Media File Size Calculator

Size = bitrate × duration ÷ 8. Real encoders use variable bitrate, so treat results as ±20%. Sizes use decimal units (1 GB = 1,000 MB) like drive makers and most upload limits; Windows displays binary GiB, which reads ~7% smaller for the same file.

The preset is the answer; the duration is just multiplication

Duration is the box people fuss over, and it is linear — an hour is exactly twice half an hour, at every preset, always. The dropdown is not. One hour of content, top to bottom:

Top to bottom is 312.5×: 4K30 H.264 against MP3 128, which is the ratio of their bitrates and nothing else. A 25 MB attachment limit holds 5 seconds of 4K, or 26 minutes of music. Same limit, same arithmetic. Every capacity in the table orders the presets identically, because only the bitrate is doing any work. Which means one wrong choice in the dropdown outweighs any plausible mistake about how long the recording is.

Resolution is a bigger lever than codec, by two and a half times

Starting from 4K30 H.264 and using this page's own presets, both choices have a price:

Switching codec halves it; dropping the resolution divides it by 5. The resolution change is worth 2.5× the codec change. And they multiply rather than adding, so doing both gives 10×. Re-encoding to a newer codec is the change people reach for first because it costs nothing visible. Dropping the resolution is the one they resist, and it is the one that does the work.

Whether you can upload faster than you record has nothing to do with length

Upload time divided by recording time is bitrate divided by upload speed — the duration cancels out completely. A ten-second clip and a ten-hour recording sit at the same ratio. So there is one clean threshold, and it does not move:

Break-even is where the upload speed equals the bitrate — 8 Mbps for 8 Mbps video. Below it, sending takes longer than filming did; above it, less. Shortening the clip does not help, because it shortens both sides equally. Worth checking the upload number rather than the download one: consumer connections are asymmetric, and the figure advertised is almost always the download.

Lossless compression saves less than most people expect

FLAC at around 1,000 kbps is 71% the size of the CD-quality WAV it came from. A lossless codec can only exploit redundancy that is genuinely there, and music has less of it than text does — 29% off is a good result, not a disappointing one. MP3 at 320 kbps is 23% of the same WAV, and most listeners cannot tell it apart. That is 2.7× the saving, bought with information you were unlikely to use — which is the entire case for lossy compression in one comparison.

The 1,411 kbps figure is not a convention either: a CD stores 44,100 samples a second, 16 bits each, across 2 channels, which is 1,411,200 bits a second exactly. Everything above is derived from that one number.