CES 2017: DIRAC VR Dynamic 3D Audio Rendering System

Dirac is well known for their room correction software and more recently among headphone enthusiasts for their HD Sound headphone compensation software that EQs headphones. At this year's CES, the company unveiled their new DIRAC VR, a dynamic 3D audio rendering platform.

Man, is there a lot to talk about here. First of all, I think we're still a few years away from really convincing 3D audio on headphones. There are still too many unknowns that need sorting through—personalized HRTFs; better headphone transducers; how to economically get all number crunching with head movement in order to commercialize it. To my mind, any solution out there right now will be largely incomplete, bleeding edge technologies. Previous to CES, that idea tainted my expectations; I really didn't think I'd hear anything really impressive when it came to 3D audio. I was wrong.

Mathias Johnsson, CEO, explained their new DIRAC VR system in quite some detail. Outwardly, it was just a laptop and some electronics connected to a Sennheiser HD 650 with a head tracker mounted in the headband. Inwardly, meaning all that software running on the laptop, abundant numbers were being crunched. Room acoustics, head movements, and HRTFs were continually being combined to filter the incoming audio signal with intent to fool you.

What sets DIRAC apart is not the number crunching itself, but the HRTF database they start with. Very simply, most HRTFs are acquired with sound coming at various angles towards a standing person who is facing forward. The problem, DIRAC claims, is that in a VR goggle application the person is moving their head in all different directions relative to their torso. HTRFs with your head in its normal position are quite different than HRTFs with your head tilted and turned in various directions. Holy smoke, that makes so much sense. Johnsson sent me some proprietary internal studies showing the changes to HRTF with oddly angled headpositions and it was quite apparent they have a lot of effect.

How good was the demo? I'd say quite good. Very often I hear what I can only call a squeaky character in the treble. I chalk it up to the fact that a lot of the HRTF cues are in the 4kHz-10kHz range, and there are probably a lot of DSP artifacts showing up there. I didn't hear that as much in this case, I felt the DIRAC system was just more natural sounding than many I've heard. Maybe it was due to the fact that they chose a good headphone to start with, but I liked what I heard.

On the down side, I still didn't get tight forward localization—which is the hardest thing to do. As is usually the case, when I turned my head from left to right, the image rose when directly in front of me. But I have to say, I've never heard a generic HRTF system get frontal localization right, and I felt the sense of space otherwise with the DIRAC VR pretty good. Off axis sound sources were stable in seemed well placed in space.

Don't know if DIRAC will be one of the lucky, or just damned good, companies to make it from the bleeding edge to the leading edge in virtual audio, but I'm going to keep me ear on them.

COMMENTS
Richter Di's picture

Was quite surprised that something which is on the market since some years, like the Smyth Reseach A8, hasn't been mentioned here.
I heard a Demo of their new A16 on Paris son et image and I was very impressed. I had allready pre-ordered mine via Kickstarter.
Would be highly interested to hear more about you Ossic X experience as I have also pledged for them.
So far I asume both will arrive at the same time April/May.

Hi-Reality's picture

Tyll, how was DIRAC VR's frontal out-of-head performance compared to Smyth Research A8 or A16?

Babak
Hi-Reality Project
www.hi-reality.com

castleofargh's picture

pretty excited by all the guys starting to look into head related anything and head tracking.
still can't get why VR googles had to come up, for audio people to pretend like they care, but better late than never.

about the center image going up, I've tried to get around it with crossfeed alone and failed. using stock 30° azimuth HRIR that kind of worked for me made it all much better but still far from actual speaker experience, some tiny bit of room reverb helps, having actual speakers to look at, and of course head tracking. with them all it's not bad after a little tweaking here and there. but I agree with your comment, custom HRTF is probably the only option for people without the absolutely average body.

JE's picture

Why Tyll completely ignores Smyth Realiser A8/A16?

Jackkevin's picture

It's what isolates reality, experienced for all intents and purposes, from an undeniable representation of such. Until today, the constraints of the 3D sound innovation utilized as a part of VR/AR programming have kept this genuine inundation from being made. Please Visit Our Website EssayThinker

raina852's picture
britney's picture
suji99's picture

Super website with a great sharing and amazing stories is ur web. Women's Day 2017 WhatsApp Dp please keep doing what u do now.. Holi WhatsApp Dp
thanks to you, It's a great sharing and really informative.. will keep following ur web. Holi Shayari in Hindi
thanks to you and ur great website.

deannawilliam's picture

Deep down, which means all that product running on the portable workstation, inexhaustible numbers were being crunched. Room acoustics, head developments, and HRTFs were persistently being joined to channel the approaching sound flag with goal to trick you. Agents of Shield Jacket Buymoviejackets