Today I heard my second demonstration of Dolby Atmos, the latest and apparently the greatest in multichannel surround sound. This second time was in a proper home cinema demonstration room with acoustic treatment and fixed in-ceiling speakers, whereas the first had been a touring show set up in a hotel function room, with ceiling speakers hung from an impressive construction of ladders and beams hired from the local Bunnings. This is not to denigrate the first demo - indeed I was hard pressed to pick a preference between the audible results (they used the same Dolby demo Blu-ray, so comparison was meaningful), which speaks well for both demos, and for this new format in general.
For the record, the first demonstration was from QualiFi, using Jamo speakers and Marantz processing and amplification, the second from Amber Technology, using NHT front speakers, 6.5-inch two-way in-ceiling speakers (most probably Sonance) and an Onkyo receiver.
Dolby Atmos adds either two or four ceiling speakers to an existing or new multichannel audio system. If you currently have 5.1-channel surround, you could add two ceiling speakers to have a 5.1.2 Dolby Atmos system, or four to make 5.1.4. You needn’t stop there - home systems can have up to 24 speakers on the floor and 10 overhead, while the first-generation professional Dolby Atmos CP850 processor supports up to 61.3 channels, the configuration used in the top commercial cinema installations for Dolby Atmos. Australia currently has only one of these - at Crown Casino - but Sydney will shortly enjoy Atmos in George Street. If you go, make sure they’re showing a Dolby Atmos-mixed soundtrack. And sit as dead centre as you can.
This ability to scale to the size of system is at the heart of Dolby Atmos, because Atmos is not about adding channels, it is about creating a near-sphere of sound using however many speakers you have available. Remarkable as the proposition sounds, the final audio mixdown now happens in the AV processor, in the Atmos decoder, not in the studio where the soundtrack was fashioned.
There are still audio channels as we currently understand them - nine of them, now being called ‘beds’. But the sound engineers can then add audio objects, their sound separately encoded along with metadata that describes their position in three-dimensional space, their physical spread, and the sound’s intensity. Only at the decoder will these audio objects be placed into specific channels, this depending on however many speakers are available, and how they’re set up. So exactly the same soundtrack used in the cinema can be used at home, with Dolby Atmos able to scale down to home layouts - indeed, we gather, all the way to mono if required. (“For Indian cinemas”, someone whispered to us rather uncharitably. We suspect they’ve never visited Jaipur’s Raj Mandir.)
This is extremely impressive - it shows the remarkable computing power that lies within today’s AV processors and receivers, for one thing, but is most remarkable in delivering an entirely new way of storing and transmitting audio, as separate fragments tagged with metadata to be downmixed at the end user. With 128 audio objects available, the soundtrack guys can have a fair bit of stuff whizzing stuff around the auditorium - or so you’d think. But there is a significant limitation on this, which we thought was audibly evident in both the Atmos demonstrations we heard.
It’s all around you
Atmos is already being described as ‘3D sound’ - though given the faltering failure of 3D video, this might not be too smart a marketing slogan. It certainly adds an additional dimension in providing height channels, and the demo material comes steeped in ceiling sound, just as the early demos of rear surround sound sent huge amounts of stuff to the rears to make the point. Any demo you get to hear will likely have the ceiling speakers much louder than a calibration would set them, so you get bombarded with sounds flying above and around you.
But what it doesn’t do is come AT you, like an axe in a 3D horror movie. The most surprising thing about the Atmos demos was that while stuff was whizzing around everywhere from front to ceiling to rear, there was nothing IN the room. Everything is AROUND the room. The second demo delivered some fairly localised positioning behind us, but never did anything go right in front of our face, buzzing like a fly before our nose.
Why? Because in a cinema, if you mixed a fly to be an audio object in front of your nose, then it would appear to be in front of the noses of those in the dead centre rows, but progressively out of position for everyone else in the cinema. Ditto for front to back movement - the point at which it goes behind you will be different for everyone in the room depending on their relative position to the nearest side and ceiling speakers. In the cinema that’s an unfixable problem which can only be addressed by not allowing it - no flies in front of your nose. So Dolby Atmos seems to deliver a sphere - the sound moves around the outside, it does not enter far inside.
Home cinemas have far fewer seats - just one sweet spot seat, if you like. So Atmos could potentially deliver a far more localised sonic experience at home than in the cinema. But it won’t, precisely because the exact Atmos soundtrack from the cinema is being scaled down to the home without engineers remixing it for a different environment.
It’s not below you
And the sound is not quite a sphere but a truncated sphere. One change that may affect those upgrading an existing home theatre is that whereas rear and surround channel speakers are often mounted fairly high on walls, Atmos requires them to be brought down as low as possible. Previously putting them high helped create just that - height. Now we have dedicated height speakers, Atmos will work best with the biggest possible separation between high and low. And it matters. However high your speakers are, that’s where the sphere of sound will be truncated.
One of the demonstrators told us that they had visited Dolby’s research division some years earlier - they were surprised how open Dolby had been, allowing them even to take pictures. A circular rig of speakers plus ceiling speakers were set up, clearly a precursor to delivering Atmos. And at one point the demonstrator turned to the visitors and said words to the effect of - ‘Now we just need a way to get it coming from the floor.’
And this was what we felt in the Atmos demos, that we now had a complete sphere of controlled sound around us, but we were conscious that the sphere was cut off at seat level, effectively the tweeter level of the front speakers. There’s not much to be done about that, and it’s not new - there’s never been any sound down there. But gaining the sound above seemed to create a contrast with the lack of sound below, a kind of emptiness as the soundfield fades to black.
Did we like it?
Here we are pointing out what this new technology doesn’t do, and what was missing. Typical audio journalists, you’re thinking, never happy. But hey, we loved it. There is greater immersion, a new level of impact and potentially audio control, and we can see why sound engineers must love it, because they can steer a single sound with a real or virtual joystick rather than abstractly using faders to mix sounds into already busy channels (though this is commonly done by joystick as well). Given the complete Z-space mapping of today's animated movies and others (e.g. the Atmos-mixed Gravity), we imagine that importing data from video object to audio object may even make manual steering redundant.
We very much doubt that Atmos will prove a fad, partly because it’s a clear and genuinely useful improvement, and partly because of the potentially revolutionary use of metadata rather than just audio files to create a bespoke mixdown for the individual end-user. This has potential in many areas. (We wonder if music might be delivered this way, using 128 individual tracks as specified audio objects, each with 24-bit/96kHz resolution and their position within either a stereo or surround mix. And perhaps an iPad app to allow the user to entirely remix the track positions, levels and effects to their whim, but now, yes, we’re dreamin'…)
A usual question with a new format (notably at present with 4K TV resolution) is where’s the material we can play? For once, there’s no wait - the first Blu-ray with a Dolby Atmos soundtrack is out at the end of November - 'Transformers: Age of Extinction' (don't all cheer at once). Others should follow rapidly, since Hollywood, Bollywood, China and others have been mixing in Dolby Atmos for several years already.
As for Atmos decoders and upward-firing speakers, we confidently predict that January’s CES in Las Vegas will be awash with Atmos-compatible hardware. Meanwhile the relatively simple upgrade path from an existing surround sound system makes it highly attractive to retailers and installers - we gather that hanging speakers are being used to create ‘Atmos zones’ even on busy chain store floors.
It’s also very easy to explain to consumers. After a cursory explanation of Atmos to my taxi driver (provided to and from the second event by Amber) he made a phone call while I was at the event to make sure the media room in the home he’s converting will be at least pre-wired for ceiling speakers. He understood how it could improve surround sound even without hearing it. (The rest of the journey we spent discussing vinyl and his seeing the Beatles in Sydney in 1964.)
So on the whole, after two enjoyable demos, we’d be giving a big tick to Dolby Atmos so far.
We had but one burning question to be answered, and Dolby Australia’s Michael Smithers was able to do so after the Amber presentation. With an audio object, we asked, is the position described as a vector? Does the system say ‘Now it’s here, and in five seconds it’ll be over there - you fill in the gaps’ or does it describe the object’s path granularly with a metadata position for every sample? It is, we gather, the latter. Phew. Glad that’s sorted.
FOOTNOTE: Ceiling speakers, really?