|
Dear colleagues,
we are conducting a study on the use of head-related transfer functions (HRTFs) in industry and research.
If you employ HRTFs in your research, products, production workflows, or elsewhere, we would highly appreciate five minutes of your time to respond -- if you haven't already -- to a few questions: https://forms.gle/9JwH6kjQC3BwXySo7
Please note that the survey is available until Feb. 11th 2026.
The study is conducted within the context of the NEMO initiative, which explores the prospect of identifying a default HRTF set that can be used across applications when no individualization is employed. Potentially, this would foster long-term adaptation.
One important goal of this study is to determine whether de facto defaults already exist, as in sets that are more widely used than others.
Additionally, you find the summary of the previously held online workshop on the same topic below.
Best Regards,
Nils Meyer-Kahlen, Pedro Lladó, Katharina Pollack, Fabian Brinkmann
---------------------------------------------------------------------------------------------
The first NEMO online Workshop took place on 1st of October 2025 with more than 50 participants from academia and industry.
The NEMO core team, comprising Pedro Lladó, Katharina Pollack, Nils Meyer-Kahlen, and Fabian Brinkmann, introduced the concept of a default HRTF set that could serve as a solution when an individual HRTF is unavailable. The primary motivation stems from
research on HRTF adaptation, which demonstrates that long-term exposure, and particularly training, can enhance localization ability with a non-individual HRTF. Deciding on a default set used across applications could therefore foster long-term adaptation.
Apart from potentially improving the quality of experience through such adaptation, a widely accepted default HRTF would have several potential immediate benefits. These include making it easier for companies and artists to optimize their work for common playback
chains and facilitating easier comparison of results across different research studies and projects.
KEYNOTE
Marc Schönwiesner (Leipzig University) talked about Adaptation and Brain Plasticity in Spatial Hearing. He provided examples of adaptation to altered ITD and spectral localization cues, along with correlates of this plasticity at the neural level. He emphasized
that (a) adaptation to a new HRTF set is persistent, i.e., the learned set is available to the listener even after a longer period without exposure to the set, (b) there is no after effect, i.e., the localization performance with listener’s own HRTF set is
not affected, and (c) results suggest a one-to-many mapping, in which multiple spectral profiles can be associated with a single location in the space at the same time.
JOINT DISCUSSION
A joint discussion took place in three separate breakout rooms after the keynote. The discussion aimed to foster a community effort in finding and establishing the NEMO default HRTF set. This long-term process involves suggesting candidate HRTF sets and
evaluating them through technical and perceptual testing. It was highlighted that external contributions in the form of scientific or data publication are possible for all steps in the process.
DISCUSSION SUMMARY
A discussion began regarding whether a de facto standard HRTF (such as KEMAR or KU100) already exists that could be directly adopted as the NEMO default set. We are currently investigating this question using a survey. In any case, participants agreed
that potential candidates for NEMO should be compared to existing sets that are commonly used. Multiple criteria should be considered when evaluating potential NEMO HRTF sets, including coloration, localization, externalization, and others, both before and
after adaptation, as well as adaptation speed. These criteria should be carefully weighted according to their importance, and it might be that not all criteria can be met by a single HRTF set. Also, the requirements might be different in different applications,
from music listening to gaming.
It was also discussed whether adaptation, as indicated by increased localization performance,
|