenge/1eaf5450-7572-4bfb-b4a2-1726b5590456n%40googlegroups.com.