Immersive Audio Panel
April 6, 2013, National Association of Broadcasters, Las Vegas—A panel moderated by Tom Scott from OnStream Media considered the issues for more immersive audio. Panel members included Jeff Levison from Iosono, Stuart Bowling from Dolby Labs, Wilfried Van Baelen from Auro Technologies, John Kellogg from DTS, and Kimio Hamasaki from NHK.
The audio industry has changed from monophonic to stereophonic and other formats over time. In cinema, the number of channels followed the audio and the changing technologies. The progression implies we will eventually be at an infinite number of channels. The reality is that math and tools are increasing the quality of the channels, making the multi-channel presentations deeply immersive.
Levison agreed that the tools have progressed and grown in capabilities. Now wavefield synthesis and object-based multiplexers can supply the services, software, hardware, and other IP to make an all encompassing sound. It is not just more speakers, but other areas that are getting better. Calibration tools for performance and research tool functions complement listening tests. A 300 channel system now takes only 45 minutes to set up, and less than one minute for the filler generation.
Bowling noted that the mix of channels is now a set of objects and metadata on the mix, so the full range of sound and a 12th octave equalizer can be concatenated with the room configuration files to set up a room. Sound is more than half of the experience.
Sound placement is subliminal, so you have to work with tools and the hardware to get good results. The distribution is from a single DCP and the DCP package in the auxiliary track. Exhibitions need hardware and tools to update the theater. The costs for this upgrade include a design review. An SDK is available for the tools
Van Baelen mentioned that the tools have to work in a unified workflow. The increasing number of audio formats increases the workload, but a unified workflow enables one mix to feed out to multiple formats in an object-based standard. Hardware plug-ins and software mix engines can encode and create a final mixdown. Object-based tools facilitate metering and visualization. The capabilities increase the artistic challenges.
Kellogg suggested that SDS is migrating towards DTS. Open standards are creating object-based solutions that cut costs, enable more formats, and other capabilities. Objects replace channels for flexible mapping and support both 2-D and 3-D presentations. The tools still need a common language as platforms and render are different.
There is a need for an open standard for a single deliverable. Multi-dimensional audio needs better tools, specifications, and deliverables for spatial render. The render must support objects that include a subset of channels and objects by line or channel. The tools have to export to any number of speakers and in any configuration. A good standard is adaptable, open, and amendable to deliver and render solutions.
Hamasaki claimed that 22.2 is all the audio dimensions you need. The increases in viewing angles is saturated at about 100 °, so the audio has to match. Their Super HI Vision 8K plus 22.2 sound reproduces the sound field to match the view. Synthesis and transfer functions direct the synthesis to develop the multi-channel audio. The sound has to cover the full screen area and give a wide area sweet spot.
Psychoacoustic studies have shown that the number of channels and their placement determines a minimum channel count that is now in SMPTE 2036-2. In production, they use a 45cm spherical mike to capture the sound field and, in post, do the pan, mix, etc. through visualization plus reverberation. They are working on conversion to headphones. Home 22.2 theater systems are possible in a larger room and could be implemented with minimum visible speakers via a speaker array surrounding the screen.
