Why video files are so large
A minute of 4K footage from a modern phone is around 350MB. Gmail stops at 25MB. That gap is why almost nobody can email a video straight from their camera roll, and why every messaging app quietly ruins your footage instead.
Most of that size is unnecessary for how the video will actually be watched. Recording at 4K makes sense for editing; sending 4K to someone who will watch it on a phone screen does not.
Setting a target size
Target mode divides your chosen size by the length of the video to work out a bitrate, then holds the encoder to it. The result normally lands within a few percent of what you asked for, which is what you want when a service refuses anything over 25MB.
The catch is arithmetic: a ten-minute video squeezed into 8MB leaves roughly 100 kilobits per second for the picture, and no encoder makes that look good. The estimate under the size box tells you the bitrate you are asking for before you start, so you can raise the limit or trim the clip instead of waiting twenty minutes for a disappointing result.
Trimming beats compressing
If a video is too large to process at all, cutting it down is more reliable than pushing the quality lower. Half the length means half the data to hold in memory, and it is usually the thing that turns a failed job into a successful one.
Start and end times accept plain clock format, so 1:30 means one minute thirty seconds. The section you choose is extracted before any encoding happens, so trimming also makes the whole job proportionally faster.
What to change first
Resolution gives you the biggest saving for the least visible loss. Dropping 4K to 720p cuts the pixel count by about 89 percent, and on a phone screen the difference is close to invisible.
After that, adjust quality. The scale here is inverted, which trips people up: a lower number means higher quality and a larger file. 23 looks essentially identical to the source, 28 is a sensible default, and 32 and above starts to show blockiness in fast motion.
Realistic expectations
- Phone footage at 4K or 1080p: typically 70 to 90 percent smaller at 720p.
- Screen recordings: compress extremely well, often over 90 percent.
- Already-compressed downloads: modest savings, sometimes none. A file someone else already compressed has little left to give.
Related things people need
If you only need the sound from a clip, the audio compressor produces a far smaller file than any video ever will.
And when a video simply will not fit however hard you compress it, splitting it across several ZIP parts is sometimes the more practical route.
Why video is so much harder than everything else
A photograph is one image. A video is thirty of them every second, and an hour of footage is over a hundred thousand frames. Video compression works by storing only what changes between frames — a locked-off shot of someone talking stores the background once and updates the face — but even so, the volume of data is on another scale entirely.
That is also why compressing video is slow. The encoder is analysing motion between frames, predicting what comes next, and encoding only the difference. On a server with dedicated hardware that happens quickly. In a browser tab it does not.
Resolution before quality, always
4K video is 3840 by 2160, which is 8.3 million pixels per frame. 720p is 1280 by 720, or 0.9 million. Dropping from one to the other removes 89 percent of the pixels before any compression happens at all.
The reason this is nearly free is that most video is watched on a phone, where a screen roughly 6 inches across cannot display anything approaching 4K detail. You are storing information the viewer's eyes will never receive.
Quality settings matter too, but they operate on what is left. Reducing resolution first means the encoder has far less to describe, so it can spend its budget well.
What the quality number means
The scale here is inverted, which trips people up: lower numbers mean better quality and larger files. It is a constant quality target rather than a percentage, so the encoder spends whatever bitrate each scene needs to hit that standard.
Around 20 is visually indistinguishable from the source. 23 is a common default for good quality. 28 is a sensible compromise and the default here. Above 32, fast motion starts to show blocking, and above 36 it is obvious.
Why this is slower than other sites
Other services upload your video to their servers and encode it there, using hardware built for the job. That is genuinely faster. It also means handing your footage to a company, and it is why those services cap file sizes, add watermarks and require accounts — every video costs them real money to process.
Here the encoder runs on your own machine. Nothing is uploaded, there is no size cap and no watermark, and the cost is time. For a short clip that trade is easy. For a long video on a phone it may not be, and being honest about that seems better than pretending otherwise.
Questions
Why is it so much slower than other sites?
Because those sites upload your video to their servers and compress it there. That is faster, but it means handing your footage to a company, and it is why they cap file sizes and add watermarks — every video costs them money. Here the work happens on your machine, which is slower but private and unlimited.
Can I use my phone?
For short clips, yes. For anything long, a phone may run out of memory and the browser will close the tab. A laptop is far more reliable.
Is my video uploaded?
No. The video never leaves your device. Only the compression engine is downloaded, once, from a public code repository.