size → bitrate
Work backwards from the size you are allowed
Encoders ask for a bitrate. Upload limits state a size. This turns one into the other, which is the calculation almost nobody wants to do by hand.
Processed on your own hardware · never transmitted, never stored
Bitrate is the number that decides everything
It is how many bits the encoder may spend on each second of footage, and it determines file size almost entirely — multiply it by duration and you have your answer to within a few percent.
The awkward part is that you almost always start from the other end. A platform gives you a size cap, an email server gives you a limit, a client asks for something under a certain figure. None of them tell you what to type into the encoder.
The audio field matters more than people expect on short clips. At 128 kbps audio consumes a meaningful share of a small budget, and for a speech-only recording dropping to 96 or even 64 frees real room for the picture.
A quick check on whether the answer is usable
If the figure comes back under roughly a thousand kilobits per second for 1080p material, the result will look visibly soft no matter which encoder you use. At that point you have two honest options: shorten the clip, or drop to 720p where the same budget stretches considerably further. Bitrate is always relative to how many pixels need describing.
Common questions
Why is my actual file slightly different?
Encoders aim at a target rather than hitting it precisely, and the container adds a little overhead. Expect to land within a few percent.
Constant or variable bitrate?
For hitting a size target, constant is more predictable. Variable produces better quality per byte but overshoots and undershoots along the way.
What should I upload to YouTube at?
YouTube re-encodes everything regardless, so send it something generous. Around 8000 kbps for 1080p is more than sufficient.
Is anything transmitted?
No. This is arithmetic running in the page.