RAW-to-Graded Video Pairings Dataset for AI Training
Access a large-scale paired dataset of professionally produced cinema footage linking original camera RAW files with their corresponding final colour-graded master videos. The collection includes approximately 115,000 source-to-output pairings and around 500 hours of content for supervised AI training, colour grading, RAW processing, video enhancement, restoration, computational photography, and video generation.
The dataset combines RED R3D RAW and 12-bit Blackmagic RAW source files with professionally mastered outputs from commercial productions featuring authentic people, lifestyle, business, fitness, education, family, and urban environments. Production identifiers enable reliable matching between original camera files and their corresponding final videos.
Dataset Preview
Representative source-to-output examples showing original camera RAW footage alongside its corresponding professionally colour-graded final master.
Each preview clip presents an original camera RAW source and its corresponding final mastered version within a single video.
Key Highlights
- • Approximately 115,000 original RAW to final graded video pairings
- • Approximately 500 total hours of paired cinema-quality content
- • Approximately 65,000 RED R3D RAW → final graded video pairings
- • Approximately 50,000 12-bit Blackmagic RAW → final graded video pairings
- • Reliable source-to-output matching using retained production identifiers
- • Final professionally colour-graded master videos for each verified source pairing
- • Rights-cleared content originating from commercial productions
- • No spoken language
Metadata Fields
Example Metadata Record
Each pairing retains production identifiers for reliable RAW-to-master matching. Additional grading metadata and colour pipeline information may be available subject to archive verification.
pair_id: Available in metadata package
source_asset_id: Available from production identifiers
master_asset_id: Available from production identifiers
source_format: RED R3D RAW or Blackmagic RAW
source_filename: Available where applicable
master_filename: Available where applicable
duration: Available per pairing
resolution: Available per source and master asset
description: Can be built according to project requirements
keywords: Can be built according to project requirements
grading_metadata: Available subject to archive verification
colour_pipeline: Available subject to archive verification
model_release: Available where applicable
property_release: Available where applicable
Technical Specifications
Dataset Type
RAW to Final Graded Video Pairings
Content Type
Paired original cinema camera RAW files and corresponding final mastered videos
Total Pairings
Approximately 115,000
Total Duration
Approximately 500 hours
RED Pairings
Approximately 65,000
RED Pairing Duration
Approximately 200 hours
RED Average Clip Length
Approximately 17.5 seconds
RED Source Format
REDCODE RAW (.R3D)
Blackmagic Pairings
Approximately 50,000
Blackmagic Pairing Duration
Approximately 300 hours
Blackmagic Average Clip Length
Approximately 20 seconds
Blackmagic Source Format
Blackmagic RAW (.braw)
Blackmagic Bit Depth
12-bit
Output Type
Final professionally colour-graded master video
Content Source
Commercial video productions
Content Coverage
People, lifestyle, business, fitness, education, family, urban environments, and other real-world scenes
Pairing Method
Production identifiers retained for source-to-output matching
Additional Grading Data
May be available subject to archive verification
Spoken Language
None, non-verbal dataset
Licensing, Documentation, and Delivery
Wavebreak Media can provide applicable licensing, provenance, and model or property release information for the paired RAW and final graded video content, with dataset delivery organized around agreed technical and project requirements. See our Dataset Licensing & Compliance page for more detail.
Need the Full Dataset?
Request access to the RAW to edited video pairings dataset or discuss custom pairing selection, metadata, grading information, licensing, and delivery requirements with Wavebreak Media.
