HDR10, HDR10+, Dolby Vision, and HLG: What Actually Differs

Open a television specification and you may see:

• HDR10
• HDR10+
• Dolby Vision
• HLG
• Dolby Vision IQ
• Dolby Vision 2
• HDR10+ Adaptive
• HDR10+ Advanced

The list can look like several unrelated picture technologies.

The useful way to understand it is to separate the layers.

An HDR format can define or depend on:

• transfer function
• color primaries
• signal representation
• bit depth
• chroma sampling
• codec profile
• static metadata
• dynamic metadata
• signaling
• compatibility structure
• certification
• display mapping

The four familiar names are not equal items in one simple comparison.

HDR10 is a baseline PQ media profile assembled from several standards.

HDR10+ adds scene- or frame-dependent color-volume-transform metadata to an HDR10-compatible base.

Dolby Vision is a proprietary family of profiles and workflows with Dolby metadata and content-mapping systems. Many home-entertainment workflows use PQ, but Dolby also supports HLG-based profiles.

HLG is a separate relative HDR television system defined in ITU-R BT.2100-3. Its basic display relationship does not require production metadata. ^[1][2]

The logos matter.

They do not tell the whole picture.

A format has more than one layer

Two signals can use the same transfer function and still differ in:

• bit depth
• chroma format
• color representation
• metadata
• codec
• profile
• mastering
• fallback structure
• display processing

Two signals can also use different transfer functions while targeting the same HDR-capable television.

This is why the statement:

The formats use the same pixels and differ mainly in metadata.

is too broad.

HDR10 and HDR10+ are closely related.

Dolby Vision can use different profile structures.

HLG uses a different HDR architecture.

What HDR10 actually is

HDR10 is often called a standard.

A more precise description is:

HDR10 is an industry media profile that combines several standards and metadata items.

The Consumer Technology Association definition reviewed for this article lists:

• SMPTE ST 2084 EOTF
• ten-bit video
• BT.2020 color primaries
• 4:2:0 chroma for compressed video sources
• SMPTE ST 2086 mastering-display metadata
• MaxFALL
• MaxCLL ^[5]

No one item defines the complete format.

PQ is the transfer function

SMPTE ST 2084 defines the PQ electro-optical transfer function.

Its normalized reference range maps to display luminance from 0 to 10,000 cd/m². ^[3]

The existence of that range does not mean:

• every HDR10 title reaches 10,000 cd/m²
• every mastering display reaches it
• every consumer television must reproduce it directly
• every image should be brighter overall

The display may need to map the source into its own luminance and color volume.

Ten-bit is the profile depth

The CTA HDR10 definition specifies ten-bit video. ^[5]

That is why the format name contains 10.

It does not prove:

• twelve-bit was never used in production
• every source element began as ten-bit
• the display panel is native ten-bit
• no banding can occur
• the PQ range ends at 1,000 cd/m²

Bit depth and PQ luminance range are separate properties.

BT.2020 is the signal colorimetry

The HDR10 profile uses BT.2020 primaries for the signal representation.

That does not mean every pixel uses the full BT.2020 gamut.

It does not prove that the mastering display or consumer television reproduces every color inside the BT.2020 triangle.

The display maps the signal into its available color volume.

What static metadata means

Static metadata normally applies to a complete program, asset, or defined sequence rather than changing with every scene.

The HDR10 metadata set includes two different categories:

1. mastering-display color-volume metadata
2. content-light metadata

They answer different questions.

ST 2086 describes the mastering display

SMPTE ST 2086:2018 defines metadata for the display configured and used in mastering.

The set includes:

• display primary chromaticities
• white-point chromaticity
• nominal maximum display mastering luminance
• nominal minimum display mastering luminance ^[4]

This metadata describes the mastering-display color volume.

It does not describe every pixel in the program.

It does not define:

• the mastering EOTF
• the actual brightest content pixel
• the average brightness of the program
• one required tone-mapping curve
• the complete mastering environment
• a guarantee of creative-intent preservation

ST 2086 explicitly says that it does not fully specify the information needed to preserve creative intent on displays with different color volumes. ^[4]

Mastering-display peak is not content peak

A mastering display configured for 1,000 cd/m² does not prove that the content contains a 1,000 cd/m² pixel.

The grade may use less than the display's full capability.

Likewise, a signal can contain isolated values that do not represent the artistic highlight structure the way a simple maximum implies.

Keep these separate:

• mastering-display maximum luminance
• content maximum light level
• maximum frame-average light level
• HDR reference white
• target-display peak luminance

MaxCLL

MaxCLL means Maximum Content Light Level.

ITU-T H.222.0 defines it as the value corresponding to the brightest pixel in the complete stream, expressed in whole cd/m² units. ^[6]

The formal field can also be zero when the value is unknown.

That limitation matters.

A television should not assume that every nonzero value was calculated perfectly or that zero means the content contains no light.

MaxFALL

MaxFALL means Maximum Frame Average Light Level.

ITU-T H.222.0 defines it as the highest frame-average brightness in the complete stream, expressed in whole cd/m² units. ^[6]

It is not:

• the average brightness of the entire movie
• the average brightness of the brightest scene
• the display's sustainable full-screen output
• a required television brightness limit

It is one static summary of the coded content under the specified calculation.

Static metadata can be missing or unreliable

Static content metadata may be:

• calculated from a complete asset
• supplied by a mastering system
• inherited from another deliverable
• set to a predetermined value in a live workflow
• set to zero when unknown
• lost during conversion
• copied incorrectly
• ignored by a display

ITU-T HDR guidance notes that live or broadcast services may use predetermined values rather than a complete measured pass. ^[18]

The television must still produce a usable picture when metadata are incomplete.

HDR10 does not force one fixed tone map

The original article said that a television reads static metadata and uses one tone-mapping setup for every scene.

That is not required by the metadata definition.

A television can:

• use the static values
• ignore them
• treat them as bounds
• analyze frames or scenes
• apply a manufacturer algorithm
• change mapping with picture mode
• react to panel temperature or average picture level
• adapt to the room

The metadata remain static.

The television's processing does not have to be.

This is why two televisions can produce different HDR10 images from the same signal.

What dynamic metadata does

Dynamic metadata is content-dependent information that can change during a program.

SMPTE describes the ST 2094 suite as metadata for color-volume transforms when content is presented on a display with a smaller color volume than the mastering display.

The information can vary by scene or image. ^[8]

Dynamic metadata can help a system account for changing content characteristics rather than relying only on one program-wide set of values.

The exact information depends on the application.

Dynamic metadata is about color-volume transformation

Tone mapping is part of the problem.

Color mapping is also part of it.

A target display may differ from the mastering display in:

• peak luminance
• black level
• primary chromaticities
• saturated-color luminance
• power behavior
• viewing environment

A useful transform must manage luminance and color relationships together.

This is why the SMPTE term is Dynamic Metadata for Color Volume Transform.

Dynamic metadata does not guarantee a better picture

More content-dependent information can help.

It does not guarantee improvement.

The result depends on:

• accuracy of the metadata
• source grade
• target display
• implementation
• picture mode
• calibration
• room light
• whether the metadata survived the chain
• whether the display follows or modifies the intended transform

A well-designed HDR10 mapper can outperform a poor dynamic-metadata implementation on a particular scene.

A dynamic format can also provide a more controlled result when the source and display use it well.

The format logo alone cannot settle the comparison.

HDR10+

HDR10+ builds on an HDR10-compatible PQ signal and adds dynamic metadata.

Its dynamic metadata are based on SMPTE ST 2094-40 Application #4. ^[7][9][10]

SMPTE ST 2094-40:2020 includes:

• scene-adaptive tone mapping
• scene-based color-saturation mapping
• descriptive scene-characteristic metadata
• target-display information
• optional creatively approved curve metadata ^[7]

This is more specific than saying:

HDR10+ tells the television the brightness of each scene.

Brightness statistics are part of the system.

The metadata support a broader color-volume transform.

Scene-based and frame-based HDR10+

HDR10+ Technologies states that:

• video-on-demand content generally uses scene-based data
• a scene can be one frame or a sequence
• HDR10+ live encoding uses frame-based data ^[9]

The cadence does not prove that every frame receives a unique creative adjustment.

It describes how metadata can be applied.

HDR10+ fallback

HDR10+ is designed around an HDR10-compatible base.

HDR10+ Technologies states that an HDR10-only display can play the content as HDR10 while ignoring the unsupported HDR10+ metadata. ^[9]

This is a useful compatibility feature.

The exact behavior still depends on:

• transport
• codec
• container
• receiver
• source device
• display

A logo on one component does not prove end-to-end metadata delivery.

HDR10+ licensing and certification

HDR10+ Technologies describes HDR10+ as an open standard and says its certification and logo program operates without per-device or per-unit royalties payable to the organization. ^[11]

That does not mean there are no conditions.

An adopter may need to follow:

• specifications
• compliance tests
• certification requirements
• logo terms
• patent and intellectual-property provisions
• product registration procedures

The accurate shorthand is:

HDR10+ uses a royalty-free certification and logo model under the program's stated terms.

Dolby Vision

Dolby Vision is a proprietary HDR production, metadata, delivery, and display ecosystem.

It includes multiple:

• profiles
• levels
• base-layer arrangements
• enhancement arrangements
• metadata versions
• delivery containers
• device-mapping paths
• compatibility options

The name Dolby Vision does not identify all of those details by itself.

Dolby Vision is not universally twelve-bit

Dolby Vision workflows may use high-precision masters.

Consumer delivery structures can use:

• ten-bit base layers
• enhancement information
• profile-specific processing
• metadata-guided reconstruction
• HLG-compatible base layers

The statement:

Dolby Vision uses twelve-bit color.

is incomplete without a profile and stage.

The delivered bitstream precision, reconstructed processing precision, and panel precision are separate questions.

Dolby Vision can use PQ or HLG

Many post-produced Dolby Vision movies and series use PQ-based workflows.

Dolby also states that Dolby Vision supports both PQ- and HLG-based transfer functions. ^[13]

A specific example is Dolby Vision Profile 8.4, which uses an HLG BT.2020-compatible base layer with Dolby Vision metadata. ^[14]

Therefore:

• HDR10 is PQ
• HDR10+ is PQ
• many Dolby Vision profiles are PQ-based
• Dolby Vision is not universally PQ-only
• HLG remains its own BT.2100 HDR system

Dolby Vision metadata

Dolby's content-creation documentation describes an analysis that generates core values for each shot, representing minimum, average, and maximum luminance characteristics.

The workflow can then add artistic trim metadata for selected target displays.

The combined metadata travel with the HDR image and can assist mapping to consumer devices and derived outputs. ^[12]

Dolby documentation also distinguishes:

• global metadata
• track metadata
• shot-based metadata
• frame-based metadata in selected cases ^[15]

This is more complex than one brightness number per scene.

Dolby Vision mapping is not identical on every device

The final result can depend on:

• profile
• source-led or display-led processing
• decoder
• target display capability
• manufacturer integration
• picture mode
• ambient-light processing
• firmware
• calibration

Dolby Vision provides a coordinated system.

It does not make all televisions physically identical.

A lower-peak display still has a lower peak.

A smaller color volume still requires mapping.

Dolby Vision licensing needs precise wording

Dolby Vision is proprietary.

Dolby states that content creators can use integrated analysis and mapping functions in supported grading systems without an additional Dolby license, and that it does not charge per-title or overarching studio fees for creating and delivering Dolby Vision content.

Dolby also offers paid artistic trim controls and licenses consumer implementations through its technology ecosystem. ^[12]

Avoid the blanket statements:

• every creator pays Dolby
• Dolby Vision content creation is free in every tool
• the only cost is a television license
• every device license has the same terms

Dolby Vision IQ

Dolby Vision IQ is a branded viewing feature that combines Dolby Vision information with television and viewing-environment inputs.

Its behavior can vary by:

• manufacturer
• model
• sensor
• mode
• firmware
• content
• room

It should be treated as a documented product feature, not as a new transfer function.

The article does not claim that it reproduces the unchanged reference grade in every dark room or that it should always be enabled.

HLG

Hybrid Log-Gamma is one of the two HDR television systems in BT.2100-3. ^[1]

Unlike PQ, HLG uses a relative scene-light signal.

The displayed luminance depends on:

• nominal peak luminance
• black level
• system gamma
• viewing conditions

HLG is not simply HDR10 without metadata.

It is a different system architecture.

HLG requires no production metadata for basic rendering

ITU-R BT.2390-12 states that HLG requires no production metadata for its basic system and was designed for an evolutionary production and distribution approach. ^[2]

That can be useful in live and broadcast workflows.

The receiver still needs to know that the signal is HLG.

“No metadata” is too broad

An HLG service still needs or can contain information about:

• transfer characteristic
• colorimetry
• range
• bit depth
• codec
• resolution
• frame rate
• audio
• captions
• service identity
• mastering or ambient information where a system includes it

The accurate statement is:

HLG does not require production metadata to establish its basic HDR display relationship.

HLG compatibility has limits

HLG was designed to provide a significant degree of compatibility with suitable SDR equipment and workflows.

ITU-R BT.2390-12 limits the direct display claim to SDR receivers that support the BT.2020 color container.

Legacy BT.709 distribution still needs appropriate conversion or a separate service.

Do not tell the reader that every old television can display HLG correctly.

Static and dynamic metadata are not format quality scores

Static metadata are not automatically bad.

Dynamic metadata are not automatically good.

Static metadata can be sufficient when:

• the target display reproduces the source closely
• the television analyzes the image well
• the content has modest variation
• the static values are reliable

Dynamic metadata can be valuable when:

• the source range exceeds the target
• scenes vary substantially
• the metadata contain useful creative guidance
• the display implements the transform well

The viewer sees the combined system.

Compatibility is profile-specific

HDR10+

HDR10+ normally retains an HDR10-compatible presentation when the dynamic metadata are unsupported.

Dolby Vision

Fallback depends on the Dolby Vision profile and delivery.

A Dolby Vision asset may use:

• an HDR10-compatible base
• an HLG-compatible base
• another supported profile structure
• an ecosystem that negotiates another asset

Do not assume every unsupported Dolby Vision signal becomes HDR10 automatically.

HLG

HLG can be converted or displayed through compatible HLG and selected SDR workflows.

It is not an HDR10 fallback layer.

Newer names: Dolby Vision 2

Dolby announced Dolby Vision 2 on September 2, 2025.

The announcement describes a redesigned image engine, content- and environment-aware features, and new tone-mapping tools. ^[16]

This is a developing ecosystem.

An announcement or manufacturer commitment is not proof that:

• the feature is present on every Dolby Vision television
• existing content contains all new metadata
• one model implements every feature
• the result is independently measured to be better

Use the exact television, firmware, content, and mode.

Newer names: HDR10+ Advanced

HDR10+ Technologies announced HDR10+ Advanced in December 2025 and published its white paper in February 2026.

The owner describes it as backward compatible with HDR10+ and HDR10 and adds new metadata capabilities for areas including color, motion, local mapping, and environment-aware presentation. ^[17]

As with Dolby Vision 2, support must be confirmed at the content, source, transport, and display stages.

The base explanation remains:

• HDR10 is the PQ baseline
• HDR10+ adds dynamic Application #4 metadata
• newer extensions add optional capabilities

What format the viewer actually receives

The answer depends on the complete chain.

Content asset

The provider must have the format version for that title.

App or service

The app must request and receive it for the:

• region
• account
• subscription
• device
• operating system

Source device

The source must decode or pass the profile.

Intermediate devices

An AVR, soundbar, switch, splitter, capture device, or processor must pass the required signal and metadata.

Display

The television must support the exact format or profile on that input or application path.

Firmware and settings

A feature can depend on:

• firmware
• enhanced input mode
• output setting
• refresh rate
• resolution
• bit depth
• chroma format

This is why a service logo does not prove the television will receive that format.

Automatic selection is normal, not infallible

Most systems attempt to choose a mutually supported format.

The selected result may still differ from the reader's expectation because:

• the title lacks the asset
• the app uses a different profile
• the source is configured to convert HDR
• the receiver blocks the metadata
• the television does not support that profile
• the service sends HDR10 as a baseline
• the device is in an incompatible timing
• a firmware issue changes detection

The format banner is useful evidence.

It is not a complete measurement.

When the expected format does not appear

Use this order.

1. Confirm the title and service asset

Do not rely only on a catalog badge.

The badge may describe support on some devices rather than the exact stream being played.

2. Check the app on the exact device

An internal television app and an external streaming box may receive different formats.

3. Check direct connection

Connect the source directly to the television as a diagnostic test.

If the format appears, an intermediate device may not pass it.

4. Check the television input mode

Use the exact manual for enhanced HDR or high-bandwidth input settings.

5. Check source output

Avoid forcing a conversion that prevents the native format.

6. Check firmware

Source, receiver, app, and television updates can change format support.

7. Check the cable when link symptoms justify it

A certified cable should match the required bandwidth.

Cable problems more commonly produce:

• blanking
• sparkles
• dropouts
• failure to establish the intended timing

Do not assume that a cable selectively removed Dolby Vision or HDR10+ metadata while preserving a clean HDR10 signal.

8. Verify the actual signal

Use reliable device information, test equipment, or known content.

Do not infer the format only from subjective brightness.

Metadata is input, not the finished picture

A television receives:

• image essence
• transfer-function signaling
• colorimetry
• static metadata
• dynamic metadata where present
• device and environment information in some systems

The display then produces light within its capabilities.

The result depends on:

• EOTF tracking
• display mapping
• color-volume mapping
• black level
• peak luminance
• panel behavior
• local dimming
• automatic brightness behavior
• room light
• picture mode
• calibration

A more elaborate metadata system cannot create physical capability the display does not have.

Why two televisions still look different

Two televisions can support the same format and have similar advertised peak luminance.

They can still differ in:

• sustained luminance
• black level
• near-black behavior
• saturated-color luminance
• tone-mapping priority
• highlight clipping
• local dimming
• panel uniformity
• white-point accuracy
• metadata interpretation
• ambient adaptation
• firmware

Format compatibility is one requirement.

It is not a complete performance specification.

What calibration can verify

For each HDR mode, calibration and controlled testing can evaluate:

• PQ or HLG EOTF behavior
• reference white
• grayscale
• color gamut and color volume
• clipping
• tone-mapping roll-off
• metadata response
• mode switching
• range and colorimetry
• input versus internal-app differences
• peak and sustained luminance
• ambient-light behavior

Record:

• exact model and size
• region
• firmware
• source
• app
• content
• format
• profile where known
• picture mode
• metadata state
• input
• receiver
• cable
• pattern source
• measurement date

One format logo does not replace those conditions.

What calibration cannot prove

Calibration cannot:

• reconstruct lost source information
• prove the exact creative intent from metadata alone
• make unsupported metadata work
• turn HDR10 into Dolby Vision
• turn HLG into PQ without conversion
• make every format look identical
• make a consumer television physically identical to a mastering display
• make copied panel-specific values valid for another unit

Calibration can reduce measurable error in a selected mode under stated conditions.

Where this leaves us

HDR10, HDR10+, Dolby Vision, and HLG are not four versions of the same pixel stream.

HDR10 is a baseline PQ media profile built from several standards.

Its static metadata include:

• ST 2086 mastering-display color-volume information
• MaxCLL
• MaxFALL

ST 2086 does not describe the content peak or define tone mapping.

Static metadata do not force a television to use one fixed curve for an entire program.

HDR10+ retains an HDR10-compatible base and adds ST 2094-40 dynamic metadata for scene-based color-volume mapping.

Dolby Vision is a proprietary family of profiles and workflows.

Many versions use PQ.

Some use HLG-compatible structures.

Dolby Vision is not universally a twelve-bit consumer signal.

HLG is a relative BT.2100 HDR system whose basic display relationship requires no production metadata.

It still needs signaling and correct receiver processing.

Dynamic metadata can give a compatible display more content-dependent guidance.

It does not guarantee a better picture.

The final image depends on the content, metadata, source, transport, display, algorithm, room, mode, and calibration.

The useful question is not only:

Which HDR logo is present?

It is:

What signal and metadata reached this display, how did this model use them, and how closely did the resulting light output approach the applicable target within the display's capabilities?