Post produced feature film and episodic content for the home is typically delivered to distribution channels, like OTT service providers in a mezzanine format such as Apple ProRes or the Interoperable Master Format (IMF).
This article is all about encoding such a Dolby Vision mezzanine file into a Dolby Vision bitstream.

If you are looking for information on how to encode content for a Dolby Vision delivery to a file sharing platform, like Vimeo or BiliBili, please refer to this article .
 

OTT/VOD encoding (non-Dolby Vision) 

 

 

In order to distribute content via OTT/VOD, all sources get encoded via a standard video encoder to compress and reduce the file size of the video content, followed by a packager that puts the encoded bitstream into streaming formats, like DASH or HLS. 

 

HLS – HTTP Live Streaming – a format developed by Apple splits a video file into small segments that are contained within a MPEG2 transport stream or MP4, hence the extension “.ts” or “.mp4”. HLS requires an H.264 (AVC) or H.265 (HEVC) codec and is currently the only format supported by iOS, iPadOS, tvOS and Apple’s Safari browser. 

 

DASH - Dynamic Adaptive Streaming over HTTP – also known as MPEG-DASH – splits a video file into small segments as well. Unlike HLS, DASH is codec agnostic. It supports H.264, H.265, VP9, and other codecs as well.  

 

Both HLS and DASH use a separate manifest file that links to the segments (and/or other manifests). Such a manifest file is a simple text file that is understood by players supporting HLS and/or DASH. 

 

 

OTT/VOD encoding (Dolby Vision) 

 

 

For Dolby Vision encoding, specific processing is added to the encoding stage. This processing includes  

• Dolby Vision PreProcessing 

• VES muxing  

• Dolby Vision PostProcessing 

 

Dolby Vision encoding couples proprietary Dolby Vision processes to the same standard HEVC encoders that are widely used for non-Dolby Vision UHD encoding.  

 

Encoding a Dolby Vision HEVC bitstream from a Dolby Vision mezzanine 

 

 

• The Dolby Vision PreProcessor splits the video and the metadata coming from the Dolby Vision mezzanine file into a YUV stream and an RPU stream. 

• The YUV stream is processed using a standard video encoder. 

• The Reference Picture Unit (RPU) is a special Network Abstraction Layer Unit (NALU) that carries the Dolby Vision dynamic metadata.  

• Both the H.265-encoded video stream and the RPU stream then get muxed by a VES muxer to a compliant encoded file. 

• The Dolby Vision PostProcessor smoothes out certain metadata values. This is particularly important in the case of a segmented encode. .  

 

 

Encoding an HDR10 HEVC bitstream from a Dolby Vision mezzanine 

 

 

You can also generate an HDR10 bitstream from a Dolby Vision mezzanine file using a Dolby Vision enabled encoder. The Dolby Vision PreProcessor will output the Dolby Vision video as an HDR10 compliant YUV stream and also pass the HDR10 static metadata to the HEVC video encoder. The final output is a standard HDR10 compliant encoded file. In some cases, it might be desirable to apply Dolby Vision content mapping (CM) to the incoming mezzannine to achieve a lower luminance output. If this mapping is needed, it is done during Dolby Vision PreProcessing.  

HDR10 static metadata may be present in the Dolby Vision Metadata as L6, or it can be provided by the user externally. 

 

 

Encoding an SDR HEVC bitstream from a Dolby Vision mezzanine 

 

 

SDR bitstreams can also be encoded by using a Dolby Vision enabled encoder and the respective Dolby Vision mezzanine file as an input. 

During preprocessing, the Dolby Vision Content Mapping will use the Dolby Vision metadata from the Dolby Vision mezzanine file to map to an SDR YUV stream that then gets encoded using a standard video encoder. 

 

Dolby Vision Profiles 

Dolby Vision Profiles describe the video codec and set of coding techniques used to encode a Dolby Vision asset. Those describe the bitstream format, defined by the following:  

• Video Codec and Video Codec Profile (and related constraints, etc.). 

• If there are one or two video streams (single or dual layer) 

• Cross compatibility to SDR, HDR10 or HLG 

Dolby Vision profiles inform: 

• Service Providers as to how they can create and distribute their assets 

• Device Manufacturers as to how they can play back a Dolby Vision asset 

 

Some profile examples are: 

• Profile 5 (notcross-compatible to HDR10 or SDR) – widest device support 

  • OTT/VOD 

• Profile 8.1 (cross-compatible to HDR10) 

• • Live 

• Profile 8.4 (cross-compatible to HLG) 

  • Live 
  • User generated content/iPhone 12 recording 
  • File sharing services, like Vimeo and Bilibili 

 

Using a single mezzanine for all encodes reduces the number of different mezzanine files to be sent to the service provider. 

Read in depth about this in the Dolby Vision Profiles and Levels Specification. 

 

 

The following encoders support encoding in Dolby Vision: 

Deployment	Tool	User Interface	Info	Target Customers
Cloud, on prem	Dolby Encoding Engine (DEE)	n/a	can be integrated with any HEVC encoder	Tech-savvy postproduction facilities, technology partners
On prem	Ateme Titan File Encoder	Yes		OTT service provider
On prem	AWS Elemental Server	Yes		OTT service provider
On prem	Telestream Vantage Encoder	Yes		OTT service provider
Cloud	AWS Elemental MediaConvert	Yes	AWS account required	postproduction, OTT service provider
Cloud	Bitmovin	Yes		OTT service provider
Cloud	Encoding.com	Yes		OTT service provider
Cloud	Hybrik	Yes		OTT service provider
Cloud	Moxion	yes	dailies/review platform	postproduction, studios

All info is subject to change. Please check with the respective encoding solution for their latest updates.