Dolby Atmos Demystified
Table of Contents
What is Atmos?
Recording engineers and producers have been experimenting with sound source positioning ever since stereo loudspeaker reproduction became possible. Today’s technology enables full immersion in an audio scene, with Dolby Atmos gaining adoption across consumer formats. Eventide is now producing Immersive plugins that are capable of applying signature Eventide effects to Dolby Atmos mixes. With technology evolving so rapidly, you may find yourself asking: exactly what is Dolby Atmos? And more importantly, how can I start making Atmos productions? Today, we provide an introduction to working in Dolby Atmos and how you can start producing immersive mixes today.
Dolby Atmos is a broader specification that provides tools needed for mixing engineers and consumer playback. This includes a specification for the location of playback speakers, two different track types (object tracks and the multichannel bed track), and a renderer that mixes tracks into playback channel layout formats. These tools insure that when a producer or mixing engineer applies spatial effects to a source in their mix, that same spatial information is preserved as closely as possible for the consumer to enjoy.
Objects and the Bed
Dolby Atmos mixes contain two track types: object tracks, and a multichannel bed track. Object tracks consist of mono or stereo pairs of source material. An object track sends audio content and spatial positioning information to a mixing engine, or renderer. They can be thought of as a stem in a DAW, with a three-dimensional panner. The bed track has one stream of audio for each speaker, and these streams are statically positioned at the location of each corresponding speaker in an Atmos layout. Eventide Immersive plugins were designed to create or be used on a bed, as speakers on the bed are in fixed location. But, Eventide Immersive plugins can still be used in any Atmos output format that is a subset of the bed for the entire mix, and additionally Eventide Immersive plugins can be used on object tracks with a send. For example, producers can instantiate an Eventide Immersive plugin in quadrophonic on a 7.1.2 mix, and this can be used to keep reverbs out of the center channel or pitch effects out of the elevation channels. Or, an Eventide Immersive plugin can be made at the full Atmos bed and maintain positioning of spatial information when being downmixed, so that mixes are consistent across different playback formats. Another way of putting this is: Eventide Immersive plugins can be used to partially or fully upmix when applying effects. Additionally, object tracks can be panned dynamically into an Eventide Immersive plugin as a send, and the Eventide Immersive plugin will respect the pan of the dry object as well as react dynamically to the panning information. This is particularly effective when setting up an Eventide Immersive plugin to have different parameters or properties in different spatial regions.
The speaker specification consists of up to three numbers, formatted Mains.LFE.Overheads. The first number dictates how many mains speakers are on the horizontal plane immediately around the listener’s head. The second number represents the number of Low Frequency Effects (LFE) channels that are commonly routed to subwoofers. The third and final number represents how many elevation or overhead speakers are in the given layout. For example, 7.1.2 and 5.1.4 are Atmos formats. Additionally, traditional formats such as stereo (2.0) and surround (5.1, 7.1, etc) are compatible with Dolby Atmos.
The Renderer can generate downmixed formats, so that a producer only creates one mix then delivers stereo, 5.1, and the full Dolby Atmos renders.
The specification provides speaker locations for up to 9.1.6 channels of audio. The 9 surrounds, placed laterally around the head, consist of left/right; center; surrounds (in 5.x.x, or surround sides in 7.x.x/9.x.x ); surround rears; and wide speakers. All overhead speakers are placed at 45 degrees of elevation: the top front, top side, and top rear speakers. All channel formats currently in the Atmos spec are subsets of these combinations of speakers. For example: the 7.1.2 speaker layout consists of left/right speakers; the center speaker; a subwoofer channel; surround sides; surround rears; and the top side speakers. This specification is especially powerful as it insures a consistent placement of speakers – and thus the position of sounds in the Atmos mix – for both the mixing engineer and the consumer for playback.
For a full list of Atmos output formats supported by Eventide, consult the FAQ.
Rendering in Dolby Atmos
The final piece of the puzzle when mixing in Atmos is the renderer. This is the mixing engine that takes all of the track and panning information from a DAW and sums it down into the desired Atmos channel format. The renderer takes object tracks and distributes them to relevant speakers when the object is being panned around the head, and sums this with information being sent from the bed. Eventide Immersive plugins sit on top of the bed, or when instantiated as a sub-format they mix in parallel to the bed. At the time of this writing, the Dolby Atmos Renderer can handle beds of up to 7.1.2, and certain DAWs can render beds of up to 7.1.4. Additionally, the renderer can down mix from the native bed to reduced channel counts or to binaural renders. Down mixed renders allow producers and engineers to mix in higher channel counts such as 7.1.2, but still render and deliver a consistent mix for lower channel count formats, such as 5.1 and 2.0. The binaural renderer generates the Atmos mix specifically for headphone playback, simulating what that mix sounds like inside of a real Dolby Atmos speaker array. While down mixing does remove spatial information – especially when down mixing from a format with overhead speakers to another format without them – much of the frequency-based content and some spatial properties are still preserved to the down mix. All spatial information is preserved in the binaural mix.
This processing pipeline – speaker specification, object and bed tracks, and the renderer – defines the framework for producing audio in the Dolby Atmos environment.