Quick Start Guide Dolby Vision: DaVinci Resolve Studio

Oct 19, 2023
Knowledge

Title

URL Name

Quick-Start-Guide-Dolby-Vision-DaVinci-Resolve-Studio

Created By

MichaelH, 10/19/2023 11:51 AM

Last Modified By

MichaelH, 10/19/2023 11:51 AM

Validation Status

Validated

Publication Status

Published

Assigned To

Assigned By

Assignment Due Date

Article Content

This article explains how to set up DaVinci Resolve Studio for a Dolby Vision project and covers the main steps required to create content in Dolby Vision.

We will discuss the following topics:

Project Settings
- Master Settings
- Color Management
- Dolby Vision Version & Mastering Display selection
- Analysis Tuning
Blanking
- For single aspect ratio timelines
- For mixed aspect ratio timelines
Dolby Vision Analysis
Review of a tone mapped version
Dolby Vision Trim Controls
Export in Dolby Vision
- Dolby Vision XML
- Dolby Vision Master file
- Dolby Vision IMF package

Project Settings

Create a New Project

Master Settings

Go to Project Settings – Master Settings
Set the Timeline Resolution to 3840x2160 Ultra HD or to a higher resolution
- If the Timeline Resolution is set to anything different than the final render resolution of 3840x2160 Ultra HD,
  - Navigate to Image Scaling – Output Scaling
  - Set Output Resolution to 3840x2160 Ultra HD (UHD is the usual render resolution for all Dolby Vision home deliveries)
  - The Output resolution is stored as canvas aspect ratio inside the Dolby Vision metadata and must correspond to an aspect ratio of 1.78.
Set the Frame Rate to match the output framerate
Set the Video format to match your reference display
Select Use 4:4:4 SDI
- When using a dual monitor setup (HDR and tone mapped (e.g., SDR) simultaneous output)
  - Select Use dual outputs on SDI
Set SDI configuration as per your infrastructure
Set Data levels to Full
Set Video bit depth to 10 or 12-bit

Ensure that these video monitoring settings match the input settings of your connected grading display.

Color Management

Navigate to Color management
Select a Color science you want to use

Note:
A scene-referred (and color managed) workflow is recommended for Dolby Vision projects
This decision should be made in consultation with all other departments involved, such as VFX, the post supervisor, and the colorist's preferences before production begins

The following combinations are popular scene-referred grading options for HDR and Dolby Vision deliveries:

Color science	Timeline color space
DaVinci YRGB Color Managed	DaVinci Wide Gamut/Intermediate
	ARRI Wide Gamut/LogC
	RED Wide Gamut RGB/RED Log3G10
ACEScct	ACEScct

For this article, we select ACEScct.

Color science ACEScct

Set ACES version to ACES 1.3
Set ACES Input Transform to None (or to the dominant media format)
- If the source media is a camera raw, DaVinci Resolve uses manufacturer-supplied colorimetry to identify its input color space
- Manually set the input transform of clips in the media pool, which overrides embedded color space metadata, or the default Input color space setting
Set ACES Output Transform to P3-D65 ST2084 / * nits or Rec.2020 ST2084 / * nits, P3D65 Limited when using a Rec.2020 display
Leave ACES mid gray luminance at 15.00 nits
Leave Process node LUTs in ACEScc AP1 timeline space
- This setting only applies if you’re loading LMTs as Look Up Tables into the project. Therefore, this setting will vary depending how the LMTs were created. To learn more about LMTs follow this link.
Check use color space aware grading tools
- This enables some “color space aware” palettes inside Resolve
Set Apply resize transformations to best suit your needs

Dolby Vision Version & Mastering Display Selection

Check the Enable Dolby Vision box
Set Dolby Vision Version to 4.0 or 2.9 in accordance with the delivery specs provided by your customer or studio
- If not explicitly stated, we recommend the use of version 4
Select the Mastering Display that matches the connected reference display device and its calibration.
- For this example, we set the Dolby Mastering Display to Dolby: ST 2084 PQ / P3 D65 / 1000 nits

This mastering display setting is saved as level 0 metadata.

Analysis Tuning

Legacy CM4 L1
Most Highlight Detail / Most Mapping
More Highlight Detail / More Mapping
Balanced
Less Highlight Detail / Less Mapping
Least Highlight Detail / Least Mapping

Selecting the “Legacy CM4 L1” option matches the analysis of the original (very conservative) CMv4.0 algorithm. Older projects from software versions created before this change will open with this selected as the default.
The other five options select the new analysis algorithm, each with a different strength for the spatial filter.
- The lowest strength filter, Most Highlight Detail / Most Mapping, retains the most highlight detail and should result in no visible clipping, but requires the most mapping. Clipping should be restricted to detail-less speculars, noise, and artifacts, but results in a darkest starting point of the new analysis options. This filter is still substantially stronger (brighter results after mapping) than filtering in the legacy option and the selection is far more tolerant to speculars, artifacts, and noise.
- The highest strength filter, Least Highlight Detail / Least Mapping, often results in visibly clipped highlights, but exhibits the least amount of mapping (brightest images). The benefit is a brighter overall image mapping at SDR 100 nits.
- Balanced strikes the balance of permitting clipping in some cases to reduce the strength of mapping, but keeping it minimized to preserve highlight detail. This is the default starting point for all new projects.

Picture 1.png

For more info on Analysis Tuning see the Dolby Vision Analysis Tuning in CMv.4.0 - FAQs

Click save to apply the settings.

From here you can insert your media into the media pool and start the conform process.

Color science DaVinci YRGB Color Managed

De-select Automatic color management
Set Color processing mode to Custom
Set Input color space to the dominant media format (default input color space)
- If the source media is a camera raw, DaVinci Resolve uses manufacturer-supplied colorimetry to identify its input color space
- If the source media has embedded color space metadata DaVinci Resolve uses that to identify the input color space
- If there is no embedded color space metadata, Resolve uses the default Input Color Space
- Manually set the input color space of clips in the media pool, which overrides embedded color space metadata, or the default Input color space setting
Set Timeline color space to one of the above-mentioned scene-referred color spaces
Set the Output color space to P3-D65 ST2084 / * nits or Rec.2020 ST2084 / * nits, P3D65 Limited when using a Rec.2020 display
Set Input and Output DRT as per your preference
- In the example shown below both are set to None resulting in a simple 1:1 output
- for other options please refer to the DaVinci Resolve user manual

Color science DaVinci YRGB

Set the Timeline color space to a combination that allows you to use your preferred color space aware grading tools and operations
Set the Output color space to P3-D65 ST2084 / * nits or Rec.2020 ST2084 / * nits, P3D65 Limited when using a Rec.2020 display
- This setting defines how the iCMU interprets the output from your timeline and thus affects the Dolby Vision content mapping process

Blanking

The (output) blanking value will set the active image area for the Dolby Vision analysis. This aspect ratio is recorded as image aspect ratio into the Dolby Vision metadata
The timeline blanking provides the image aspect ratio value when exporting the Dolby Vision XML
It provides a visible mask (blanking) over your image (while grading)

In cases where the viewing aspect ratio and the image aspect ratio are the same, no blanking is required.

Blanking for single aspect ratio timelines

Go to Timeline – Output Blanking and select the correct aspect ratio
- This aspect ratio is recorded as global image aspect ratio into the Dolby Vision metadata and applies to all shots in the timeline

Blanking for mixed aspect ratio scenes (using per shot output blanking - L5)

Setting the L5 Aspect Ratio

The example below demonstrates how to setup a timeline with varying shot based image aspect ratios

Navigate to the Color page
Go to Timeline – Output Blanking and select the dominant timeline blanking that applies to the majority os shots inside the timeline
Select a clip that differs from the dominant timeline blanking
Navigate to the sizing palette
Select Output Sizing
Click the Option Menu and check Show Blanking Clip Override
Select Clip and uncheck Use Timeline Blanking
Set the custom blanking for that shot
Repeat that procedure for all shots with different aspects (clip based output blanking)

Shot based aspect ratios get recorded as per shot L5 metadata into the Dolby Vision metadata.

Dolby Vision Analysis

Before you run the Dolby Vision analysis in Resolve, you must first organize your Timeline and arrange all clips on Video Track 1.
Any group of clips that form a composite in a timeline, whether it's multiple overlapping clips combined via key or alpha channel transparency, clips combined via composite or blend modes, or text generators appearing over one or more video clips, must be converted to a compound clip for Dolby Vision analysis to work correctly.

Dolby Vision Analysis options

Navigate to the Dolby Vision tab in the Color page and check the analyse options. There are several options to choose from when running the Dolby Vision Analysis.

All: Analyzes every clip that is present on the timeline.
Selected: Analyzes only selected shots in the timeline.
Blend: Analyze multiple selected shots and then average the analysis result. Blend will analyze those selected shots if they were one. This is a fast way to save analysis time and match shots
Frame: Analyzes a single frame in a selected clip that is representative of the entire shot and uses the analysis metadata for the entire length of that shot. This is a useful option for shots with varying lighting conditions.

After the Analysis is complete, the Min, Max, and Average have now been populated in the designated fields at the top of the Dolby Vision Analysis view. These values are stored as L1 metadata. Hovering over the words “min”, “max”, or “avg” in the analysis pane will bring up a tool tip identifying the Analysis Tuning.

Picture 2.png

All clips now have shot-by-shot analysis metadata. When the Dolby Vision Analysis detects a dissolve, those transitions are analyzed using frame-by-frame offset metadata.

Review of a tone mapped version

Once the Dolby Vision Analysis is done, check the mapping at SDR 100 nit

Navigate to the Dolby Vision tab
Set Target Display Output to: 100-nit BT.709, BT.1886, Full (HOME)
Set your mastering display to SDR Rec.709 (Full)
Review the mapping

You can additionally check the mapping at 600nit HDR, as this target serves as a good emulation for most Dolby Vision TVs. To do this, simply set your HDR reference display back to HDR.

Make sure to always set the input of your reference display to match the Target Display Output selection.

Note:
You need to run the analysis again if you change the HDR grading. Therefore, we recommend that you first approve the HDR grading before running the Dolby Vision Analysis.

If you have a Dolby Vision trim license loaded, additional controls show up that let you modify the mapping using the Dolby Vision Trim controls.

Dolby Vision Trim Controls

Trimming for SDR Rec.709 100 nits

To use the Dolby Vision Trims,

Navigate to the Dolby Vision tab
Set Target Display Output to: 100-nit, BT.709, BT.1886, FULL (HOME)
Set Trim controls to: 100-nit, BT.709, BT.1886, FULL (HOME)

Trimming for other Home Targets

You can also check the mapping for other targets and use trims, if needed. In this example 600-nit, P3, D65, ST.2084, Full (HOME).

Set Target display Output to: 600-nit, P3, D65, ST.2084, Full (HOME)
Review the mapping at 600 nits (the current Rec 709 trims will be interpolated)
- Optional:
  - Set Trim controls to: 600-nit, P3, D65, ST.2084, Full (HOME)
  - Create new trims for 600 nits, if required

Note:
As the trims interpolate, it’s advised to review and trim the 100nit SDR version first, check the interpolated versions as described above, and only create new trims if needed!

The Dolby Vision Trim controls in version 4 offer primary and secondary controls to adjust the mapping for each target display output. Ensure to select the corresponding Trims Controls for the selected Target Display Output!

Mid Tone Offset – L3

Dolby Vision CMv.4.0 trim controls allow you to offset the initial analysis using the Mid Tone Offset control. Therefore, this trim control is in the Dolby Vision Analysis section and it is particularly useful for matching the overall exposure of the mapped SDR signal to the HDR master. It acts as an offset to the L1 mid values and adjusts the image’s mid tones without affecting the blacks and highlights.
It is recommended to first trim the mapped output with the Mid Tone Offset control before adding additional primary and secondary trims.

This global offset is per-shot and applies to all targets for the shot and gets recorded as Level 3 metadata.

Unlike the L3 Mid Tone Offset control, the primary and secondary trim controls only affect the selected trim target.

To learn more about the Dolby Vision Trim controls, visit this Best Practices Article.

Note:
On some consumer display devices, the current implementation of subtitles over HDMI applies Lift trim values to the black level of the letterbox.

Consequently, avoid positive Lift values over 0.025. If the scene contains positive Lift values over 0.025, this may cause the letterboxed area to have elevated black levels during playback.
If the image aspect ratio matches the canvas aspect ratio, so there is no letterbox, positive Lift can be used without constraints.

Both primary and secondary trim values are stored as Level 8 metadata when using Dolby Vision CMv.4.0.

Dolby Vision Trim Controls – edit functions

The Dolby Vision trim controls can be applied on as many shots as needed. Trim metadata can be copied and pasted onto any number of shots.
Resolve also lets you copy and paste analysis metadata. However, analysis metadata should be calculated individually for each shot and copied from one shot to another only for very specific use cases.

Export in Dolby Vision

Export a Dolby Vision XML

You need to supply a Dolby Vision XML if your deliverable is TIFF 16-bit RGB or ProRes 4444 (XQ).

Go to the Media or Edit page and select the current Timeline
Right-click – Timelines - Export – Dolby Vision XML
Select a filename for the Dolby Vision XML
Press save and select an export destination
Select XML Version 4.0
- Select v5 XML only to include L11 as well as Cinema trims metadata to your XML
- If not explicitly requested and outlined in the delivery specifications of your customer, you should export a version 4 XML
Set the Color Encoding Primaries to match the output color space and press export
- in this example P3-D65

Export a Dolby Vision Master file

Navigate to the Delivery page
Select a suitable Dolby Vision Master or Mezzanine file format

Supported formats include

ProRes 4444 (XQ)
TIFF - RGB 16-bit uncompressed
IMF Package (mxf)
- Generic App2e Dolby Vision
- Netflix HDR - Dolby Vision
- Supplemental IMF - Dolby MXF

Export in ProRes or TIFF RGB 16-bit

From the Render Settings:

Select Custom Export
Choose a File Name and render Location
Select Format Tiff for a TIFF export
Choose Format QuickTime and Codec ProRes 4444 (XQ) for a ProRes export
Set Resolution to 3840x2160 UltraHD
Leave the Data Levels set to Auto for all Format and codec options
Leave the Color space and Gamma Tag set to Same as Project
Select Add to Render Queue
Select Render All from the Render Queue

Export a Dolby Vision IMF Package

From the Render Settings:

Select IMF Netflix from the IMF dropdown
Choose a File Name and render Location
Set Preset Type to Dolby Vision UHD
Set the correct Frame rate
Leave the Data Levels set to Auto for all Format and codec options
Leave the Color space and Gamma Tag set to Same as Project
Set the composition settings as defined by the delivery specifications.
Select Add to Render Queue
Select Render All from the Render Queue

Note:

Other presets, such as the Disney preset, do not include the Dolby Vision metadata in the video essence of the IMF package. At the time this video was recorded, this studio requires a standalone sidecar for IMF packages in their delivery specifications.