Netflix Expands Camera File Processing

Netflix integrated the FilmLight API (FLAPI) into its Media Production Suite (MPS) as a core processing engine, standardizing the inspection, metadata processing, and conversion of camera original files. Once media is uploaded, its integrity is verified with ASC MHL, then camera metadata is extracted via FLAPI and mapped to Netflix's normalized schema for search and downstream tasks.

FLAPI is also used to generate deliverables for post-production and VFX. It processes original files with the correct debayering parameters, applies framing decisions via ASC FDL, and applies ACES Metadata Files (AMF) to maintain a repeatable color management pipeline. In the end, deliverables in various formats such as OpenEXR are automatically generated.

Netflix runs this processing not on its own high-performance equipment but as Linux Docker-based functions on top of its Cosmos cloud infrastructure. It's integrated with Java and Python, and designed to work even on CPU-only instances, allowing it to leverage a wide range of compute resources.

The key requirements are as follows.

• Must be able to run and terminate a single task like a serverless function
• Must support headless execution
• Must operate statelessly so it can easily restart in the event of a failure
• Must be able to elastically scale to handle uneven throughput driven by production schedules

Thanks to this architecture, Netflix can handle thousands of parallel renders and transcodes at low idle cost, providing a consistent workflow for both experienced on-set staff and teams with less color management experience.