File Transfer Time Calculator

Two units, one very common mistake

Every file browser, encoder and NLE reports file size in bytes — MB, GB, TB. Every ISP, router and speed-test site reports connection speed in bits — Mbit/s, often shorthanded "100 meg" or "1 gig." The two differ by a factor of exactly 8, and confusing them is the single most common error in transfer-time planning. A "100 meg" line does not move 100 megabytes per second — it moves at most 100 megabits per second, which is 100 ÷ 8 = 12.5 MB/s peak. A "1 Gbit/s" (gigabit) line tops out at 1,000 ÷ 8 = 125 MB/s. Mix the two up and every estimate is off by 8×, which on a wrap-day offload is the difference between "back in an hour" and "still uploading at midnight." Every readout on this page shows both units side by side so that mistake can't happen quietly. The underlying arithmetic is one line: seconds = (file size in bytes × 8) ÷ (connection speed in bits per second). File sizes here use decimal units — 1 GB = 10⁹ bytes, 1 TB = 10¹² bytes — matching how operating systems, cameras and marketing report capacity, not the binary GiB/TiB some tools quietly substitute.

Why "realistic" almost never matches "advertised"

An ISP's headline number is the physical-layer ceiling under ideal lab conditions. Real sustained throughput on a real transfer is lower, for reasons that stack: TCP/TLS handshake and encryption overhead eats several percent before a single media byte moves; the disk at each end has to keep up (a spinning drive or a saturated NAS can bottleneck well below the network's ceiling); Wi-Fi, VPNs and shared office/ISP contention all shave more off; and the transfer tool itself — its chunk size, retry logic, checksum verification — adds its own pacing tax. None of that is a defect to "fix" — it's normal, and pretending it away is how delivery deadlines get missed. This calculator bakes a realistic figure into each named connection tier (visible in the dropdown) instead of the advertised one, and exposes the overhead as an explicit 0–40% slider (default 15%) for the Custom tier, so you can see exactly what's being subtracted and why.

Preset derivations

The size presets below are back-solved from real, sourced codec data rates, not guessed — the same nominal ProRes figures used across VoiceDeck's Memory Card Record Time Calculator:

The "one shoot day" figure isn't 24 continuous hours of ProRes — it's a realistic accumulated rolling total across a shoot day's setups and takes, which is why it lands near 4.5 hours of actual recording rather than a full day.

The worked example

Take a 1 TB shoot day of footage going up over gigabit fiber. The tier's advertised number is 1,000 Mbit/s, but the realistic sustained figure is 700 Mbit/s (see the tier table above). seconds = 10¹² × 8 ÷ (700 × 10⁶) ≈ 11,429 s ≈ 3h 10m. That's nearly five times longer than the "1 Gbit/s = instant" assumption a client might make from the plan's marketing name — and it's the number worth setting expectations against before you start the upload, not after it's still running at the end of the day.

The ship-a-drive crossover

Andrew Tanenbaum's line still holds: "never underestimate the bandwidth of a station wagon full of tapes hurtling down the highway" — updated here to an overnight-couriered SSD. A 2 TB drive that arrives in 24 hours moves data at an effective rate of 2 × 10¹² × 8 ÷ 86,400 ≈ 185 Mbit/s — comfortably beating every home-connection tier on this page, and even beating some office lines, once the payload gets large enough. This calculator computes that crossover live for whatever tier and size you've selected: it finds the file size at which your connection's realistic rate would take exactly 24 hours, and everything past that point is faster to physically ship. On gigabit fiber (700 Mbit/s realistic) that crossover lands around 7.6 TB; on a typical home upload (30 Mbit/s realistic) it's under 350 GB — which is why couriering drives is still routine practice for large RAW or DPX deliveries, not a relic.

When to split deliveries — and a better option

For anything past the crossover, or with a hard deadline, splitting one massive archive into per-file, per-shot delivery links lets a client start reviewing while later files are still moving, and gives you a retry unit smaller than "the whole project" if one file fails checksum. But the workflow that actually removes this math from your day-to-day isn't a faster upload — it's uploading a master once and letting every reviewer stream it, instead of re-exporting and re-sending a new file every time notes come back. That's the workflow VoiceDeck is built around.

1. Enter your file size, or load a production preset such as a 1 TB shoot day or a 166 GB ProRes master.

2. Pick the connection and dial it down to the sustained speed you actually see, not the advertised one.

3. Read the total, the 25/50/75% ETAs and the ship-a-drive crossover — or shrink the file first with Compress Video to Size.