If Dolby Atmos is Object based and doesn’t need a fixed number of speakers, considering the cost of Atmos monitoring why would someone choose 9.1.6 over 7.1.4, or even fewer speakers? Julian investigates ‘speaker density’.
We’re all pretty used to stereo monitoring and how it works. Two speakers can reproduce a remarkably convincing stereo soundstage and the reason this can happen is because of virtual, or ‘Phantom’ sound sources. Convincing stereo imaging allows sounds not only to be located at either speaker but anywhere in between, but this stereo placement is fragile and is very location-dependent.
We are used to hearing references to “wide sweet spots” in stereo monitoring but in this context we’re really talking about varying degrees of narrow. A monitor with a particular waveguide may have improved stereo imaging but its sweet spot, where an even spread across the panorama can be found, is never going to be truly ‘wide’. The stereo image will collapse when I move significantly off centre. If I move left, everything will quickly sound as if it is coming from the left speaker only. Why?
Phantom Centre
The answer is to be found in the process which placed sounds between the speakers in the first place. The level based panning which places sounds in the centre is being cancelled out by the opposing level offset introduced by my moving closer to the left hand speaker.
The inverse square law governs the attenuation of sound with distance and the speed of sound is relatively slow. Although rooms affect this, the fact remains that the further you are from a speaker, the quieter it becomes and the longer it takes to reach you. The sweet spot within which phantom centre sources can exist is, in terms of level and in terms of differences in time of arrival, small. A handful of dB and a few milliseconds difference between the left and right channels is all it takes to move a sound all the way across the stereo soundstage.
To create this ‘phantom’ centre source, magically placing sounds somewhere in between two speakers in a way which matches the original recording, the two channels creating it have to be as close to identical as possible. As well as the listener being the same distance from both speakers and being appropriately placed so that the two speakers are approximately 60 degrees apart, the speakers have to be matched. It sounds obvious and we take it for granted that you can’t build a stereo system with speakers of differing model or brand. Both the frequency response and the time domain response would be different enough to damage the stereo reproduction being sought. We’ll return to this later.
How To Fix A Sound In Place So It Doesn’t Move When You Do
This location-dependence of the stereo image has been well known in cinema sound for a long time. If you want your dialogue to sound as if it’s coming from the characters on screen you can’t rely on a phantom centre image to do that. It will work for people in the sweet spot, in the centre of the cinema, but for those at the extreme sides the dialogue will be coming from a speaker in front of them and the actor’s mouth will be over to the side of them on the screen. Physics is physics and the only way to fix a sound in space for all listeners is to put a speaker there.
In cinema sound dialogue is usually reproduced by a centre speaker for exactly this reason. The LCR format allows conventional stereo reproduction in terms of presentation, but with a truly wide sweet spot because the phantom centre created by the accurate matching of the output of the left and right speakers is now represented by a dedicated centre speaker, the location of which is unambiguous to all regardless of where in the room they are.
What Does Stereo Reproduction Have To Do With Immersive?
The role of phantom sources is just as important in immersive formats as it is in stereo. Using the well understood example of a stereo system and applying those same principles to immersive arrays is an interesting exercise. Looking specifically at Dolby Atmos, the whole point of this object-based format is that material can be rendered for reproduction on a system with an arbitrary number of speakers at the point of presentation (i.e. at playback). Many of us think of 7.1.4 when we think of Dolby Atmos. It is a baseline configuration but Dolby Atmos can be reproduced with more or fewer speakers and how many speakers you have in your Atmos monitoring system depends on a number of factors. Dolby recommends at least 5.1.4 for home entertainment content but 7.1.4 is the preferred minimum, 9.1.6 is found in large studios. There are options both above and below this but considering the cost of a Dolby Atmos monitoring system why would someone choose to install 15 speakers and a sub in a 9.1.6 system when you can do the same work with the 9 speakers and a sub of a 5.1.4 system? The answer lies in localisation and phantom centres.
In a Dolby Atmos monitoring system Objects can be placed anywhere in the surround + height soundfield and if the location the Object’s panning metadata places that Object at is somewhere between the speakers, the rendering process will send the appropriate signals to the appropriate channels to create a ‘phantom centre’ to represent that location. For point-source sounds this is likely to involve three speakers. The more speakers there are in an immersive array the less distance there will be between these speakers and the more robust the localisation on this source will be.
Localisation Of Objects
In much the same way as with a stereo system, if the listener can guarantee that they are in the ideal ‘sweet spot’ for that monitoring system then these delicate relationships between speakers, which are being used in combination to place a sound somewhere between their actual positions, are more likely to be able to present a convincing localisation. If the studio needs to accommodate multiple people then, in the same way that it is difficult to accommodate multiple listeners in the sweet spot of a stereo monitoring system, it is more difficult to accommodate multiple listeners in the sweet spot of a 7.1.4 system than it is under a 9.1.6 system. This is one of the reasons you are more likely to find such a system in a larger studio, the others include the cost and space requirements of such a system.
Another reason a designer might choose a greater density of speakers is that such a system offers inherently better localisation so if a space is acoustically challenging then a greater speaker density offers a more robust system which is more tolerant of difficult environments.
So the rule is more speakers is better, but more speakers are also more expensive and are more challenging to install, particularly in small spaces. A space designed for a single user which is reasonably acoustically suited to an Atmos install would potentially offer acceptable performance with a significant saving on equipment costs. It probably goes without saying that it’s a very good idea to have the same model of speaker in all channels. This isn’t always possible but in these cases systems are more forgiving, particularly with the advent of speaker calibration systems.
It’s common practice to have speakers all from the same product family but to have larger, more powerful full range monitors across the front and to use smaller models in the surround and height positions. The reason for going with larger monitors for the LCR was dictated by the requirement to deliver a certain amount of SPL as the listening position. This used to be an issue in Post Production where the screen channels were 3 meters away so you would have to go with something with enough horsepower to throw to the mixer and satisfy Dolby’s target of 85dB SPL for the Home Entertainment certification. This is less of an issue in music as the listening position is usually fairly close and the vast majority of Atmos music rooms are aiming for 79dB. This rather contradicts the idea that to create effective phantom sources between speakers the speakers need to be identical, in the same way as they do for a stereo pair.
Monitoring For Good Immersion
In a properly tuned system which uses speakers with adequate headroom and exploits speaker calibration to compensate for frequency and time of arrival differences between channels these compromises can be adequately compensated for. The system I’m most familiar with is that of Genelec SAM monitoring used in combination with the GLM loudspeaker management system. The GLM software automates the measurement process and with the new 4.2 version, the software generates a comprehensive report on the acoustical response of the room. An important factor in optimising the immersion of a system is in the contribution of the room, particularly the off axis sound and having an optimised system with all the same model of monitor is of course the ideal. In the Genelec SAM range that would be ideally be a system entirely comprising The Ones of the appropriate size for the listening distance and bandwidth requirement. Such a system is out of reach for many (including me). A very effective compromise would be to use the same model for the front and surround channels and then go a size smaller for the height channels. With a product family with a really consistent sound you could pair larger monitors in the LCR with a size smaller for the sides and rears. Possibly even two sizes smaller for the heights if the ceiling height is on the low side.
More Speakers Is Better But Only If The Room Can Accommodate It
One significant factor in the ability to perceive the size/realism of objects between monitors is how effectively the directivity is controlled and how coloured the off-axis information is. There is an argument for improving the ‘resolution’ by adding wide channels (9.1.4). This can be helpful if the monitors are the same flavour and if the room is wide enough. If you have to sacrifice the placement of the L & R to accommodate ‘wides’ this will probably negate the advantages introduced by the additional wide speakers.
Is Using More, Smaller Speakers Better Than Fewer Larger Speakers?
Given the significant expense of setting up a quality Atmos monitoring system and the guaranteed matching between speakers of using the same model of monitor throughout the system the idea of a system built using the smallest monitors which meet the SPL requirements seems attractive. This approach can work well. Using the Genelec range as an example again, an Atmos systems for smaller rooms based around 8330s and 8320s is popular and can work well. The key is being able to position them correctly to satisfy the Dolby criteria and ensure that the flow of the room is preserved. (i.e. not having to move speakers when they aren’t in use/de rig for stereo work etc). These monitors work particularly well for difficult installs because they have so many mounting options available for ceiling and wall mounting. One very significant saving can be found by using GLM 4.2 as an alternative to an expensive hardware monitor controller for Dolby Atmos. The 4.2 update introduces MIDI CC mapping so you can control the monitor level using a MIDI controller. Another benefit when considering the use of smaller speakers is found in bass management. installing a more powerful subwoofer and using something like the Genelec 9301B allows the subwoofer to support the bass content across all chandelles, not just the LFE.
The principles outlined here illustrate the complexity of setting up an Atmos monitoring system. A nearfield stereo monitoring system doesn’t really have to be planned and designed for a mix space in the same way as an Atmos system does but the same principles apply. If your room can accommodate more speakers you can potentially get better results but the install needs to work in the room both in terms of physical installation and acoustic performance. Luckily more and more options are opening up and many people who thought the costs around Atmos monitoring put such systems beyond their reach might find it to be more achievable than they thought. To answer the original question, for all these reasons, higher speaker density is better in Atmos systems but it does depend on what those speakers are and whether they will work with your room.
Images courtesy of Genelec and Sandy Kawadkar on Unsplash