H.264 (also called AVC, Advanced Video Coding) is the most widely deployed video codec in the world. It's used by YouTube, Netflix, Blu-ray, most phones, and virtually every streaming platform. If you've watched a video in the last decade, it was almost certainly H.264 or its successor H.265.
Understanding how it works isn't just for engineers. The settings you choose when compressing a video — quality level, bitrate, encoding speed — only make sense if you understand what the codec is actually doing.
The core idea: only store what changes
Raw video is every pixel of every frame stored in full. For a 1080p video at 30 fps, that's 2,073,600 pixels × 30 frames per second = about 62 million pixel values per second. Uncompressed, one minute of 1080p video is roughly 10 GB.
H.264 exploits the fact that most video frames don't change much between them. Instead of storing every pixel of every frame, it stores a small number of "keyframes" (full frames) and then only the differences between those keyframes and the frames in between. Static backgrounds, slow pans, and talking heads require very little data between keyframes. Fast action, scene cuts, and confetti require more.
What CRF means
CRF stands for Constant Rate Factor. It's a quality target, not a bitrate target. You tell the encoder "I want this level of visual quality" and it figures out how many bits to use in each part of the video to meet that target. Simple scenes get fewer bits; complex scenes get more.
CRF values range from 0 (lossless, largest file) to 51 (worst quality, smallest file). In practice, 17–28 is the useful range. CRF 18 is visually near-lossless for most content. CRF 28 is where compression becomes noticeable on high-detail footage under close inspection. CRF 35 is for rough drafts and previews.
Each CRF step of 6 approximately doubles or halves the file size. CRF 28 produces roughly half the file size of CRF 22, which produces roughly half the file size of CRF 16.
Why H.264 is still dominant despite H.265 existing
H.265 (HEVC) offers 40–50% better compression than H.264 at the same quality. However, H.264 has two advantages that keep it dominant: universal hardware decoding support (every device from 2010 onwards handles it efficiently) and no patent complexity. H.265 has historically had complicated licensing, though H.264 has its own licensing via MPEG LA. For most uses, H.264 at a sensible CRF is still the right choice.
Why browser-based H.264 encoding is slower than native
When you compress a video in a browser tool, the H.264 encoder runs as WebAssembly — a portable binary format that runs inside the browser sandbox. WebAssembly is fast, but it can't access hardware video encoders (the dedicated silicon chips in modern CPUs and GPUs that make H.264 encoding near-instant in native software). Expect browser-based encoding to be 5–15× slower than a native application like HandBrake. For most video files under 500 MB, this is still fast enough to be practical.
Try H.264 compression in your browser
Choose High, Medium, or Low quality. Up to 1 GB input. No upload, no install.
Open Video Compressor