1
$\begingroup$

This question follows from Modelling propagation of sound wave by particle simulation

This last fortnight I have started experimenting with Binaural audio; it really blows everything else out of the water.

The main reason it hasn't taken off, I think, is that every individual needs to get their own HRTF calculated, which consists of sitting in a soundproof room with microphones in your ear canals, while something makes popping noises at thousands of locations around you.

I'm only aware of one service provider that may calculate your HRTF: http://www.physiol.usyd.edu.au/~simonc/hrtf_rec.htm and they are in Australia!

I'm wondering whether a more practical method may emerge, which would involve taking a 3-D scan of one's head (maybe using something like http://www.david-3d.com/) and shipping it off for a heavy dose of distributed computing which would return HRTFs for that individual.

I wonder if we may see a day where people have their own HRTF data stored in the cloud, and can enjoy binaural sound on their iPods.

How far away is such a technology? And other any other contenders for measuring HRTF?

π

PS could we possibly have some more tags like "binaural", "hrtf/hrir"

|improve this question
$\endgroup$
  • 1
    $\begingroup$ Yay, HRTF tag. Thx Peter :) $\endgroup$ – P i Oct 23 '13 at 22:09
  • $\begingroup$ I think the only way I can contact you is by answering here as I can't even add comments yet. I'd be really interested in taking a read of your paper, I might reference it in a PHD proposal I'm working on. Thanks. $\endgroup$ – conor Nov 10 '14 at 13:32
3
$\begingroup$

There are various school of thoughts about this. A German guy named Klaus Genuit came up with a parametric model of HRTFs that's based on the actual physcial geometry of head, shoulder, pinna, etc. He actually now runs a company that craeates binaural products and measurement gear, see for example http://www.head-acoustics.de/downloads/publications/binaural_technology/DAGA_99_Physical_Modeling_of_individual_HRTFs.pdf

Based on this work you could potentially create individual HRTFs using a some 3D scanning scheme. I do vaguely remember a company coming doing exactly that for a bunch of people at my employer at the time. I didn't think the HRTFs that came out that way were particularly good but I didn't do a lot of work with them either.

Individualized HRTFs only help to a certain extent. Let's say you want to record a concert. You can do this binaurally using a dummy head. However, there is currently no way to convert this recording to your individual HRTFs after the fact, even if you have them.

|improve this answer
$\endgroup$
  • $\begingroup$ Yes, a concert wouldn't work. But if you could record each instrument in its own separate channel, then you could combine the channels separately for each individual listener using their own HRTF data. Fascinating that you have worked in this field! Many thanks for sharing your knowledge! $\endgroup$ – P i Oct 24 '13 at 13:31
  • $\begingroup$ Even if you have individual instruments, the HRTFs are anechoic. You would need a full blow binaural room simulation of the performance space to get this "indivualized". That actually was my Ph.D. thesis :-) $\endgroup$ – Hilmar Oct 24 '13 at 13:51
  • $\begingroup$ id love to get a copy of your thesis himar! where can it be downloaded? thanks! justus $\endgroup$ – user10179 Jun 11 '14 at 9:48
  • $\begingroup$ It's been a while so I don't think it can be downloaded anywhere and it's in German. See biblio.com/book/… $\endgroup$ – Hilmar Jun 11 '14 at 18:38
1
$\begingroup$

In the end I managed to do this myself, working in collaboration with http://ir-ltd.net/ to get the scan of my head, Blender to refine the mesh and position a cloud of microphone points, http://www.waveller.com/Waveller_Cloud/ to compute the frequency responses for these points, and finally some Python/NumPy scripting to convert these into impulse responses.

We have collaborated on a joint paper which is being presented at the forthcoming AES conference.

Please leave a message after July '14 if you would like me to link the paper. There is currently no link as it is still in draft form.

I'm using the results of the simulation and I'm happy with the results.

EDIT: http://www.aes.org/e-lib/browse.cfm?elib=17365

|improve this answer
$\endgroup$
  • $\begingroup$ i've only now seen this question/thread. i'm happy to defer any technical fact and detail to @Hilmar's experience and research. there were two "caveats" (or the antonym of "caveat") that sometimes softened the problem of "everybody's ears are different": 1. the inter-aural difference is the most salient 2. there were some basic commonalities to our pinnae that results in two major notches in the frequency response of the HRTFs. as the azimuth angle increased, i think both notches went up in frequency sorta together and as the zenith angle increased, the spacing of the two notches increased. $\endgroup$ – robert bristow-johnson Jun 13 '14 at 1:03
  • $\begingroup$ How come I get a downvote the day after providing a link to the paper? Step up, downvoter. Let's hear what you have to say! $\endgroup$ – P i Nov 11 '14 at 2:36
  • $\begingroup$ i'm not the downvoter (in fact, i'll cancel the downvote), but i wouldn't mind a free copy of the paper, if i may have one. i'm at rbj@audioimagination.com . $\endgroup$ – robert bristow-johnson Nov 11 '14 at 2:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.