How To Optimise An Audio Mix For Delivery To Netflix
To help the growing number of audio professionals creating content for Netflix, the streaming company has produced a tutorial video designed to provide useful advice on how to optimise a mix for their streaming audiences, insights on how Netflix members consume Netflix content, as well as sound mixing guidelines and deliverable requirements for original version sound mixes, M&E mixes and alternate language sound mixes.
To help you access this advice from Netflix you can either watch the video presentation or read a transcript. We have only made minor adjustments to what Netflix say in the video to make it easier to read. Please remember that when reading this transcript, it is Netflix speaking, not Production Expert.
The Transcript
In this video, we're going to talk about sound mixing for Netflix. We'll answer these questions…
How should I prepare sound mixes for Netflix?
How should I prepare M & E mixes for Netflix?
How should I prepare dub mixes for Netflix?
What is the purpose of Netflix sound specifications?
How may I ensure a sound mix plays well on the service?
How does Netflix process and stream sound mixes?
How does Netflix optimize sound experiences for members at home?
Our goal is to enable a clean sound pipeline from the microphone onset to the speaker in a member's home by protecting the sound pipeline, we're able to provide members with the sound mix that the creator's intent.
How should you prepare sound mixes for Netflix?
For the main sound mix that members hear at home, there are three ways to deliver your mix…
Atmos delivered within an IMF
Atmos delivered outside of the IMF.
5.1 and stereo mixes without an Atmos delivery.
When delivering Atmos within the IMF, it's not necessary to conform precisely to the final picture master.
However, when delivering Atmos as a sidecar, outside of the IMF, leader in-sync pops must be removed and the Atmos mix must be the precise length of the final IMF picture. This is because we encode the Atmos next to the IMF and expect the two to play in sync throughout.
When Atmos is delivered in an IMF, the IMF packaging vendor will sync Atmos to the picture and the sound facility will deliver Atmos as they would traditionally deliver 5.1 with an academy leader and two pops.
When Atmos is delivered in an IMF, both the 5.1 and the stereo mix are optional. Netflix can create those automatically, according to the Dolby downmix parameters selected in the renderer at the time of mastering. When Atmos is delivered outside of IMF, the stereo mix is still optional as it is when just a 5.1 mix is delivered.
Netflix also requires 5.1 and stereo dialogue, music and effect stems, which is a traditional D&ME for marketing and some dubbing use cases. Netflix requires a 5.1 discreet wav mix delivered for a dubbing reference. Netflix asks for the final mix project files, which are stored in the archive because they are the best archival deliverable for sound.
How should I prepare M&E mixes for Netflix?
Most content delivered to Netflix will be dubbed for a global audience. These servicing assets include a print master and fully filled M&E along with any optional tracks to be created in accordance with Netflix Partner Help Center guidelines.
For mixes finished in Atmos, we will require the final mix to be delivered as an ADM BWF and the M&E separately as either an ADM BWF or Pro Tools session. You may provide any number of optional tracks in any channel configuration from mono to 7.1.2, also folded down to 5.1, if any of the channel configurations exceed this, regardless of whether Atmos mixes are delivered as IAB or via sidecar.
We continue to ask at this time that 5.1 WAV fold downs are authored for all projects to service the needs of our dubbing and audio description productions. For 5.1 surround only projects, we simply require the mix M&E and optional assets to be delivered with this channel configuration.
You will be informed by a Netflix representative if your project does not require M&E creation. In these rare circumstances, we will continue to require final mix deliverables to produce audio description. Non-servicing assets are also required for the M&E QC process, including a dialogue guide and the reference video used during the M&E mix. Careful preparation and labelling of optional tracks are critical to a successful dubbing strategy, generic guidelines, and this M&E creation tutorial video can be found on the Netflix Partner Help Center.
M&E Creation Guidelines
For certain projects, a representative from the Netflix dubbing team may reach out to you for specific optional track configurations to better align with the creative approach of our dubs.
Grunts, sighs, laughs, cries, on-screen singing performances, songs produced for the show, dialogue spoken outside of the primary language, film clips, old TV shows, radio programs and Wala with intelligible dialog are all examples of audio that should be separated out from the M&E and delivered as optional.
You can deliver as many separate optionals as you need to ensure the separation of all of these elements.
When in doubt, it should go in the optional track to give dubbing mixers as much flexibility as possible.
Providing supporting documentation to explain optional track content and labelling is highly advised secondary audio assets, including any alternate language and audio description tracks are delivered independently to the Netflix platform via Backlot, always conformed to the primary digital audio-video assets delivered.
Their channel configurations must match those of the original language mix. Acceptable formats for the alternate language audio tracks are stereo or multi-channel WAV extensible format or iTunes QuickTime format as described in the iTunes package asset specification.
As part of the dubbing and audio description workflow, the partner responsible for the fulfilment of the service asset must also deliver a number of archival assets, including a fully mixed digital audio workstation session, and a fully mixed dub dialogue WAV stem of matching channel configuration.
What is the purpose of the Netflix sound specifications?
Dialogue is the anchor element of a Netflix mix. Our core sound specification is for loudness at -27LKFS within a 2LU window. This means you could deliver an average loudness of -25 to -29LKFS or LUFS. Netflix asks that you measure this with dialogue intelligence, according to BS 1770-1. The meter will judge what is dialogue and it will not measure sound effects and music. If you want to learn more about Netflix Loudness specifications, check out our article Has Netflix Turned The Clock Back 10 Years Or Is Their New Loudness Delivery Spec A Stroke Of Genius?
This allows mixers to work as creatively as they like with sound effects and music elements using the full dynamic range and not change the resulting loudness measurement. Many common meters are available with the Netflix preset and there's a list of meters in the sound section of the Netflix Partner Help Center. TAt the time of writing, they are…
Plug-ins
Enterprise Solutions
Dolby DPLM
Interra Systems Baton
Emotion Systems Eff
Minnetonka Audio AudioTools Server
Netflix also requires peaks that do not exceed -2dBFS true peak. For open codecs, such as AAC. this ensures that peaks don't exceed zero causing audible clipping on a device. For Atmos, peak overs are an FYI and not an issue. This is because it's very difficult to control peaks in the automatic renders that the Dolby tools make. So peaks are strictly enforced for 5.1 and stereo, but are an FYI and not an issue for Atmos.
The sound specification is a promise. For titles delivering within the spec tolerance, Netflix won't apply limiting to an audio stream and the platform does not normalize audio beyond -2dBFS true peak. All audio on Netflix is aligned to precisely -27LKFS or LUFS for dialogue.
So if you're delivering within spec, Netflix, won't turn it up or down by more than 2dB. This ensures that all titles on Netflix play at a similar loudness level. When a member switches from series to a feature or back, the dialogue will be in a similar range.
How may I ensure a sound mix plays well on the service?
Dynamic range is central to providing a great experience for members. Netflix recommends that the full mix is between 4 and 18 LRA and that the dialogue is within a window of 10 LRA or less. It's important to create a very full dynamic mix that tells the story and provides a premium experience. But excessive dynamic range requires members to ride the volume on their remotes and this can take people out of the story.
A moderate dynamic range will play best in the home. Netflix recommends using a meter that provides LRA measurements and particularly dialogue LRA, a moderate or tight LRA for dialogue will allow a mixer to still use the full expressiveness of dynamic range for music and effects elements while keeping the core anchor of the dialogue, very reasonable. Members will be able to clearly understand the dialogue and we'll still have a great premium experience of a full cinematic mix.
How does Netflix process and stream sound mix?
Our primary audio streams are HE-AAC, the most commonly streamed, Dolby Digital Plus for 5.1, and Atmos. Atmos streams are in the form of Dolby Digital Plus JOC (Joint Object Coding). xHE-AAC is a recent codec addition that provides stereo experiences for Android mobile devices.
This includes a -16LUFS target for a dynamic range control to provide a good experience for members watching on mobile, where it can be difficult to fully understand the dialog. In the case of Dolby Digital Plus, rather than gaining up or down the mix to align to -27LKFS, the codec will use dialnorm metadata to do that on decode at the device.
How do members listen?
A significant majority of members listen to stereo in some form, a large number, receive 5.1, and a small, but growing number receive immersive, which is in the form of Dolby Atmos. Immersive audio is steadily growing as a percentage of Netflix viewing. About 72% watch on a TV in the home, about 20% on a mobile device and about 8% on laptops.
In conclusion, how should you mix for a Netflix title?
Create one near-field mix with moderate dynamic range, checking the 5.1, and stereo derivatives to ensure that they are providing a great experience.
The Video
Want to learn more?
There you have it, all the advice from Netflix. There is more from them in the Sound section of the Netflix Partner Help Center.
We also have a significant amount of information Do check them out…