Why is my video file so large?
Four things decide how big a video is. Only some of them are worth changing, and the one most people reach for first is usually the wrong one.
Published
Size is bitrate times duration. That is the whole formula.
A video's size is its bitrate — how many bits are spent on each second — multiplied by its length. Everything else, including resolution, only matters because of how it pushes bitrate around. A ten-minute clip at 10 Mbps is roughly 750 MB, and no amount of tweaking other settings escapes that arithmetic.
The codec matters more than the resolution
A codec is the compression method. H.264 is the default almost everywhere. VP8 is the royalty-free option used in WebM. VP9 was released royalty-free by Google, but a third-party patent pool now claims it. Newer codecs are dramatically more efficient: the same visual quality that costs 10 Mbps in an old format might cost 4 Mbps in H.264.
This is why dropping a 4K video to 1080p often disappoints. You cut the pixel count by three quarters and the file barely halves, because the encoder was already spending fewer bits on the detail you removed. Re-encoding with a better codec, or simply a lower bitrate target, usually does more.
Why phone video is enormous
Phones record at high bitrates on purpose. The camera has no time to think — it is writing frames in real time, with no opportunity to analyse the whole clip — so it over-provisions to guarantee quality. A minute of 4K phone video can easily exceed 400 MB.
Re-encoding that same clip offline, where the encoder can take its time, routinely cuts it by half or more with no visible difference. That is the single biggest saving available, and it costs you nothing but processing time.
What CRF means, and why it beats picking a bitrate
Most encoders offer two ways to decide quality. You can name a bitrate, or you can name a quality target and let the encoder spend whatever each scene needs. The second is almost always better: a static shot of a wall needs very few bits, a confetti explosion needs many, and a fixed bitrate wastes the former and starves the latter.
For H.264 this target is CRF, roughly 18 to 28, where lower means better and larger. This site converts to MP4 at CRF 23, which is the usual balanced default — visually very close to the source for most footage.
The audio is not your problem
People often try to shrink a video by attacking the audio. A typical stereo track is 128 to 192 kbps, against video at several thousand. Removing audio entirely from a 500 MB video saves perhaps 10 MB and makes it useless. Leave it alone.
When GIF is the wrong answer
Converting video to GIF makes files dramatically larger, not smaller. GIF has essentially no compression between frames — every frame is stored almost independently — and it is limited to 256 colours. A three-second clip that is 400 KB as MP4 can be 4 MB as GIF. Use GIF when you need it to autoplay somewhere video is not allowed, never to save space.
To shrink a clip in practice, compress the video or reduce its resolution; if a short loop is all you need, turn it into a GIF and compare the sizes.