Video Compression Explained: Resolution, Bitrate, Codec and File Size

Video file size is not determined by duration alone. A minute of slides, a minute of handheld night footage, and a minute of fast sport can demand very different amounts of data. Resolution, bitrate, codec, frame rate, audio, and the visual complexity of the source all influence the result.

The useful mental model

For constant-bitrate media, approximate size is bitrate multiplied by duration. Real encoders often vary bitrate over time, spending more data on difficult scenes and less on simple ones, but the relationship remains useful: longer duration or higher average bitrate generally means a larger file. Resolution and frame rate affect how much visual information the encoder must represent; the codec determines how efficiently it can represent that information.

Resolution: how many pixels each frame contains

Resolution is the frame’s width and height, such as 1920 × 1080 or 1280 × 720. Moving from 1080p to 720p greatly reduces the number of pixels an encoder must describe, but it also reduces the available spatial detail. That trade can be sensible for a small embedded player and harmful for a software tutorial whose interface text must remain readable.

Use Resize Video when output dimensions themselves need to change. The separate Compress Video page exposes compression level and CRF rather than an output-resolution control.

Bitrate and quality targets

Bitrate measures data per unit of time. At the same codec, resolution, and content, a higher bitrate usually gives the encoder more room to preserve detail and creates a larger file. A quality-based control such as CRF takes the reverse approach: you choose a quality target and let bitrate vary with scene difficulty. With the CRF convention used by common H.264 tools, lower values preserve more detail and usually create larger files; higher values are more aggressive.

There is no single correct value. Animation with flat colors may stay clean at a setting that damages noisy camera footage. Evaluate faces, text, gradients, shadows, and fast motion rather than judging only the opening frame.

Codec and container are not the same thing

A codec defines how video or audio is encoded and decoded. H.264, H.265/HEVC, AV1, VP9, and AAC are codecs. A container such as MP4, MOV, MKV, or WebM packages one or more streams with timing and metadata. Renaming an extension does not transcode the streams, and two MP4 files can use different codecs internally.

Frame rate matters when motion matters

Frame rate describes how many frames appear each second. Higher frame rates can preserve smooth motion but give the encoder more frames to represent. Reducing frame rate may save data, yet it can make pans, gameplay, demonstrations, or sport look uneven. Avoid changing it merely because the number is available; preserve the source timing unless the destination calls for a different rate and you have checked motion quality.

Why equal-duration videos differ in size

Screen recording

A mostly static slide deck has large unchanged regions that inter-frame compression can represent efficiently. But tiny interface text and sharp cursor edges expose ringing or blur quickly. Reduce frame rate only if motion is unimportant, and protect legibility when adjusting quality or resolution.

Camera footage

Handheld movement, water, leaves, low-light sensor noise, film grain, and frequent scene changes are harder to predict from nearby frames. The encoder may need a higher bitrate to maintain comparable quality. Stabilizing or denoising can affect compressibility, but those are editing decisions with their own visual consequences.

Animation and graphics

Flat colors can compress efficiently, while sharp lines and text can reveal artifacts. A codec and pixel format may also change edge appearance or color precision. Always check the actual output rather than assuming animation is automatically small.

Audio can be small—or significant

Total size includes video, audio, and container overhead. In a high-bitrate 4K clip, audio may be a small share. In a low-resolution talking-head clip or nearly static presentation, a high audio bitrate can represent a more noticeable fraction. Speech often needs less data than complex music, but excessively aggressive audio compression can make consonants, ambience, and music sound brittle.

If the picture is unnecessary, Extract Audio can produce an audio-only workflow instead of spending bytes on unchanged video frames.

Practical compression decisions

A repeatable checking process

1. Keep the source and note its duration, dimensions, frame rate, codecs, and size.
2. Remove footage that the recipient does not need.
3. Choose a compatible container and codecs for the destination.
4. Test a representative short section containing motion, text, faces, dark areas, and audio.
5. Adjust quality or bitrate, then resolution only if appropriate.
6. Play the complete output, confirm duration and audio synchronization, and measure its actual size.

Decision summary

Reduce duration when footage is unnecessary, resolution when the destination needs fewer pixels, and bitrate or quality when the same frame dimensions can tolerate less detail. Choose codec and container for the complete playback chain, not just the filename. Preserve enough frame rate for motion and enough audio quality for the material. Because sources vary, practical compression is an inspect-and-measure process—not a guaranteed percentage.