T.H.E. Show Newport 2014: Darin Fong Audio "Out Of Your Head" Virtual Audio Software

Getting headphones to sound like speakers outside your head is not easy...not easy at all. But on short listen under show conditions, I found myself favorably impressed with Darin Fong Audio's "Out Of Your Head" software.

Basically, you play your music or movies just as you normally would. The "Out Of Your Head" software is installed as a virtual sound card. Audio from your playback app goes through "Out Of Your Head" where some digital signal processing (DSP) filters add the sound of the speakers and room. There are a number of systems like this that use filters made from mathematical models of speakers in a room. "Out Of Your Head" on the other hand, uses models made from real measurements in real rooms using real heads. Darin's usually, but other folks have done the measurements as well.

The procedure is that Darin finds a room worthy of measurement; he puts special microphones into his ears at the entrance of the ear canal and sits at the focus of the sound system; a frequency sweep is made and recorded from each speaker; and then a whole bunch of math is done. The result is a specific DSP filter synthesizing the signal relationship between each speaker in the room and both your left and right ears. The "Out Of Your Head" software will turn on and off speakers as needed depending on the media your playing.

With headphone virtualization systems that are mathematical models the user is able to adjust things like room size and speaker angle. Because "Out Of Your Head" is measurement based, you can't adjust these parameters. You can however measure multiple rooms, large and small, using various high-end speakers, which is exactly what Darin has done. The cool thing about "Out Of Your Head" is the DSP filters derived for headphone localization of speakers in the surround experience are done naturally not synthetically. Darin says that makes a huge difference. I'm going to have to spend some time with it, but what I heard sounded pretty good. Currently 21 systems have been measured, descriptions can be found on the FAQ page.

One thing I'll forewarn you about: With these types of virtualization systems, both synthetic and natural, it will often take a bit of time for your head to figure out what it's hearing well enough for good localization. In other words, sometimes these systems will sound weird to you at first, but after a few good listening sessions start to get really convincing. Your brain has to have a little time to build its filters. So if you try the demo, give it some time.

"Out Of Your Head" software currently requires windows 7 or higher and 64-bit operating system. Mac, android, and iOS versions are expected in future.

Check out Darin Fong Audio here.

COMMENTS
jherbert's picture

I managed to downloaded the app some months ago and could not get it up and running. Unfortunately there was no support at all from Darin. Great concept, somewhat underwelming execution.

This site now is the slowest I can remember. Maybe its run on https only. Will not try to contact Darin again, but maybe he reads it here. I believe the product deserves better support.

Tyll Hertsens's picture
It's a young outfit, I'd give him another chance.
HiFiArt's picture

Hi, Tyl,

What is Mr. Fong's headphone setup? Headphones, amp and stand? :)

I'm listening to the demo now, https://fongaudio.com/demo/, and the surround decoding parts are interesting.

I don't like the speaker / room simulation, but that's a personal preference.

Thanks!

Darin Fong Audio's picture

I had 7 different listening stations with all kinds of headphones and amps and DACs just to show that Out Of Your Head works with anything.

The station in the video was:

  • Windows 8.1 desktop computer running JRiver Media Center 19
  • Vostok Sound ES-21 prototype electrostatic headphone amp and DSD DAC combination.
  • Stax SR-009 electrostatic headphones
  • Brainwavz acrylic headphone stand

Thank you for trying the pre-rendered demo files.

I understand that for 2 channel listening, many headphone-philes want untouched, bit-perfect audio fed straight to their headphones. Many people like the sound of headphones and aren't bothered by the sound being "in their head".

However, as you said, even for those people, watching/listening to a movie in full 7.1 surround in a high end home theater system using just headphones is a lot of fun. Most people don't have access to a high end 7.1 system but being able to watch a movie in one with just your headphones is very attractive to even the most purist headphone enthusiasts.

Darin Fong Audio's picture

Sorry about the slow server right now. I am working with my VPS provider to figure out why it's running slow. However I was able to test the trial download and it's working.

As far as support goes, I pride myself in responding to support requests within minutes, not days, even late at night or on the weekends, etc. I don't know your e-mail you used to contact me, but please reach out again and I am sure we can get you up and running. It's possible I never received your support request e-mail for some reason.

I can give you countless examples of support issues being resolved very quickly. Alternately I am happy to call or Skype or remote control to resolve any problems quickly.

Thanks for your post.

-Darin Fong

jherbert's picture

Darin sent me a link to a different server, and instead of crashing the download went through in less than two minutes. I had to switch jrivers output to Direct Sound from WASAPI.

Zapping through the various setups I have to say that this is amazing stuff.

jherbert's picture

I surely will, because this is really interesting stuff. Sort of Smyth Realizer without the hazzle - great stuff. Maybe I will get a chance to hear it working the other day. Until now, no way to download the demo. I tried chrome and ie several times.

Darin Fong Audio's picture

If you are still having trouble downloading the trial on our website, please e-mail directly at info@fongaudio.com and I can give you a trial download link. -Darin

jherbert's picture

I surely will, because this is really interesting stuff. Sort of Smyth Realizer without the hazzle - great stuff. Maybe I will get a chance to hear it working the other day. Until now, no way to download the demo. I tried chrome and ie several times.

Hjelmevold's picture

I want to comment on two of the terms used in the article:

1) "Whole bunch of math" is probably a Fourier transform. In the video, it is inferred that doing a frequency sweep only results in gathering frequency data. But there is time data in such a frequency sweep as well, and this is why you do a Fourier transform - to convert a frequency sweep into an impulse response. FFTs are used all the time in DSP, and most low-level audio APIs that I've seen have built-in functions for performing FFTs.

2) The use of a "DSP filter" is from what I understand in this article actually a convolution algorithm, as made famous by the Altiverb reverb plugin in the early 2000s. Explained in simple terms: With convolution you apply an impulse response onto every single sample of the incoming PCM audio, which shapes the sound in time, amplitude, and therefore also frequency. There could be some filters involved in the process, but they're not the defining part of the algorithm.

But don't misunderstand me. Even if the concept is not new, I'm not dismissing this product. It's not so much about the technology, as it is about all the time and effort spent doing all these measurements. And there might be some interesting algorithms dealing with how surround sound is downmixed to binaural stereo, combining all those impulse responses into a True Stereo convolution.

I've used the Focusrite VRM Box (a similar product) a number of times to do mix translation referencing, and I thought it worked pretty well, so it's very likely that I'll check out OOYH quite soon.

ultrabike's picture

I dunno how Darin does it exactly, but what I heard was pretty good IMO! Specially with movies.

Now...

1) "Whole bunch of math" may be many things, including or excluding Fourier Transform.

2) More than an algorithm, convolution is a linear operator, usually performed between an input signal (electrical or acoustical or whatever) and a system's (headphone, analog circuit, digital filter...) impulse response. Things proly get interesting when 2 or more dimension are involved, and when one has all sorts of cross channels.

The concept might not be new, but the implementation might. I don't know the details.

Hekeli's picture

Come on, it's no secret Darin owns Realiser and more than likely uses it to do the initial measurements, nothing wrong or illegal with that (but I guess he likes to dodge the question just in case).

It's easy to capture the impulse responses from Realiser PRIRs, all this is described in detailed howtos and even lots of impulses are already shared in the head-fi threads.

Hjelmevold's picture

You could make your own convolution-based stereo speaker simulation quite easily, provided that you have binaural microphones (either in-ear such as the SoundMan OKM or a dummy head such as the Neumann KU100). Then, using room measurement software such as RoomEQWizard, record a four channel (True Stereo) impulse response of the speakers in your room: First the left speaker into both ears, and then the right speaker into both ears. Drag and drop the impulse response audio files direcly into Altiverb or similar audio convolution software, and enjoy the effect. But you need a lot of patience and attention to detail in order to achieve a decent signal-to-noise ratio on the impulse response.

And then there's the final link in the chain: My experience tells me that headphones that follow the Harman Headphone Target Response Curve, end up sounding too bassy and dark when used with the Focusrite VRM Box, while headphones with more energy in the presence region (2kHz-4kHz) sound closer to how the speakers would sound in real life. I suspect that similar considerations would have to be made when choosing headphones to use with OOYH. This is probably where the Smyth Realiser has the upper hand, since the headphones are actually calibrated to your ears.

A more general note:
One limitation with traditional convolution-based binaural speaker simulation, is that while you can get quite accurate time/amplitude/frequency results, non-linear attributes such as distortion do not transfer well using this method. But there are some interesting alternatives to traditional convolution; I would be very intrigued if someone had the time and patience to do measurements similar to those done for OOYH using the Nebula Application Tool by Acustica Audio - that could possibly bring binaural speaker simulation to a completely new level.

LibiSC's picture

You really need more bandwith... your product was featured but you couldn't handle the success :=)

Darin Fong Audio's picture

Yes, Murphy seems to be at work...

The day Tyll put up this post, my VPS provider that hosts my site and my store had two major systems go down due to drive failures and then they are currently under a Denial Of Service (DDOS) attack on them which has been slowing everything WAY down. I am surprised it's working at all right now. :-(

Yes, I know I need to have a backup VPS server in cases like this, but this is the first time this has happened with this ISP.

Sorry about the slow server. I am told they are working on it but I have no idea when it's going to get fixed. Please bear with us!

Tyll Hertsens's picture
...sorry about your wallet, Darin.
X