Topics covered in this article:
- Pre-production practices
- Post-production practices
What follows are some suggestions and tips for working on Dolby Atmos Home Entertainment content. This is not meant to be prescriptive. Every project is different, and every facility and mixer has or will develop techniques and methods best suited to them.
A mixer may not always be involved in a production from the start. However, depending on the production, it may be possible to convey some requests that will assist in working with Dolby Atmos in post.
Gather audio assets for the final mix that will allow the sound editor and mixer the highest level of flexibility for panning and mixing. Dolby Atmos relies on mono or stereo elements that can be dynamically panned and moved, and in this sense, is no different in concept to layered editing and mixing in Stereo, 5.1, and 7.1. If only 5.1 stems are received, the output will likely be a static, mostly 5.1 mix, as it is more challenging to separate individual, movable Objects once a premix has been flattened to 5.1.
If possible, request music stems in sets of stereo and mono files, as opposed to a finished 5.1 mix, or 5.1 stems that are hard to “unpack.” If a composer is accustomed to delivering finished 5.1 mixes, ask for a 5.1 guide track and stereo stems broken out by instrument group.
Multichannel microphones are not always practical in scripted production where clean, discrete recordings are preferred, but they can be invaluable for post-production of recorded live events. Soundfield/ambisonics microphones, like those from Sennheiser (Ambeo) SoundField, CoreSound, Røde, Schoeps (ORTF-3D) et al., are becoming increasingly popular for immersive capture. The multichannel output of these systems can be decoded to Dolby Atmos to reproduce the height plane. Placement of these microphones is more convenient than hanging mics over the audience, but the result can be more diffuse than using multiple directional microphones. You are also ‘baking in’ the directionality of sound sources within a scene, meaning you are unable to pick and choose elements within the recording.
There are several best practices that can make Dolby Atmos post-production workflows more streamlined and predictable. Planning is key to finding an approach that works best for a given mix/show, and can be maintained across multiple mixes if working on an episodic.
It is prudent to closely study delivery specifications and plan on the best way to achieve creation of all the deliverables required.
The goal of these practices is to:
- Organize Pro Tools sessions and the Dolby Atmos Renderer in a way that is easy to understand and facilitates downstream localization work.
- Minimize the number of mastering passes that are required; this can save time and effort.
- Minimize the number of deliverables and converge towards a common workflow if working on an episodic.
Minimize the effort required to derive localization assets.
Using multiple Beds is a recommended practice and is key to successfully generating stems and mastering-ready assets. It is important that both the team managing production and the audio post-production team are aligned at the start of the audio production process. When working in Pro Tools, each stem (such as Dialog, Music, Effects, etc.) can be assigned a Bed (stereo up to 7.1.2) and a set of Objects.
In some workflows, these stems are all combined into a composite Bed for output to the Dolby Atmos Renderer. This allows up to 118 Objects to be used in the mix. However, this approach makes it impossible to derive channel-based stems without making multiple real-time mastering passes.
Assigning each stem Bed in Pro Tools to a unique set of outputs feeding the Dolby Atmos Renderer means that one master can be created that will allow for all the channel-based stems to be derived in a single pass. This master, including the multiple Beds, can be delivered as-is for encoding and distribution.
A second master may also be created with a fully filled music and effect (M&E) stem to create a single localization master. A fully filled M&E includes the non-dialog sound elements from the dialog stem, as well as additional Foley to cover elements of production sound lost when removing the Dialog stem. The original language can be included as a custom group of beds and/or objects in the localization master for guide purposes, and removed by re-rendering all other elements to provide clean localization assets.
This simplified approach saves time and reduces the number of master file sets.
A couple examples of a multiple Bed workflow:
- Six Bed and five Object groups: All 7.1.2 except 5.1 Narration – 72 available Objects
- Intl ME Fill Helper
- Four Bed and Object groups: All 7.1.2 except 5.1 Narration – 92 available Objects
- ME Comp (FX, BG, MX)
- Intl ME Fill Helper
In the first example, 6 Beds are used. Separating out the music can be useful as it facilitates replacement, should there be regional licensing issues.
None of this means that additional Beds can’t be used in the master Pro Tools session, but they will be bussed to the main Beds and therefore, mixed into the master. These could include reverb returns, Foley, etc. to provide as much control as needed.
Previous: Module 9.1 – Module Learning Objectives