How Much Storage Does a Film Production Actually Need? A Data-Driven Guide

The Storage Crisis That Happens at 3:00 PM on Day Seven

You're midway through a 12-day feature. The DIT has been backing up to the same two 8 TB drives that were full by the end of day four. Someone ordered two more drives, but they shipped to the wrong address. The director is pitching a new scene that wasn't in the breakdown, which means an extra hour of shooting today. The 1st AC is on the walkie asking whether to keep the camera rolling or swap cards. Nobody has a verified backup of days two through four.

This is not an exceptional situation. It happens on productions that never thought carefully about storage before cameras rolled. The mistake is always the same: estimating storage from a gut feeling instead of calculating from the actual codec data rate and shooting ratio.

This post works through the math for real configurations -- from a solo Blackmagic Pocket Cinema Camera 4K shooter to an ARRI ALEXA 35 feature -- so you can arrive on day one with the right amount of storage and a backup plan that won't fail mid-shoot.

Storage capacity figures in this post are based on published manufacturer codec specifications from Blackmagic Design, ARRI, and Sony, verified against independent codec benchmark data from Digital Cinema Society reports.

The Storage Formula Every DIT Uses

The core calculation for total raw acquisition storage is:

Total Storage (GB) = (Bitrate in Mbps / 8) x 3,600 x Shoot Hours Per Day x Shoot Days x Shooting Ratio / 1,000

Breaking that down: divide bitrate by 8 to convert megabits to megabytes, multiply by 3,600 to get per-hour storage, then scale by your daily shoot hours, number of days, and shooting ratio. The shooting ratio is the ratio of footage recorded to footage used in the final cut -- a 10:1 ratio means you record 10 minutes for every 1 minute in the finished film.

A worked example: BMPCC 4K shooting Blackmagic RAW at 3:1 compression (approximately 400 Mbps), 6 camera-hours per day, over 8 shoot days, at a 12:1 shooting ratio.

(400 / 8) x 3,600 x 6 x 8 x 12 / 1,000 = 50 MB/s x 3,600 x 6 x 8 x 12 / 1,000
= 103,680 GB = approximately 101 TB

That number is only camera originals. Before factoring backup copies, the production needs over 100 TB for acquisition alone. Most solo BMPCC shooters severely underestimate this. The Storage and Footage Calculator runs this calculation automatically once you enter codec, days, and ratio.

Storage Requirements by Camera and Codec

The most important column is the 10-day shoot at 8:1. These figures represent raw camera data only -- before any copies. The ARRI ALEXA Mini LF shooting ProRes 4444 generates more raw data per shoot day than any camera in this table, despite recording to a "compressed" codec, simply because of its higher per-frame bitrate.

Three Real-World Storage Plans

Example 1: Micro-Budget Short, Solo BMPCC 4K Operator

A director-DP shooting a 3-day narrative short solo on a Blackmagic Pocket Cinema Camera 4K (BMPCC 4K) in BRAW 12:1 (approximately 130 Mbps). Delivery target is 4K streaming. Shooting ratio estimated at 15:1 (generous for a one-person shoot with no dedicated continuity).

Calculation: (130 / 8) x 3,600 x 5 shoot-hours x 3 days x 15 / 1,000 = approximately 1.3 TB of camera originals.

Storage plan: Two 2 TB Samsung T7 SSDs. One acts as the primary card dump drive. The second is cloned from the first at the end of every shooting day using rsync checksums to verify the copy. Total cost: approximately $180 for both drives. No RAID needed at this scale -- but both drives travel in separate bags so a single theft doesn't wipe all footage.

Key decision: BRAW 12:1 instead of 3:1 keeps the daily data volume manageable for a one-person operation. The quality difference between 12:1 and 3:1 BRAW is visible only under very aggressive color correction at 100% zoom -- irrelevant for this project's delivery target.

Example 2: Low-Budget Documentary, Sony FX6 Dual-Operator Crew

A two-camera documentary with two Sony FX6 bodies, each recording XAVC-S 4K at 150 Mbps. A 15-day observational shoot with an estimated 40:1 shooting ratio (standard for observational documentary -- the camera runs when subjects are present, regardless of whether any given minute makes the cut).

Calculation: (150 / 8) x 3,600 x 7 camera-hours x 15 days x 40 / 1,000, multiplied by 2 cameras = approximately 75 TB of total camera originals.

Storage plan: On-set: two 4 TB Samsung T5 SSDs per camera (eight drives total) for card dumps. At base camp: a 72 TB Synology NAS in RAID 6 configuration (effective capacity approximately 48 TB) plus a second identical NAS as offsite backup shipped weekly to a secure facility. LTO-8 tape archive ordered after the shoot for long-term storage at approximately $35 per 12 TB tape.

Key decision: The production hired a part-time DIT at $250 per day to manage backup integrity and generate QC reports. At 75 TB of irreplaceable observational footage, this was insurance, not overhead.

Example 3: Mid-Budget Feature, ARRI ALEXA 35

A 20-day narrative feature on the ARRI ALEXA 35 in ARRIRAW (4.6K, approximately 310 Mbps). The production has a DIT, a data manager, and a post supervisor. Shooting ratio is 10:1, camera runs 8 hours per day.

Calculation: (310 / 8) x 3,600 x 8 x 20 x 10 / 1,000 = approximately 22 TB of camera originals, plus one additional camera on select days adds approximately 4 TB.

Storage plan: On-set: SSD offloads to OWC Thunderblade 8 TB arrays via Silverstack, with md5 checksum verification per card. Overnight: full copy to a second Thunderblade stored separately. End of week: LTO-8 tape archive of verified originals, managed by the post house. Edit storage: 40 TB Thunderbolt RAID 5 (Synology or G-SPEED Studio XL) at the post facility for the ProRes 422 HQ proxy edit set.

Key decision: The production budgeted $12,000 for storage -- drives, tapes, and DIT labor combined. That represented less than 0.5% of the production budget and was among the lowest-risk line items in the below-the-line.

How to Build Your Storage Plan: A Delivery Checklist

Step 1: Calculate acquisition storage. Enter your codec bitrate, daily shoot hours, shoot days, and shooting ratio into the Storage and Footage Calculator or run the formula manually. Round up to the nearest drive size -- never buy exactly the capacity you calculated.

Step 2: Multiply by three for the 3-2-1 backup rule. You need three copies of every file: one working copy (the edit drive), one on-set backup (a second drive that leaves set separately), and one offsite backup (a drive at a different physical location, or a cloud upload). Calculate total storage budget as 3x your raw acquisition figure.

Step 3: Decide between individual drives, RAID, and tape. Individual SSDs work for shoots under 5 TB. RAID arrays (RAID 5 or RAID 6 minimum) are appropriate for shoots over 10 TB where drive failure is a real risk. LTO tape is the cost-effective long-term archive solution above 20 TB -- current LTO-8 tapes cost approximately $35 each and hold 12 TB native. Do not use RAID 0 (striping without redundancy) for camera originals -- a single drive failure loses all data.

Step 4: Plan your card management system. Decide on a protocol for card labeling, dump order, verification, and re-use. A camera card should not be reformatted until two verified backup copies exist and the DIT or data manager has signed off. Build this into the daily call sheet as a production step, not an afterthought.

Step 5: Budget drive costs as a production line item. A 10 TB external SSD costs approximately $400-500. A 72 TB Synology NAS (drives included) runs approximately $3,000-4,000 when rented from a production house. LTO-8 drives rent for approximately $500-800 per month. Include these in your budget alongside cameras and grip.

Step 6: Assign a named person to storage oversight on every shoot day. If nobody owns this responsibility, it doesn't get done reliably. On micro-budget shoots, the director or DP takes this role. On professional productions, it's the DIT or data manager. The role includes: verifying checksums after every dump, logging card numbers to a manifest, and confirming offsite backup delivery.

Pro Tips and Common Mistakes

Pro Tip: Always calculate card-to-edit-drive storage separately. The question "how many cards do I need on set?" is different from "how much total storage does the project need?" Cards rotate -- you dump them, verify, reformat, and reuse. Your edit drive accumulates. A 20-day shoot might cycle through four CFexpress cards while accumulating 20 TB on the edit drive.

Pro Tip: For shoots using multiple camera formats (A-camera shooting RAW, B-camera shooting H.264), calculate storage for each camera separately and sum them. Don't average the bitrates -- the distribution of shooting time between cameras varies day to day, and averaging underestimates total storage on days when the RAW camera runs more.

Pro Tip: When using cloud backup, test your upload speed before the shoot begins. A 100 Mbps symmetric upload connection can push approximately 1 TB in 2.2 hours. A typical domestic upload of 20 Mbps pushes 1 TB in 11 hours. If you're shooting 4 TB per day and your upload is 20 Mbps, cloud backup is not a viable primary backup strategy unless you can run overnight uploads every night without interruption.

Common Mistake: Buying storage sized for one copy of the footage rather than three. The budget line says "8 TB hard drive -- $200." The actual cost of responsible storage on a shoot that generates 8 TB of camera originals is three drives (or equivalent) at $600, plus the labor to manage them. Productions that budget for one copy and then scramble mid-shoot for additional drives pay retail prices and express shipping fees that exceed what the proper budget would have cost.

Common Mistake: Relying on RAID as a substitute for backup. RAID protects against a single drive failing mid-project. It does not protect against: accidental deletion, ransomware, theft, flood, fire, or multiple simultaneous drive failures. RAID is uptime insurance, not data insurance. Offsite copies are data insurance. Both are necessary -- neither replaces the other.

Frequently Asked Questions

How many camera cards should I bring to a shoot?

Calculate your daily acquisition volume, then divide by card capacity to get the number of card cycles per day. Add at least two extra cards as contingency. For a 5 TB shooting day with 512 GB CFexpress cards, you need approximately 10 card cycles -- meaning you can get through the day with six cards if you dump and reformat three times, but you should have eight cards on set as a buffer against delayed dumps or verification failures.

Is cloud storage reliable enough to be one of the three copies?

Yes, if your upload speed is sufficient and you're using a production-grade cloud storage service (Backblaze B2, Wasabi, or Amazon S3, not consumer Dropbox). Cloud storage is particularly strong as the offsite copy because it's geographically separated from your on-set drives automatically. The risk with cloud is upload time -- verify that your daily upload volume is achievable within the available overnight window before committing to it as a primary backup strategy.

What RAID level should I use for the edit drive?

RAID 5 is the minimum for single-drive redundancy with decent write speed -- it can lose one drive without data loss. RAID 6 provides two-drive redundancy and is appropriate for edit arrays where drive failure during a busy finishing period would be catastrophic. Avoid RAID 0 entirely for any project storage. RAID 10 (mirrored stripes) offers the fastest performance with single-drive redundancy but requires twice the number of drives for the same usable capacity.

How long do hard drives last, and should I buy new drives for every project?

Consumer hard drives have an average lifespan of 3-5 years under normal use, per manufacturer MTBF (mean time between failure) data from Backblaze's annual drive reliability reports. For short-term production storage (weeks to months), used drives are acceptable if tested with a SMART health check before use. For long-term archive, buy new drives and rotate them out every 3 years. SSDs have different failure profiles -- they degrade with write cycles rather than age, but for camera data (write-once, read-many), SSDs are highly reliable over long periods.

When should I use LTO tape instead of hard drives for archive?

LTO tape becomes cost-competitive with hard drives at approximately 20 TB and above. Below 20 TB, hard drives are cheaper and more accessible. Above 20 TB, tape offers a lower per-TB cost (currently $2-3 per TB for LTO-8 native capacity) and is the most durable long-term storage medium -- LTO tapes stored correctly are rated for 30 years. The tradeoff is that tape requires a dedicated LTO drive to read, costing $2,000-4,000 for a standalone unit. Most productions at this scale rent tape archive service from a post house.

Related Tools

The Storage and Footage Calculator is the fastest way to turn your codec and shooting parameters into a total storage requirement. For understanding how codec choice affects data rate in the first place, the Codec Storage Calculator lets you compare codecs side by side on storage per hour. The post on video codecs explained covers the technical reasons behind each codec's bitrate and helps you choose a format that balances quality against manageable data volume. If storage planning intersects with your shooting ratio decisions, shooting ratio: what it means and why it matters shows how ratio targets translate directly into post-production workload and storage costs.

Conclusion

Storage math is not optional. Every TB of camera original that goes unplanned adds real cost -- either in emergency drive purchases or, worse, in lost footage that can't be re-shot. The formula is straightforward: bitrate drives volume, shooting ratio multiplies it, and the 3-2-1 backup rule triples it into your total storage budget.

This guide covers standard single-camera and dual-camera configurations. Multi-camera live event productions and VFX-heavy shoots with on-set plate photography generate fundamentally different data volumes and require separate planning frameworks.

What's the largest amount of data you've managed on a single production, and what was the one thing you wish you'd planned for before the shoot started?