The Missing Measurements – Challenges of Measuring Immersive Audio
The University Atrium and Presentation Space that was Simulated in the Immersive Listening Room (Photo courtesy of Acentech)
Earlier this week I attended the AES Boston section meeting at Acentech’s state-of-the art facility, lured by the promise of a demonstration of their 3D acoustic listening simulation. I was not disappointed! After some drinks and conversation with other attendees, we broke into smaller groups to experience their immersive listening room. Although only the size of a large conference room, this room simulates much larger spaces such as lecture halls, restaurants and offices.
We listened to a simulation of a University atrium and presentation space. There was a dining space on a balcony above it, and offices and classrooms above and behind that. We heard how it would sound both with and without diners during a presentation, and if you closed your eyes it was easy to believe that you were actually there. This impressive auralization was created by first recording a presenter in an anechoic environment, then applying sophisticated predictive models to apply reflections, reverberations and background noise to account for the room configuration and construction materials. It was then reproduced in the immersive room on an Ambisonic playback system. This model permits not only evaluation of architectural features, but also experimentation with different materials for walls, windows, furniture etc. Architects, designers and sound consultants can evaluate and adjust the acoustics of a room at the design stage, ensuring acoustic perfection in the finished product.
What I found surprising was the lack of feedback to the model based on real measurements. It seems that there’s a prime opportunity to make actual measurements in the room once it is built, and compare this real-life data to the model. This was also apparent when I spoke to another attendee, a recording engineer, about mixing in Dolby Atmos. We discussed how he could master something until it was perfect on his Dolby Atmos system, but there was no way of knowing whether other listener’s Dolby Atmos systems would reproduce the sound in the same way he intended.
Spatial audio measurements are still in their infancy, and they are a complex combination of perceptual measurements such as localization, intertwined with room acoustics. Until we devise reliable and accurate ways of measuring spatial audio, we have no way of measuring room simulations to compare them to actual measurements, and no way of comparing the fidelity of various Dolby Atmos playback systems. This is a source of frustration for many product design engineers who are accustomed to having good measurement data to drive their product development.
Recently, I’ve been investigating how a binaural head perceives virtual source location and envelopment, and we’ve developed some rudimentary ways to measure sound localization and envelopment using interaural level difference and interaural cross-correlation measurements with a head and torso simulator. What are your thoughts on spatial audio measurement and what measurement techniques and metrics you are working on? Let me know in the comments and please reach out for me if you’d prefer an offline discussion.
Learn more about Acentech.
Learn more about stereo soundfield measurements for quantifying sound localization and envelopment.
