HDR Video Processing and Delivery: A Beginner’s Guide to Capture, Encode, and Stream High Dynamic Range Content
High Dynamic Range (HDR) video is revolutionizing visual content by delivering richer highlights, deeper blacks, and a broader color spectrum compared to Standard Dynamic Range (SDR) video. This guide is designed for beginner and intermediate video producers, streaming engineers, and developers interested in HDR workflows. You will learn how to capture, encode, and stream high-quality HDR content, preserving the creative intent throughout the entire process.
What is HDR? (Concepts)
HDR increases the noticeable range between the darkest and brightest parts of an image while often combining it with a wider color gamut, resulting in more realistic and vibrant visuals.
Key Concepts:
- Dynamic Range (Luminance/Nits): The ratio between the darkest and brightest areas of an image, measured in nits (cd/m²). HDR content typically aims for highlights between 1,000 to 4,000 nits.
- Wide Color Gamut: Utilizing color primaries like Rec. 2020 versus Rec. 709 (SDR) allows HDR content to display a more saturated and richer color palette.
- Transfer Functions (EOTFs): These map scene-referred light to display-referred signals, with common standards including PQ (Perceptual Quantizer) and HLG (Hybrid Log-Gamma).
- Mastering Display & Metadata: The reference display used during color grading and metadata like MaxCLL (maximum content light level) and MaxFALL (maximum frame-average light level) help ensure accurate presentation on playback devices.
For a detailed overview of technical standards, refer to the ITU‑R Recommendation BT.2100.
HDR Formats & Metadata
Choosing the right HDR format is crucial due to its impact on device compatibility and final appearance. Here’s a summary of common HDR formats:
Format | Transfer | Metadata | Licensing / Notes |
---|---|---|---|
HDR10 | PQ (ST 2084) | Static (SMPTE ST 2086 + MaxCLL/MaxFALL) | Open, baseline for many services |
HDR10+ | PQ | Dynamic metadata | Open-ish (Samsung/AMZN push), per-scene optimizations |
Dolby Vision | PQ (extended) | Dynamic metadata (scene/frame-level) | Proprietary, licensing required; excellent tone-mapping |
HLG | HLG | Typically no metadata | Designed for live/broadcast, backward-compatible with SDR |
Why Metadata Matters:
- Static Metadata (HDR10): Defines characteristics of the mastering display across the entire file, which may not be optimal for scenes with varying brightness.
- Dynamic Metadata (Dolby Vision, HDR10+): Offers detailed information for each scene/frame, enabling superior tone-mapping on compatible devices. You can read more about the advantages of dynamic metadata in the Dolby Vision whitepaper.
Practical Distribution Strategy:
- For online streaming, many providers offer several HDR variants (HDR10 + Dolby Vision or HDR10+) ensuring optimal quality across devices.
- For broadcasts, HLG is often preferred due to its compatibility with SDR displays.
Capture and Production Workflow
Effective HDR starts with proper capture. If the initial footage lacks detail, it cannot be compensated for during grading.
Shooting Considerations:
- Camera Capabilities: Utilize sensors that capture wide dynamic ranges; consider RAW or high-bit-depth formats like Arri RAW, RED, or log formats.
- Exposure Management: Use zebras, false color, and histograms. Aim to expose properly for highlights while preserving shadow detail.
- Quality Gear: Select high-quality lenses and neutral density filters to control exposure while maintaining depth of field.
Color Pipeline and Grading:
- Utilize a consistent, scene-referred color pipeline. ACES is recommended for effective color management across devices, especially with multiple camera setups.
- Target a mastering display of 1,000 nits for many workflows but push for 4,000 nits in premium projects. Note down your display settings.
- Be cautious with LUTs; prefer 3D LUTs from controlled workflows to maintain highlight roll-off and color fidelity.
Mastering Decisions and Deliverables:
- Identify your target mastering display before final grading. Record MaxCLL and MaxFALL values during QC.
- Produce a mezzanine/master file in high bitrate and upper bit depth, commonly in DNxHR, ProRes, or high-bitrate HEVC formats.
Deliverables Checklist:
- Raw/camera original files
- Conform/timeline project files
- HDR master with metadata annotations
- SDR deliverable or LUT for SDR conversion
- Documentation: including mastering display details and MaxCLL/MaxFALL values.
Encoding and Delivery
Proper encoding of HDR ensures metadata, color space tags, and bit depth are preserved, allowing seamless playback.
Codecs & Containers:
- Common HDR-supporting codecs include HEVC/H.265, AV1, and VP9. For mezzanine formats, consider ProRes or DNxHR.
- Suitable containers are MP4/M4V, MKV, or MPEG-TS for broadcasts, with CMAF recommended for simplified HLS/DASH compatibility.
Bit Depth & Chroma Subsampling:
- Use a minimum of 10-bit depth for HDR encoding to prevent banding. 12-bit is advisable for professional pipelines.
- Consumer streaming typically employs 4:2:0 chroma subsampling but 4:2:2 or 4:4:4 is recommended for high-quality masters.
Metadata Embedding:
- Static metadata (HDR10) is included via HEVC SEI messages and container tags, while dynamic metadata may need extra streams. Use reliable packagers to maintain metadata integrity; recommended options include AWS Elemental, Bitmovin, or Shaka Packager.
Streaming Formats & Manifest Notes:
- Both HLS and DASH can support HDR if streams are correctly labeled. Use CMAF to standardize chunk formats.
- In HLS, utilize EXT-X-DISCONTINUITY and include HDR tags to ensure compatibility across devices.
Bitrate Ladders & CDN Considerations:
- HDR typically demands higher bitrates than SDR, so perform tests across different scenes. CDNs should support your packaging choices effectively.
Example FFmpeg Command:
ffmpeg -i input_hdr_master.mov \
-c:v libx265 -pix_fmt yuv420p10le -preset slow -x265-params "hdr-opt=1:colorprim=bt2020:transfer=smpte2084:colormatrix=bt2020nc" \
-c:a copy output_hdr10.mp4
Ensure to validate metadata with an appropriate tool after processing.
Playback, Tone-Mapping, and SDR Conversion
HDR playback varies across devices, with tone-mapping adapting visuals to display capabilities while striving to maintain creative intent.
Tone-Mapping Strategies:
- Display-Referred Tone-Mapping: This technique adjusts the final image based on playback device capabilities.
- Scene-Referred Workflows: Used during grading (e.g., ACES) to preserve creative intent that is then mapped to the display.
- Devices with dynamic metadata can perform per-scene tone-mapping more accurately.
Creating SDR Versions:
- When converting PQ/HLG to Rec. 709 SDR, ensure careful tone-mapping to avoid clipping highlights or crushing shadows. It’s best to grade for HDR first and then create a dedicated SDR variant or use a high-quality tone-mapping LUT.
Device Compatibility and Negotiation:
- Playback systems negotiate the HDR formats they can handle. Ensure to provide an SDR fallback for devices incompatible with HDR. More details can be found in Microsoft’s HDR overview.
Tools, Testing, and QA
Authoring and Mastering Tools:
- Helpful tools for HDR grading include DaVinci Resolve, Baselight, and Adobe Premiere.
- For HDR metadata authoring, vendors provide specific tools for HDR10+ and Dolby Vision.
Encoders and Packagers:
- FFmpeg: Effective for testing, but requires precise configuration. For production packaging, use Shaka Packager, Bento4, or commercial solutions like AWS Elemental.
Measurement Tools:
- Employ scopes and meters (e.g., waveform, parade, vectorscope) to measure luminance, color balance, and ensure no clipping occurs. Validate MaxCLL and MaxFALL values using specific QC software.
QA Checklist and Test Matrix:
- Verify metadata presence, bit depth, and color space tagging for a successful HDR experience. Test playback across multiple devices: high-brightness TVs, mid-range phones, and streaming sticks.
Best Practices and Checklist
- Work consistently at 10-bit (or higher) throughout the entire process.
- Maintain a uniform color pipeline, utilizing proper documentation for mastering display settings and values.
- Deliver both HDR and SDR, or provide engineered tone-mapping solutions.
- Utilize trusted packagers to ensure metadata integrity and proper distribution outputs.
Common Pitfalls and Troubleshooting
- Missing or Incorrect Metadata: Check for SEI messages and correct embedding in the container if images appear washed out or too dark.
- Banding Artifacts: Ensure the use of a continuous 10-bit pipeline and avoid excessive compression.
- Incorrect Color Space Tagging: Verify color space tags to prevent oversaturation or misrepresented colors.
- Inconsistent Device Playback: Offer multiple format versions to accommodate device differences, and utilize SDR masters when necessary.
Future Trends and Further Reading
- Watch for rising support of AV1 for HDR and the expanding use of dynamic metadata like HDR10+ and Dolby Vision.
- Real-time HDR production and cloud-based grading workflows are set to mature as technology advances.
Authoritative References:
Plan your next steps: Test your HDR workflow with DaVinci Resolve, and build a home lab for encoding and playback by following relevant guides.