Introduction:

This article overviews Dolby’s recommended specifications and testing criteria for HDR displays used for Dolby Vision content creation and workflows. The information provided in this guideline can be utilized by display manufacturers who want to ensure their products meet Dolby’s quality standards and by individual users who are performing their own display validations. 

 

Overview:

 

1. Dolby does not certify specific displays or publish a list of preferred HDR displays.

 

2. The recommended minimum display specifications outlined in section 1 of this document are highly encouraged by Dolby but ultimately, this is up to the discretion of the facility based on their client’s needs. Dolby always recommends using the highest quality (reference grade) display available when creating Dolby Vision content for film or television.

 

3. This guideline is focused on evaluating display performance of calibrated displays, please ensure the specific display is calibrated prior to evaluation measurements.

 

4. These guidelines were designed for evaluating direct view flat panel displays, other types of playback devices (such as projection systems or head-mounted displays) are outside of the scope of this document.

 

5. Proper testing equipment and display setup are required to obtain accurate measurement readings during an evaluation, refer to Dolby’s best practices below.

 

1. Recommended Specifications:

For creating Dolby Vision content, Dolby classifies HDR displays into two distinct tiers:

 

Tier 1: Reference-level HDR display well suited for any professional color grading or mastering scenario, quality control, as well as all applications outlined below for Tier 2 displays.

 

Tier 2: Mid-end or ‘prosumer’ level HDR display that can be utilized for applications outside of final color grading or mastering such as pre-grading, editing, VFX, review and approval, or educational purposes.

 

Note: Pre-grading refers to a basic color pass to get the general look/feel of the content. Depending on the content, Tier 2 displays may clip certain shots that contain luminance values above their peak luminance specification, making any grading or mastering decisions made around these areas inaccurate.

 

These tiers should serve as general guidelines or starting points when reviewing HDR display options and are also fairly aligned with the EBU’s HDR display classifications (as specified in EBU Tech 3320). Dolby recommends the following minimum specifications for these types of HDR displays:

 

	Tier 1	Tier 2
Peak Luminance	≥ 1,000 Nits	≥ 600 Nits
Black Level	≤ 0.001 Nits	≤ 0.005 Nits
Intra-Frame Contrast Ratio	≥ 1,000,000:1	≥ 200,000:1
Color Gamut	P3, BT.2020, BT.709
White Point	D65 (0.3127, 0.3290)
EOTF	SMPTE 2084 (PQ), BT.1886 or Gamma 2.4
Grayscale △EITP	Average ≤ 2.0	Average ≤ 3.0
Color Checker △EITP	Average ≤ 2.0	Average ≤ 3.0
Supported Signal Bit Depth	12 Bit	10 Bit
Chroma Sub-Sampling	RGB 4:4:4

 

Note: The specifications listed above should be measured in accordance with their corresponding test criteria outlined in section 4 of this document.

 

2. Testing Equipment:

When performing display evaluation and calibration measurements, certain tooling is needed to measure light output from your display:

 

1. Measurement Device (colorimeter or spectroradiometer) to take readings from display.
2. Pattern Generator to drive the display with the appropriate and accurate test patches.
3. Control Software is utilized to streamline the calibration process and report/record results.

 

Note: It is possible to use color correction software such as DaVinci Resolve to output the test images to your display if your color management pipeline and output hardware is configured correctly

 

There is a plethora of variables that can cause inconsistencies when performing display measurements. It’s important to ensure you’re utilizing the correct equipment and that it is configured properly to avoid improper signal management that could skew measurement results. To achieve the most accurate results, Dolby recommends the following best practices:

 

• Use a high precision spectroradiometer measurement device such as the Colorimetry Research CR-250/300 or Photo Research PR-740 to minimize errors caused by panel variability and differing display backlight technologies (with varying spectral power distributions).
• If utilizing a colorimeter device, always create a calibration (correction) matrix profile for the meter with a spectroradiometer for each specific display you measure (regardless of the displays backlight technology or model). Variability can be seen between two displays of the same model and between displays that utilize the same panel.
• Regularly send measurement tooling in for manufacturer calibration (to minimize drift) as per manufacturer’s recommendations.
• Ensure your pattern generator output matches your calibration targets.

 

For more information on display calibration principles and tooling, refer to our corresponding articles: Display Color Calibration and Colorimeters vs. Spectroradiometers.

 

3. Display Setup:

When evaluating a display’s performance, several image quality specifications should be measured to verify the display is outputting accurate color and luminance values that adhere to industry standards for content creation. To ensure the most consistent measurement results, Dolby recommends the following best practices prior to performing an evaluation:

 

• Allow the display to warm up for 30 minutes before taking any measurements (this is often referred to as ‘ageing’ the display). Display dynamic (moving and variable) content to avoid over-heating and burn-in.
• Disable any dynamic image settings (if applicable) that impact image rendition such as: auto brightness limiters, noise reduction, motion smoothing, dynamic contrast, etc. These invalidate the source signal and will cause erroneous measurements.
• Ensure the display is configured properly for the given test parameter or calibration target. Example SDR & HDR calibration targets can be seen below:

   

  
  SDR: ITU-R BT.709 

  Signal Format	RGB 4:4:4
  Signal Range	Full
  EOTF	Gamma 2.4/BT.1886
  Black Level	0.05 nits
  Peak Luminance	100 nits
  Color Gamut	BT.709
  White Point	D65 (XW 0.3127, YW 0.3290)

   

  HDR Tier 1: P3 PQ and ITU-R BT.2100 PQ

  Signal Format	RGB 4:4:4
  Signal Range	Full
  EOTF	SMPTE ST 2084
  Black Level	0.001 nits or lower
  Peak Luminance	1000/2000/4000 nits (depending on display’s capabilities)
  Color Gamut	P3 or BT.2020 (depending on delivery specifications)
  White Point	D65 (XW 0.3127, YW 0.3290)

  
  Note: Dolby advises against the use of modified or adapted white point targets such as Judd-Voss (unless explicitly required by your client or delivery specifications).
  
  
• Perform a routine calibration before running the Dolby Vision evaluation criteria. Unless you’re explicitly looking to measure out of the box performance, a basic calibration is recommended to ensure the display is configured accurately.
• Always perform display testing in the environment the display will eventually reside and be used. Different ambient conditions can greatly influence measurement results and display performance.

 

4. Testing Criteria:

The following parameters encompass Dolby’s recommended testing criteria for HDR display evaluations. For these tests, configure the display to the HDR: P3 PQ settings described above in section 3. To generate the necessary test patches, Dolby recommends utilizing software such as Portrait Display’s Calman or Light Illusion’s ColourSpace for a streamlined evaluation experience.

 

Peak Luminance (L_W)

DESCRIPTION: Peak Luminance is one of the first specifications to consider when choosing a display as it’s one of the main components that describe HDR imagery. Peak luminance describes the highest possible luminance level the display can produce.

 

EXPECTED BEHAVIOR: It’s expected that this measurement would yield a peak luminance value that matches closely with the manufacturer’s advertised specification. Likewise, when presented with a code value that corresponds to a luminance output that his higher than the limitation of the display, the display should hard clip and not perform any sort of highlight roll-off.

 

Note: Luminance values may drift with operation and overall usage time.

 

TESTING PROCEDURE: Measure peak luminance with an L18 Loading Pattern. The L18 loading pattern is a white rectangle centered in a black surround that covers ~3% of the total screen area. The white patch should correspond to full drive luminance code values (4095 for 12-bit image, 1023 for 10-bit image, and 255 for 8-bit image).

 

Note: The number associated with the Loading Pattern (ie. ‘L18’) refers to the diagonal of the pattern in respect to the display diagonal (refer to IDMS 5.24: LOADING). This can be calculated via the following equation:

 

 

Note: A 3% window differs from EBU’s recommended 1% as HDR display specifications have evolved since the original date of that publication. Tier 1 displays should be capable of achieving 1,000 nits at a 3% window without any power limiting.

 

Black Level (L_K)

DESCRIPTION: Black Level describes the darkest value that the display can output (while turned on).

 

EXPECTED BEHAVIOR: It’s expected that this measurement would yield a black level value that matches closely with the manufacturer’s advertised specification. For HDR content creation, Dolby recommends utilizing displays with at least a black level of 0.005 or lower (and ideally 0.001 or lower for Tier 1 displays)

 

TESTING PROCEDURE: Measure black level with a static full screen black test chart that corresponds to code value 0.

 

Note: Some displays (especially QD-OLEDs & OLEDs) may produce a black level that is beyond the sensitivity of your measuring device. In these instances, refer to the manufacturer’s specification.

 

Note: The Tier 1 black level specification value of 0.001 outlined above differs from EBU’s recommended 0.005 as HDR display specifications have evolved since the original date of that publication (especially with the increasing number of OLED-based displays and improvements being made to full array local dimming algorithms seen on more advanced LCD displays).

 

Contrast Ratio (CR)

DESCRIPTION: Contrast Ratio refers to the relationship (ratio) of the peak luminance to that of the black level the display can reproduce. An important distinction between a display’s contrast ratio is the difference between intra-frame and inter-frame contrast ratio. Intra-frame contrast ratio describes a display’s contrast ratio measured within a single frame. Inter-frame contrast ratio describes the display’s contrast ratio measured with two different frames.

 

EXPECTED BEHAVIOR: Typically, manufacturers will disclose inter-frame contrast ratio specifications (as these values are typically higher and imply added benefit/performance). Intra-frame contrast is more relevant for content creation applications as these scenarios are more indicative and relevant for real-world scenes (with varying dynamic ranges and average picture levels). In an ideal world, intra-frame and inter-frame contrast ratio values would be the same.

 

TESTING PROCEDURE: For the purpose of this evaluation, we’ll use intra-frame contrast ratio as it most relates to real world HDR Dolby Vision content creation scenarios. To calculate this contrast ratio, measure the center of the corner contrast ratio patch (pictured below) this is the contrast ratio black level L_KCR.

 

        

Example corner contrast ratio test pattern used for calculating intra-frame contrast ratio (left) vs. full screen black pattern used to calculate inter-frame contrast ratio (right)

 

Compute your contrast ratio via dividing your previously measured peak luminance to this black level:

 

CR = L_w / L_KCR

 

If you wish to calculate the inter-frame contrast ratio, you can simply divide the previously captured peak luminance and black level values (CR = L_W/L_K)

 

Gray Scale (PQ) Tracking

DESCRIPTION: This measurement ensures that the display is tracking the ST 2084 (PQ) HDR EOTF properly across the entire luminance range.

 

EXPECTED BEHAVIOR: The measured DE_ITP average should be less than 2.0 up to the display’s peak luminance. Test patches that correspond to values above the display’s peak luminance capabilities should result in clipping at the peak luminance value. There should be no roll-off of peak luminance.

 

Note: DE_ITP is a color difference metric developed by Dolby to better quantify the performance of HDR and WCG displays. Refer to whitepapers in the Appendix for more information.

 

Note: It is common to see most of the error in the low code value areas of this measurement (as the display comes out of black). This can be attributed largely to display variability and probe measurement tolerances.

 

TESTING PROCEDURE: Load a series of L18 gray scale images. Dolby recommends utilizing at least a 20-point grayscale series. Example code values for these test images can be seen below. Measure each step and compute the DE_ITP average of the series.

 

Test Image	12 Bit Code Value	Display Luminance (nits)	10 Bit Code Value	Display Luminance (nits)
1	0	0.000	0	0.000
2	64	0.005	16	0.005
3	128	0.022	32	0.022
4	256	0.101	64	0.101
5	481	0.500	120	0.500
6	614	1.000	156	1.000
7	771	2.002	193	2.015
8	952	4.006	238	4.016
9	1069	6.009	267	6.006
10	1157	8.016	289	8.012
11	1228	10.020	307	10.051
12	1462	20.000	366	20.176
13	1717	40.000	429	40.030
14	1875	60.080	469	60.439
15	1990	80.080	497	79.975
16	2081	100.100	520	100.230
17	2371	199.700	592	199.153
18	2672	399.700	668	401.505
19	2851	599.600	712	598.320
20	3078	998.400	769	998.932
21	3388	1999.000	847	2009.799
22	3696	4000.000	923	3987.985

 

Note: Refer to ITU-R BT.2124 on how to calculate DE_ITP or utilize one of the calibration tools that automatically measure and calculate DE_ITP (such as Portrait Displays’ Calman or Light Illusion's ColourSpace software).

 

 

RGB Additivity

DESCRIPTION: RGB Additivity testing verifies that the display allocates appropriate luminance for each color channel in a linear fashion (to match the peak white luminance, as white is a sum of RGB)

 

EXPECTED BEHAVIOR: The sum of the luminance of Red, Green, and Blue is approximately equal to the luminance of peak white (L_R + L_G + L_B = L_W)

 

TESTING PROCEDURE: Measure full drive R, G, B, and white with corresponding L18 patches in CIE XYZ. Calculate the additive white W_{A{X, Y, Z}} by summing the measurements of R, G, and B:

 

R_X + G_X + B_X = W_A,X

R_Y + G_Y + B_Y = W_A,Y

R_Z + G_Z + B_Z = W_A,Z

 

Compare the calculated additive white W_A{X,Y,Z} against the measured display white W_X,Y,Z  and calculate the additivity ratio A_{X,Y,Z} via:

 

A_{X,Y,Z} = W_A{X,Y,Z} / W_{X,Y,Z}

 

If the display is additive, all three values will be close to 1.0. If the values are substantially larger, then the white luminance level is higher than the sum of the primaries. Similarly, if the values are negative, this means the display white is darker than the sum of the primaries (and there may be some active luminance compression active such as power management/dimming). The accepted tolerance is [-0.01 to +0.05] which represents a luminance discrepancy of -1 to 5%.

 

Note: Some display types that feature designs containing more than 3 subpixels (such as WOLED) may struggle with additivity measurements

 

Note: Calman users can measure RGB Additivity via the 'Additivity' page found in the SDR & HDR Analysis workflows and ColourSpace users can calculate RGB Additivity throughout the entire luminance range and view any deviations in the 'RGB Separation' graph.

 

Color Checker

DESCRIPTION: The Color Checker test verifies the display’s ability to accurately render colors that are commonly associated with content creation (skin tones, daylight, etc.).

 

EXPECTED BEHAVIOR: The measured DE_ITP average should be less than 2.0.

 

TESTING PROCEDURE: Measure L18 color checker patches found in Portrait Displays Calman “HDR Analysis” workflow and compute the DE_ITP average. For ColourSpace, use a .csv patch set with defined patch colors and the resulting DE_ITP average will be calculated and displayed automatically.

 

HDR Color Gamut Coverage (P3/2020)

DESCRIPTION: HDR Color Gamut Coverage describes the total percentage of a given color gamut (P3/BT. 2020) that the display in question can reproduce. 

 

EXPECTED BEHAVIOR: The measured coverage matches the advertised specifications. For typical reference grade HDR displays, 95-100% P3 coverage is expected.

 

TESTING PROCEDURE: Dolby recognizes there is significant active research being worked on to develop a proper test that accurately summarizes this coverage in a more volumetric way (rather than a traditional single luminance 2D plot). Until the ideal metrology is more developed or standardized, users can opt into using the gamut coverage tests found in industry calibration software such as Calman.

 

Peak Luminance vs. Window Size

DESCRIPTION: This test is designed to determine when a specific display implements dimming in respect to peak brightness. Knowing this threshold is useful for content creators that deal with scenes that feature high APL as their display may not be showcasing the full brightness capabilities of the panel. Likewise, many manufacturers do not disclose how or at what patch size their peak brightness specification was calculated.

 

EXPECTED BEHAVIOR: Typically, peak luminance decreases with window size. This behavior is more common on OLED displays whereas LCD-based devices are more often able to retain their peak luminance at larger patch sizes.

 

TESTING PROCEDURE: Measure a series of increasing in size full drive white patches (L10, L18, L20, L32, etc.) up until full screen white and record the luminance levels at each step. Compare the results to advertised specifications and keep a record of when significant drops in brightness occur.

 

 

Additional Considerations:

 

Viewing Angle

Viewing angle variance is very common and an important factor to consider when evaluating any display for content creation. Many professional grading or mastering environments are utilized by multiple people during collaborative review sessions that are viewing the same display from different positions, which can cause differences in content perception

 

When viewing content off-axis of a given display, there is a perceived change in both color and luminance appearance (commonly manifested in hue shifts or drops in luminance). These shifts become more apparent the further off-axis you view the display. These viewing angle issues are mostly seen on LCD-based displays, with OLED technologies exhibiting much better off-axis performance.

 

Uniformity

Panel uniformity is another important aspect to consider when evaluating an HDR display for content creation. Ideally, the display can showcase both fullscreen color and white patches (at various luminance levels) uniformly, with no visibly obvious dark spots or hue/brightness shifts. This is especially important at low luminance levels (such as fades from black, etc.).

 

For more information on evaluating viewing angle and uniformity performance, refer to sections 1.5.8  and 1.5.14 of EBU Tech 3320.

 

5. Appendix:

Dolby Professional Support:

https://professionalsupport.dolby.com

 

Dolby Vision HDR Display Considerations:

https://professionalsupport.dolby.com/s/article/What-display-should-I-use-for-creating-Dolby-Vision

 

Dolby Vision Self-Guided Curriculum:

https://learning.dolby.com

 

EBU Tech 3320:

https://tech.ebu.ch/docs/tech/tech3320.pdf

 

SMPTE ST 2084:

https://ieeexplore.ieee.org/document/7291452

 

ITU-R BT.709:

https://www.itu.int/rec/R-REC-BT.709

 

ITU-R BT.2020:

https://www.itu.int/rec/R-REC-BT.2020

 

ITU-R BT.2100

https://www.itu.int/rec/R-REC-BT.2100

 

ITU-R BT.2124:

https://www.itu.int/rec/R-REC-BT.2124

 

Portrait Displays Color Difference Metric Whitepaper:

https://www.portrait.com/resource-center/eitp-is-now-itu-r-bt-2124-is-the-industry-ready-to-move-on-from-e2000/

 

International Committee for Display Metrology

https://www.sid.org/Standards/ICDM