9.1 Auro-3D encoded Blu-ray Disc to be released in 2012

Red Tails was apparently mixed in 11.1

Only a handful of theaters in the U.S. (none of them anywhere near me) were actually able to present it in 11.1

Having seen Redtails in an 11.1 audio mix in Calabasas, CA, it was hard to get excited at all. In fact, the trailers that preceded the film sounded awesome because they were so immersive.

As to this new Auro-3D mix in 9.1 for home, ho-hum. The actual BR media would have to present something that was extraordinary because with the a well recorded and mixed CD, 9.2 audio with heights is hard to beat.

The problem is the compatibility between the Dolby Atmos layout and the Auro-3D layout. If, Dolby (which adds top rows only), wants to be compatible with Auro-3D, it will have to add the height layer (at 30° angles) to their setup. It is this heigh layer around the listener, that creates the natural immersive 3D sound.

Also, there is a natural compatibility between all Auro-3D formats (Auro 9.1, 10.1, 11.1) which are controlled during the mix in the Auro-Codec Plug-in.
At home, people don't need the ceiling speakers above their head since in smaller rooms, the height channels at 30° around the listener are key to achieve a full immersive experience at home, similar as an Auro 11.1 system in a large cinema theater.

It is already supported by Atmos. That is the beauty of object-based content. The playback processor is made aware of every speaker's position in the theater, and it automatically maps the source to them optimally, such as with VBAP. If there is a subjective difference in the spatial perception between Atmos and Auro3D, that can come down to three main causes:

1) The speakers are not optimally positioned or in sufficient numbers to create the best subjective effect.

2) The mixing engineers responded to the specific attributes of the system at hand, emphasizing its unique capabilities, resulting in a soundtrack that does not necessarily translate the intended mix optimally to other "3D" speaker arrangements.

3) Special psychoacoustic processing is added in one system's processor that is not emulated in the other's. (Such as Ambisonics or Trinnov remapping, but the possibilities are much broader.)
Object-based playback is inherently compatible with any speaker configuration, and unlike normal downmixing, direct rendering avoids unintended buildup of correlated sounds when electrically downmixed, thus preserving the subtle balances in the mix.
I have always found the putting speakers on the medial plane (other than in front of the listener) is a bad idea, whether directly behind (6.1) or overhead. Directionality is ambiguous (subject to reversals). An array of, say, 4 speakers high on the sides is much more effective. They may have to inclined more than 30° however to effectively distinguish their value relative to surrounds that are already somewhat elevated. Perhaps 45°.

The problem with all of the above is that this is only true for simple objects. The key behind the two separate layers is to use them for ambient sounds where the interaction and correlation between all the speakers is crucial. This is what creates a true 3D listening experience. It is not only about being to pinpoint sounds from a certain direction, it is even more about being enveloped by the sound, like a blanket. This not possible with object-based sound and is actually the most probable reason why Dolby Atmos is hybrid. But there the 'ambient' layer is likely to be limited to the horizontal 7.1 plane only, so not true 3D.
As said before, this might be true for simple, monophonic objects, but not for ambient sounds. The downmix issues you mention are actually taken care of by the Auro codec because the artist has full control over how this 'downmix' needs to happen.
Actually not. Inclining them above 30º indeed makes them more individually noticeable, but that is completely not the point. Experiments have shown that at 30º individual sounds can be distinguished, but that this also generates a sense of envelopment for ambient sounds where the speakers seem to dissapear. The latter does not happen aymore with speakers at 45º where there is a disconnection between the lower and height layers, similar to the font-back disconnection in 5.1 surround.

I think you have made an assumption that objects must be mono. That is like saying channels are mono. They are, but only until you use several of them together. There is no sound that can be delivered by a channel that cannot be delivered by an object equally as well.
Having full control implies that the mix engineer must review each downmix permutation and potentially make some adjustments because the correlation issues are not automatically handled as they are with objects. That affects the production time/cost element.
If the surrounds are some 20º -ish elevation (as in non-ITU home theaters or commercial cinemas) the effect of adding another layer at 30º is rather diminished. No?
Edited by Roger Dressler - 6/25/12 at 12:44pm

So, Auro and MDA are object-only, while Dolby Atmos is hybrid (a mixture of objects and "fixed channels"), is that correct?

No. Auro3D is channel only. Atmos and MDA can use objects and channels.

Ah, OK, thanks. So what's so special about Auro-3D, then? Haven't such "incremental" systems been around for a long time now? (though not put to much use beyond some demos) Isn't the objects approach much more revolutionary, as well as much more flexible? (no [?] loss in quality when "scaling down" for a fewer-loudspeakers setup)

Good questions.
Auro3D has a means to carry their height channels downmixed into a standard 5.1 PCM payload. But since current D-Cinema specs already define up to 16 channels of 48 kHz 24-bit PCM, and there's room on the DCP HDD's for that plus a standard 5.1 mix, it makes one wonder.

To expand on what Roger said, both Atmos and MDA can be called "hybrid" technologies. Either one can be fully object-based or fully channel-based, or some mix of the two, depending on what the mixer wants. The hybrid approach is always going to be necessary, because any movie soundtrack will have some audio elements (for example, ambient noises that are supposed to fill the entire soundscape) that don't make sense to encode as discrete objects.

 

May be one day it will trickle down to the masses like me :) $30K for a processor ouch!!