Why Video Files Are So Big and How to Shrink Them

A single photo might be a few megabytes. A two-minute video at the same resolution can easily be a hundred times larger. The reason isn't that video is worse at compression. It's that video is a much bigger problem to begin with.

It's just a lot of images, very fast

A video at 30 frames per second is, underneath, thirty separate images shown every second. A two-minute clip is 3,600 frames. To see why compression is essential, do the math on raw footage. One uncompressed 1080p frame has 1,920 by 1,080 pixels, about 2 million of them. At 3 bytes per pixel for color, that is roughly 6 megabytes per frame. At 30 frames per second, that is close to 190 megabytes every second, and a two-minute clip would be over 20 gigabytes. Nobody stores or streams video like that, so every video file you've ever seen is already heavily compressed.

Modern video codecs are smarter than compressing every frame

Rather than compressing each frame on its own, video codecs like H.264 mostly store the differences between frames. Some frames, usually called keyframes, are stored in full. The frames between them describe what moved or changed. If a scene is mostly static, such as a talking head with a still background, only the parts that actually change (the mouth, a hand gesture) need new data. The unchanging background is referenced from earlier frames.

The codec also predicts motion. If a car moves across the screen, it stores a note saying roughly "this block moved this way" plus a small correction, instead of redrawing the car. This is why a static screen recording compresses far better than fast action footage, a handheld shot, or anything with film grain, rain or confetti. There is simply less that is predictable from frame to frame.

What bitrate and file size have to do with each other

File size is mostly bitrate multiplied by duration. Bitrate is how many bits are spent per second of video. As a rough example, a video encoded at 8 megabits per second takes about 1 megabyte per second, so two minutes is around 120 megabytes. Halve the bitrate and you roughly halve the file. Longer video means a bigger file, which is why trimming is such a reliable way to save space. The audio track adds some size too, though usually much less than the picture.

What you're actually trading when you compress

Video compression is lossy, which means some information is permanently thrown away. Many encoders expose a quality number called CRF (constant rate factor). Lower numbers mean less compression and bigger files. Higher numbers mean smaller files with more visible artifacts: blockiness in fast motion, softer detail, banding in gradients like skies, and smearing around moving edges. There is no universally correct setting. It depends on how much visible loss is acceptable for where the video is going.

Also remember that each re-encode adds loss. Compressing a video that was already compressed, then compressing the result again, can degrade it noticeably. Keep the original and always compress from that.

Practical order of operations

If a video needs to be smaller, the biggest wins usually come in this order:

1. Trim what you don't need. Cutting dead time at the start and end is free quality-wise. Our trim video tool exports just the segment you choose.
2. Lower the resolution. A video meant for a phone screen doesn't need 4K. Try the change resolution tool to scale down while keeping the aspect ratio.
3. Lower the frame rate if the content allows. Screen recordings and slideshows rarely need 60 frames per second. The change FPS tool does this.
4. Raise the CRF value. Increase it in small steps and check the result.

Changing everything at once makes it hard to tell which change mattered, so adjust one thing at a time if you need to hit a specific size.

Common mistakes

• Compressing a file that is already small and highly compressed, then wondering why it barely shrank.
• Chasing a tiny size at the cost of a blocky, unwatchable result.
• Upscaling a low-resolution video, which makes the file bigger without adding detail.
• Overwriting the original instead of keeping a master copy.
• Assuming the file extension decides the quality. MP4, MOV and others are containers. The codec and bitrate inside decide the size.

Our video compressor re-encodes with H.264 using a quality (CRF) slider, and runs entirely in your browser through ffmpeg, so nothing is uploaded to a server. Because it runs on your own device, large files can take a while and use a fair amount of memory. If the file is in a format other tools don't accept, the video converter can change it to MP4.

Frequently asked questions

Why is my phone video so large?

Phones record at high resolution and often at a high bitrate so they can encode quickly and keep quality up without much processing time. Footage from a phone is usually far less compressed than what you'd get from a streaming service, which spends a lot of computing effort to achieve small sizes.

Does a smaller video always look worse?

Not always. Many videos are encoded at a bitrate higher than their content needs, so there is often room to shrink the file with no visible difference. Past a certain point, though, the artifacts become obvious. View the result at the size it will be shown, not zoomed in.

What is the best format for small files?

Newer codecs such as H.265 (HEVC) and AV1 can produce smaller files than H.264 at similar quality, but support varies by device and software, and they can be slower to encode. H.264 in an MP4 container remains the most widely compatible choice, which is why it is a safe default.