One of the hot topics at AES this year…and by “hot,” I mean a subject that had multiple conference sessions devoted to it…was the concept of adding height to the spatial information presented by multi-channel surround formats. I’m sure a fair bit of the enthusiasm for this subject is caused by the announcement and release of Dolby Atmos earlier this year.
My experience with Dolby Atmos prior to AES was non-existent. To date, there are only 14 theaters in the U.S., and one in Canada, currently equipped for Atmos playback. The closest theater to me is in New York, and that’s not exactly a short trip from the Washington, DC, metro area. Thankfully, my trip out to San Francisco for AES provided me with two opportunities to listen to the system at work. The first was a technical demonstration at Dolby Laboratories, scheduled as a “Technical Tour” within the AES events program. The second was the AMC Metreon, which had two daily showings of Chasing Mavericks; the latest film release to be mixed in the new Atmos format.
Just imagining all that could be done in creating subtle backgrounds and ambiences, I was excited to hear what this system could do…though I fully expected the bulk of the examples that Dolby would be showing would tend toward spectacle. That proved, for the most part, to be true. Which made the opportunity presented by Chasing Mavericks all the more important; a chance to truly hear how editors and re-recording mixers would make use of the system throughout the course of a story. Before I get too deep into those experiences though, let’s talk about some of the interesting technical abilities of the system.
The system does not store your typical “channel” information. In the base system for panning, there is no dedicated “Left,” “Right,” “Center,” etc. Rather than store channel information, the system emulates something that has been used in game audio for years, spatialization within a 3D model. The reason for this, is that the system is scalable up to 128 channels. Yes, you read that right…128. The decoder has to be programmed for the playback environment it is installed in. Once it has that information, it creates the playback channel information in real-time during the film. In this manner, a film can be mixed in one room with a given set of playback channels, and be distributed to any number of theaters in any number of configurations; though Dolby has suggested that the minimum installation for an Atmos system is 9.1. The system also has the ability to output printmaster versions in what would be, at least with respect to Atmos, downmix versions of 5.1 and 7.1 (with the ability to go in and tweak those particular mixes). While this capability was described, I don’t believe a specific example was played during the demonstration. I have a hard time believing any re-recording mixer is going to inherently trust that system. It may be a starting point, but I imagine many are going to want to confirm and adjust the mix personally.
Some other interesting facts about the system…
The surrounds extend further to the front of the theatre, with individual units being placed between the audience and the screen. [I’ll touch on their reasoning for this shortly.] Additionally, the surround speakers can be treated as an array (what we are familiar with in 5.1 and 7.1) or as individual point sources. This all depends on how you use them, similar to how one might apply divergence across the front speakers in a 5.1 or 7.1 system. Each of the individual surround speakers can also be Bass Managed to provide for a more natural response in the room.
Continuing in that vein, the yet to be released cinema processor (all theaters are currently using the same decoders that are used in post-production) will include EQ management. This unit will allow the room’s response to be measured, and a unique EQ profile can be established for each individual channel in the system. This processor will also be somewhat “self-aware” [forget the T1000, SkyNet’s infiltrating the cinema business] in that it can identify and report whenever there is a failure in the playback system. Harman, QSC, and other manufacturers of speakers and amplifier systems are in the process of developing units optimized for Dolby Atmos systems.
OK, I know. These are all interesting technical notes, but how does it sound?
To be perfectly honest, the Dolby presentation left me with mixed feelings. A number of examples were played for us. Rather than go into extreme detail on all of them, I’d prefer to point out the moments that left the biggest impressions on me…both good and bad. As a point of reference, I was sitting nearly dead center in the room.
Two examples were audio only, no picture: an approaching thunderstorm that passes by,and a musician singing while playing guitar. The thunderstorm sounded great. There was a lot of detail, and gave a great sensation through the overhead channels. It truly felt like rain was hitting the roof; which is silly when you consider the fact that we were in a carefully designed, acoustically isolated, theater. What this example didn’t do so well, was show off localization in the overhead plane. There was one, maybe two, thunderclaps that did not feel as though they “hit directly overhead.” This was a missed opportunity. The sound of thunder coming from overhead is wonderful, but the sound of it localized in the center of the room, nearly every time, only draws attention to itself…lessening the immersion.
The musician was simultaneously my favorite and least favorite demonstration. They played the musician back twice; once in 5.1, and once in Atmos. Predictably, the transition across the rear (really, the surround arrays, because there is no rear) in 5.1 had that characteristic “jump” from the left to the right. Unless treated carefully, the phantom center between the surrounds does have a tendency to fall apart. And it did. Cue up the Atmos pass, with additional playback channels in the rear. The transition truly was in the rear, well behind me, and incredibly smooth. It left me very impressed…which was good for the Dolby folks, because the transitions across the sides (from front to rear) did not. There was a very noticeable shifting coloration that occurred as the sound passed from one speaker to the next. You could hear the frequency response drop as it moved into a phantom center between two side speakers, and return to nominal as it localized again in one alone. Asking them about this later on, they admitted that the room was not properly designed for an atmos system. It is a theater that was designed for their previous systems and utilizes cavity mounted surrounds. It was stated that, ideally, they would be mounted in a plane on a flat wall with no architectural structures between them (such as you would see in your neighborhood theater) that would cause interference. That was how the rear speakers were mounted, and, remember, I found that transition incredibly smooth. Nonetheless, it had me disturbed. I kept listening for it in all of the demonstrations that followed. I did not pick up on it again in the demonstrations afterwards, and I attribute that to masking effects associated with the use of more complex elements in the left and right side speakers, less movement across the arrays and a greater emphasis on point sources and the use of separation. To be honest, I also stopped listening for it. At this point, I was more concerned with the system’s use in relation to the scene in the next examples. It’s highly probable it was still there, merely less present.
Several of the other demonstrations were segments from the films Rise of the Planet of the Apes, Mission Impossible, The Woman in Black, and Brave.
Rise of the Planet of the Apes (fight on the Golden Gate Bridge) is where I return to that item that I mentioned several paragraphs ago, the side speakers that reside between the screen and the first row of seats. These were described as being beneficial for creating more sonic space in the front of the theater. Using these speakers with the overhead arrays, it is possible to create a sonic plane that is in front of the audience but separated from the screen…leaving more ‘space” in the speakers across the front. Rise did this with the music, and I wish I could have heard it in a larger theater. The screening room at Dolby is relatively small, with not much space between the audience and the screen. I found the music in the overheads to be completely distracting. This might work better in a larger theater, where that forward “off-screen” plane can be firmly established, but I fear it’s a technique that people are going to stumble with for a bit before anyone gets it “just right.”
Mission Impossible (sandstorm chase scene) made excellent use of the side and rear channels, creating well localized points in the spatial environment without becoming too distracting. Like the scene from Rise, this was a very action-oriented sequence. The demo conformed to what I expected out of a piece used to promote Atmos: spectacle, bombast, and “Whoa that was cool!” This was not something overly difficult to achieve given the material chosen. Please don’t let that convince you that I did not appreciate it in any way. It was extremely well done. It was just exactly what I had expected them to do, and I wanted to hear something a little more subtle that would make use of the potential to have many localized quiet sources.
The next two demos moved closer to that ideal.
The Woman in Black (end of the film, in the room full of creepy wind-up toys) was not so much a “quiet” piece, it actually was quite active, but it was a piece that left me with little impression about the sound afterwards. I don’t actually consider this a mark against, just the opposite really. All of the sounds were present and a part of the viewer’s spatial awareness. The sound was enveloping, present, and in no way distracted me from the events and story being told on screen. Atmos was used to create point sources and a matching perspective that grounded us in the view the camera was giving, and then, ever so slowly, spun the sound field around as the tension built.
Brave (beginning of the film with, before the angry bear attacks) came the closest to what I most wanted to hear in an Atmos demo…subtler use of sounds to create an immersive space. In particular, as the young Merida runs off into the forest to search for the arrow that she has just fired, there are little touches in the sides and rears that really flesh out the forest. You can hear that there is life and movement, but lt’s shying away from Merida with little hints of foreboding to imply that it is not tame. For as well done as this particular example was, I still wanted to hear a use of Atmos that was less “present,” one that is purely background and scene establishing. Brave was the closest example presented by the Dolby team to that ideal, but it still had moments that “pushed” in from the surrounds.
As I mentioned, the Dolby presentation left me with a big old bag of mixed feelings. I wanted an experience that was less focused on the marketing and selling of Atmos, and more focused on its use in the broad context of a single film. Thankfully Chasing Mavericks is out, was mixed in Atmos, and just happened to have a few showings in the new format at a movie theatre right next to the Moscone Convention Center where AES was being held. Kind of a no-brainer decision here.
Dolby should have just done it’s talk in this theatre, with this film, because it sold me on the system in ways that their demonstration couldn’t.
There was one scene in particular that was exactly what I had been waiting for. It was early morning on the suburban street where Frosty (Gerard Butler) and Jay (Johnny Weston) live. Jay is trying to convince Frosty to train him to surf the Mavericks (monster waves), and Frosty is decidedly disinterested in being party to the boy’s death. I swear, I’ve experienced that ambience in real life. There are little touches of human activity off in the distance, the neighborhood isn’t truly awake yet, but the birds are lively and everywhere without forcing themselves on top of the conversation. Dolby should look into swapping this scene into their demo any time they’re presenting to sound editors and re-recording mixers, because it was simply superb!
Naturally, this film had it’s share of spectacle as well; after all, it’s a story about surfing some of the largest waves in the world. The visual edit definitely left space for the sound to shine. The waves are visually large, imposing, dangerous and long lived. The entire sound team, from field effects recordists to sound editors through the re-recording mixers, did their jobs commendably with these waves. In the Atmos presentation, you feel the oppressive nature of these mammoth waves, because they utterly envelope you. There was a level of discrete detail and power in that theater that I am not confident could be reproduced in a 5.1 or 7.1 mix. That detail and power gave the waves a life of their own. They were not just an ambience to spatialize the presentation, they were a character contributing to the story.
This was what I had needed to hear. There is a lot of potential in this format. Of course, there are dangers in trying to approach those potentials too quickly. There are obviously issues that can arise from the architectural features of a room, as was evidenced by the coloration I experienced as sounds moved across the side speaker arrays. I also still have concerns about the concept of creating a second “forward” sonic plane using the front-sides and overhead channels. In the Dolby demonstration, it was distracting and showed (at least in that context) that it couldn’t truly create a “plane.” It was more of an arch that extends from the sides overhead. I did not notice this in the screening of Chasing Mavericks, but I do not know if that’s because of the room or the mix itself. That is something I would like to explore in the future, if given the opportunity, in an “optimized” room with knowledge of the intended mix behavior.
Overall, I’m excited about the possibilities this system presents. As I’ve indicated, that wasn’t exactly how I felt after Dolby’s presentation. I do, however, think that the information and critiques that I left that experience with gave me a greater appreciation for the accomplishments achieved in Chasing Mavericks. It will probably be a long time before I get to listen in that format again, but I can’t wait to experience the different approaches that will be attempted with the Atmos system in the future.
Personally I would be very curious to hear a comparison between this and Ambisonics, which seems to be very similar and has existed much longer, to see which is better and in what circumstances.
My suspicion is that you might be able to localize sound a bit better in Atmos but I have no technical details to back that up really, just my limited knowledge of Ambisonics and my desires in mixing for video.
Thanks for this interesting account on your experience with Atmos. I’m looking forward to the day I will experience it as well.
Colin Hunter says
Nice review Shaun! I feel that Dolby Atmos is a big can of worms that may pose more problems than deliver solutions? If the thunder localisation in their demo was a missed opportunity, how likely is it that we will continue to see (well hear actually) missed opportunities in future productions? Perhaps in the long run we’ll experience better immersive soundtracks but how long will the learning curve be? Although I think a step up from 5.1 and 7.1 is an interesting prospect, I have the feeling Dolby might have gone a bit too far a bit too soon.
Shaun Farley says
I think “too far, too soon” might be a bit of an overstatement. Dolby has created an incredibly flexible system with the potential to remain useful and relevant for a long time. I see it as more of an issue of, “just because you can, doesn’t mean you should”…and that applies primarily to the people crafting the mix. It’s up to us, not Dolby, to use the system effectively in helping to craft the story. Yes, it will take time before people really start using the system to its full potential, but it’s something that we as professionals can grow into.
Will Stowell says
Dagnabit, I was at AES and totally missed this! Oh well. Very good read!
John Loose says
Thanks for all the feedback. I am one of the people that is usually tapped for these demo, but I was out of the country when this one happened. I have been following Designing Sound for a long time, so I’m sorry I missed you. I am one of the few mixers in the world who has worked in the format and I can say it is really fun to work in. Next time you’re in SF, lmk. I’ll take your comments into account for the next demo. Right now, one of my crew is in NZ with Peter Jackson doing the Hobbit in Dolby Atmos, which should be excellent. Stay tuned.
Rune Palving says
Nice writing. After the AES weekend I ended up with almost same conclusion. I can see the potential in the system especially the discreet surrounds with the front sides and bass management can be a help to tell the story. I still feel that the height surrounds are not that a powerful tool. There was a discussion later at AES where Dolby diplomatic said that some people hear heights better than others and another manufacturer claimed that we have to learn audience to listen in the height…!?!
Seablade – regarding Ambisonics there where some music engineers presenting recordings with different height surround techniques. They started the experiment with Ambisonics, one of the mikes is pointing upwards so….but they were not to happy about the result, so they ended up with a triple MS set-up where the last S is on the vertical axis. And the examples they played were convincing. You really felt you were at the venue.
Another discussion at the AES was the reverb in Atmos. Unlike some of the other 3D sound systems it does not have a reverberator to do spatialization. It is the a dry signal it spreads. Someone was anxious talking about a 128 ch Altiverb…!
How did you solve that John? Very interested to hear more from the Hobbit mix….
Gabriel Guy says
Great insights Shaun, I’d like to touch on a couple although I know this article is a couple years old at this point.
In regards to dong the subsequent downmixes (7.1, 5.1, 2.0) from an Atmos mix, we do use the Dolby RMU to create some initial 7.1 downmix stems but then it’s up to the mixers’ ears to adjust them accordingly to make the 7.1 printmaster. It’s very much the same process as making the Dolby 2-track mix from a 7.1/5.1.
Panning through Atmos can exhibit some of the phantom center shoft between surround field speakers that you heard. Part of this can be minimized with the Atmos panner Size setting which can give a slight spread at low setting and minimize this effect, especially during slow pans.
In regards to front speaker phantom center shift, I’ve found that placing a sound as an Atmos object midway between two screen speakers (even in a 3 speakers front array setup) has greatly clarity that using a traditional panner, both when monitoring in Atmos and subsequent downmix versions. I hope this is helpful.