On-Set Data Management: The Workflow That Protects Your Footage

Why Data Management Is the Most Critical Non-Creative Job on Set

Every other mistake on a production can be worked around. Bad lighting can be fixed in post. A missed line can be ADR'd. A performance can be replaced. But lost footage cannot be recovered. When a data failure happens — and they do happen, to productions of every size and budget — the only question is how many days, and how much money, you lose before you figure out what can be salvaged.

The good news is that data loss is almost entirely preventable with the right workflow and the right discipline. Productions that follow a rigorous data management protocol essentially never lose footage. Productions that improvise or cut corners because they're running behind schedule are the ones that end up in catastrophic situations. The workflow I'm describing here is not theoretical — it's what we run on every production at Rayvn Films, and it has never failed.

The Complete Offload Workflow

The core principle is simple: a card is never formatted and returned to camera until you have confirmed, verified copies on at least two separate physical media, and those copies have been checksum-verified against the original. Here is the step-by-step:

• Card arrives at the DIT cart. The loader or data wrangler hands off the card. It's logged into the camera report immediately — camera, reel name, time received.
• Ingest into Silverstack Lab. The card is mounted and ingested into Silverstack Lab, which reads the entire card and generates a checksum for every file. This checksum is the fingerprint of the original data.
• Copy to primary RAID. Silverstack copies the files to the primary RAID array — typically a multi-drive RAID 5 or RAID 6 — while simultaneously verifying that the copied files match the original checksums byte for byte.
• Copy to secondary backup. Simultaneously or immediately after, a second copy goes to a separate physical device — a second RAID, individual drives, or LTO tape depending on the production's setup. The secondary backup is verified the same way.
• Confirmation both copies pass verification. Only when both copies are confirmed verified does the card get cleared to be formatted and returned to camera. This is the only moment at which it's safe to reuse that card.
• Camera report updated. The log is updated with the time both copies were confirmed and the checksum status. This documentation is as important as the data itself.

Organization Structures That Actually Work

How you organize your media on the RAID matters as much as whether you copy it. A well-organized media structure means editorial can start working from dailies the same day, and anyone in post can find exactly what they need without asking the DIT. A disorganized structure means hours of cleanup in post and a much higher chance of editorial errors.

The structure we use: Project Name / Date / Camera Designation / Reel. Under each reel, the original camera files are preserved exactly as they came off the card — no renaming, no reorganization. A separate folder structure holds dailies and transcodes. Camera reports live in a Documents folder at the project level.

Software Tools

Silverstack Lab (Pomfort) is the industry standard for professional media offload and verification. It handles checksum generation and verification, side-car metadata, and camera report export. If you're doing professional production work, you should be using Silverstack. Alternatives include Hedge (simpler interface, good for solo data wranglers), and YoYotta (strong for LTO tape workflows). For productions without a dedicated data budget, Checksum+ or rsync with hash verification can work in a pinch, but they require more manual oversight.

RAID Levels Briefly Explained

Not all RAIDs are equal. RAID 0 stripes data across drives for speed but provides zero redundancy — a single drive failure loses everything. Never use RAID 0 for primary storage on a production. RAID 1 mirrors data across two drives, so one can fail without data loss, but you get half the storage capacity. RAID 5 distributes data and parity across three or more drives, allowing one drive failure without data loss, and is the most common choice for DIT primary arrays. RAID 6 extends this to allow two simultaneous drive failures, useful on larger arrays where the rebuild time after a single drive failure increases the risk window.

RAID is not a backup. A RAID is a redundancy mechanism against hardware failure within a single array. The second copy on a separate physical device is the backup. Both are required.

When to Format Cards

The answer is simple: never before confirmed backup on two separate physical media. Never when you're running behind schedule and the camera department is pressuring you for cards. Never when you think you probably copied it but didn't verify. The words "I'm pretty sure it copied" have preceded some of the most expensive conversations in production history.

If the camera department needs cards before you've completed and verified the backup, the right answer is to have more cards in the production's budget. Cards are cheap. Reshooting isn't.

What a Loader vs. a Data Wrangler Does

On large productions, the DIT often works with a camera loader who physically handles the cards — taking them off cameras, bringing them to the cart, returning them after formatting. The DIT or a dedicated data wrangler runs the software and manages the verification workflow. On smaller productions, the DIT does both. The important thing is that whoever is running the verification workflow understands exactly what they're doing and why, not just how to click the buttons.

For any production serious about protecting its footage, the data management workflow is not a place to improvise or save money. Contact Rayvn Films to discuss how we can support your production's data management needs, from single-day shoots to multi-week features.