Login

Sampling my Head (HRTF)

Tips & tricks, working results, technical support

Sampling my Head (HRTF)

Postby Superfly76 » Wed Oct 17, 2012 6:41 am

I'm a former Nebula customer and am concidering purchasing Nebula 3 or 3 pro for this project.

I'd like to sample my HRTF (head related transfer function). I believe NAT will be up to the job but have a few questions.

I'll explain a little before jumping into my questions. I won't go into all the technical info but just give a broad explination of what and why I want to do this. Our auditory system consists of two imputs, our ears. Yet we are able to locate sounds spacially. We can hear when something is in front of us, behind us, next to us. When we hear a sound on our right side, the sound goes into our right ear, but also goes into our left ear a bit delayed. Our head, and the shape of our ears also alters the frequency of the sound arriving at both ears. Our brain is able to take that time and frequency information, and translate that into space. This happens automatically, our brain is cool that way.

Headphones are also a two channel system, like our ears. But when we listen on headphones, the soundstage is collapsed to the middle of our brain, between our ears. If a sound is panned from left to right that sound appears first to be right at the left ear, then travel through our skull and ultimately end up at our Right ear. This is a bit unnatural, and is not the way sound sounds in real life. When listening through speakers the sound fills the room and comes from around you. When listening through headphones, that sound is contained within the space between your ears. There have been various attemps to take the sound out of your head ranging from simple crosstalk and delay boxes, to hardware units costing several thousands of dollars. There are also DSP like Dolby Headphones and SRS WOW that attempt to do this by using their propiatary process and a generic HRTF.

The problem with off the shelf products is that each head and ear are different. My brain is used to processing spacialization through my unique ears using my * shape. Dolby headphone has worked the best for me, but even it sounds artificial and the appearace of an outside sound sometimes collapses back into my head.

What I want to do is sample my head including my ears, sampled in my room, on my monitor setup. This can be done using standard convolution, by placing a measurement mic into the entrance of each ear canal and sampling each speaker. My thought is that NAT and nebula may do a better job at this as it doesn't just take a static picture. With HRTF it isn't just time dependant, but it's also very much time and frequency dependant.

So I'd like to know if I can do this with Nebula. I want to create both a 2 channel program with me at the mix position, and a 5.1 surround program with me sitting in the middle of an optimal 5.1 system. With these programs I'd be able to put nebula on the mix buss and mix with headphones and have it sound just like I'm at the mix position of my studio.

So here are a few questions:

Could nebula do this?

Can nebula deal with surround sound like this? It would take the 6 channels of the 5.1 system, process them and sum to 2 channels for my headphones ( could sum seperately on my own later.

I'd also like to sample my over ear headphones with my measurement mics in my ear canals so that I can correct the sound of the headphones. This would be a seperate program as I could do one for each kind of headphone I wear. This would compensate for any coloration that the headphone is adding to the HRTF program.

Which version of nebula would I need to create these programs? Would Nebula 3 Free work?

I have friends who have done this with more standard static impulse response processors. I can attest that this works with Voxengo's Voxformer. I would assume that I can do this with nebula too and that the results would be better.

I'm currently abroad for the next few months, so won't be able to start the project for a while. I'm wanting to reaserch now and get everything in order so that I can get to work on this right away when getting home.
Superfly76
User Level 0
User Level 0
 
Posts: 6
Joined: Wed Oct 13, 2010 3:27 am

Re: Sampling my Head (HRTF)

Postby yr » Wed Oct 17, 2012 11:29 am

If I understand the question correctly, the answer is no- you can't do that with NAT/Nebula. I think what you describe would require true-stereo/surround support and cross-feed simulation. You could try the Externalizer plugin in Wavelab or the free VNOPhones to see if they help you mix more accurately using headphones.
Reuven | post-production & sound-design | scenography |
website | nebula presets
User avatar
yr
Beta Tester
Beta Tester
 
Posts: 438
Joined: Mon Mar 29, 2010 9:04 am
Location: Amsterdam

Re: Sampling my Head (HRTF)

Postby mathias » Thu Oct 18, 2012 3:07 am

there are a few other solutions, that could help you in headphonemixing.
i use ToneBoosters Isone Pro and made good experiences with it. most of the time it translates pretty well to my monitors.

there is another solution from 112db, it is called redline monitor.
i demoed it a while ago, but can only remember that isone pro did the job better for me.

mathias
system 1: windows 8 32 bit - samplitude prox, tracktion6, reaper
system 2: mac osx yosemite - reaper(32+64bit), tracktion6(32+64bit)

both systems on: macbook pro (late 2009), core 2 duo 3,06 ghz, 4 gb ram, graphic: nvidia geforce 9600M GT 512 MB
mathias
Expert
Expert
 
Posts: 2116
Joined: Wed Mar 31, 2010 12:25 am
Location: South-West Germany

Re: Sampling my Head (HRTF)

Postby Superfly76 » Thu Oct 18, 2012 10:49 am

Does nebula do true stereo processing? If it does, then that's all I need to do this.

Edit: I don't need true stereo processing. Each channel is going to be fed in mono without panning anyway, so traditional mono in/stereo out will work perfectly.
Last edited by Superfly76 on Thu Oct 18, 2012 10:57 am, edited 1 time in total.
Superfly76
User Level 0
User Level 0
 
Posts: 6
Joined: Wed Oct 13, 2010 3:27 am

Re: Sampling my Head (HRTF)

Postby Superfly76 » Thu Oct 18, 2012 10:54 am

mathias wrote:there are a few other solutions, that could help you in headphonemixing.
i use ToneBoosters Isone Pro and made good experiences with it. most of the time it translates pretty well to my monitors.

there is another solution from 112db, it is called redline monitor.
i demoed it a while ago, but can only remember that isone pro did the job better for me.

mathias


I own Isone Pro and have also tested Redline monitor. Neither of these work as well as Dolby Headphone. All three of these are a bit of a comprimize compared to measuring your own HRTF through convolution.

I think maybe I'm not expressing my idea and method clearly.

All I really need to know is if NAT will work for natural spaces like a convolution inpulse response deconvolver would. Can NAT capture real spaces with a speaker and a stereo measurement microphone? If it can do this, then it will do what I need.

I will document the project more clearly with pictures when I get around to it.
Superfly76
User Level 0
User Level 0
 
Posts: 6
Joined: Wed Oct 13, 2010 3:27 am

Re: Sampling my Head (HRTF)

Postby yr » Thu Oct 18, 2012 2:39 pm

I think that if you are trying to capture the acoustics of your room, standard convolution might be more effective. NAT has issues dealing with complex phase shifts, which are a given in this case.

If you are trying to capture the HRTF only, you will need good measurement equipment and an anechoic chamber. Otherwise you might be capturing your monitors+room too. If your purpose is to simulate sitting in front of your monitors (in your room) while working with headphones, you still need to consider cross-feed.
Reuven | post-production & sound-design | scenography |
website | nebula presets
User avatar
yr
Beta Tester
Beta Tester
 
Posts: 438
Joined: Mon Mar 29, 2010 9:04 am
Location: Amsterdam

Re: Sampling my Head (HRTF)

Postby Superfly76 » Thu Oct 18, 2012 4:16 pm

yr wrote:I think that if you are trying to capture the acoustics of your room, standard convolution might be more effective. NAT has issues dealing with complex phase shifts, which are a given in this case.

If you are trying to capture the HRTF only, you will need good measurement equipment and an anechoic chamber. Otherwise you might be capturing your monitors+room too. If your purpose is to simulate sitting in front of your monitors (in your room) while working with headphones, you still need to consider cross-feed.


I want to get my HRTF and my monitors and room too. I have a great sounding room with lots of acoustic treatment and great monitors as well. I want to capture it all. Thanks for your answer. I will try doing this with both nat and standard convolution and compare the two. Crossfeed is not needed as it will take place in the impulse itself. I'll be doing stereo inpulse responses with microphones in each ear. The crossfading itself will occur in the capture.

Could you elaborate on NAT's issues with phase shifts? Does NAT not do well with capturing real acoustic spaces? I did a search on the forums and see that there are at least 2 libraries that contain acoustic spaces.

I've been looking for material on NAT, but I can't really find much info on it. Is there any documentation or manual that I'm missing?

Thanks a ton guys.
Superfly76
User Level 0
User Level 0
 
Posts: 6
Joined: Wed Oct 13, 2010 3:27 am

Re: Sampling my Head (HRTF)

Postby yr » Thu Oct 18, 2012 5:05 pm

Hi, yes NAT documentation is limited.. it takes some patience to figure out the basics, but it's a great tool. What I've noticed is that when there are complex phase shifts (like with room acoustics) you end up with a "colored" preset- the frequency response is not accurate. I think Giancarlo explained it in a thread about IR re-sampling once, but I can't find it anymore.

You can't capture cross-feed with NAT. You will have to add some kind of simulation. If I understand you correctly, you need something like TB Isone, but with the option to load your own IR's. I think you might also need a way to control the "strength" of room coloration in your simulation.
Reuven | post-production & sound-design | scenography |
website | nebula presets
User avatar
yr
Beta Tester
Beta Tester
 
Posts: 438
Joined: Mon Mar 29, 2010 9:04 am
Location: Amsterdam

Re: Sampling my Head (HRTF)

Postby mathias » Fri Oct 19, 2012 2:00 am

for the truestereo-attempt, it could be a workaround, to make two programs and use two instances of nebula. one with only the left speaker running and one with the right speaker, while taking stereorecordings for both. in your daw you can then arrange the signalflow into two nebulainstances.

what you have to keep in mind, when using Nat is, that you will have to sit there a long time, when you want to sample your ears with a dynamic approach (means different runthroughs with different levels).

when you move your head slightly while sitting there, it could introduce errors in the resulting program.

i did some realroom-sampling myself and put it up for free here in the forum. search for "mabene" ... and take a listen. sometimes there are chirping artefacts when using these programs, which may come from the phasingissues yr pointed out.

and yes, it takes a while to wrap your head around Nat.
make some tests with simple nondynamic programs, to get an idea. many things are not obvious and you will have to find out yourself or ask giancarlo directly. specific questions are always easier to answer, so if you put those up here, you may get answers from people, that are sampling with Nat.

good luck for your project,
mathias
system 1: windows 8 32 bit - samplitude prox, tracktion6, reaper
system 2: mac osx yosemite - reaper(32+64bit), tracktion6(32+64bit)

both systems on: macbook pro (late 2009), core 2 duo 3,06 ghz, 4 gb ram, graphic: nvidia geforce 9600M GT 512 MB
mathias
Expert
Expert
 
Posts: 2116
Joined: Wed Mar 31, 2010 12:25 am
Location: South-West Germany

Re: Sampling my Head (HRTF)

Postby Superfly76 » Fri Oct 19, 2012 9:45 am

Thanks Mathias,

Yes, that's exactly what I'm planning to do. One instance for the Left Channel, and a second instance for the Right Channel.

To do surround, I'd need an instanse of Nebula for each channel. In a 5.1 setup, I'd have 5 seperate instances of Nebula. For the .1 channel (the sub channel), I'd pass that straight through with no room simulation, as bass frequencies are omnidirectional anyway.

When I get around to it, I'll try it with Nebula free, and the version of NAT that comes with it, and also try it with traditional convolution and compare.

Yes, I'm going to have to keep my head extremly still while doing this. I'm definitately going to use an intern to run the rig while I just sit there. I'm thinking of constructing some sort of jig that will hold my head perfectly still, while not interfering with the acoustics. Maybe something that will cradle the back of my head and my neck, with a strap going around the front of my head.

I'm going to use mics like this to take the measurements:

Image

Edit: The image was cropped and I can't seem to resize it in the forum so here's a link to see the full image inserted in someone's ears.

http://www.soundprofessionals.com/mas_assets/full/MS-TFB-2.jpg

I have the elements and will construct my own in ear measurement microphones. I've built a few mics and preamps in the past. The elements I have are measurement grade elements used in some earthworks mics.
Superfly76
User Level 0
User Level 0
 
Posts: 6
Joined: Wed Oct 13, 2010 3:27 am

Next

Return to Working with Nat

Who is online

Users browsing this forum: No registered users and 4 guests

cron