Raw video is huge. A single uncompressed second of 4K footage at 60 frames per second can take up several gigabytes. Without compression, video would be almost impossible to stream, store, or share at any reasonable cost. Even a normal hard drive would fill up in minutes.
Compression solves this by finding the parts of the video that repeat and storing only what is actually needed.
Two ideas drive it.
Spatial redundancy. Within a single frame, neighboring pixels are often the same. A blue sky is mostly the same blue across thousands of pixels. Instead of storing each pixel on its own, the compression stores “blue sky here, this much area” and rebuilds the pixels on playback.
Temporal redundancy. Between one frame and the next, most pixels don’t change. A talking head against a still background only changes around the face and lips. So instead of storing every frame as a full image, compression stores one full frame, called a keyframe or I-frame, then stores only the changes from frame to frame.
Together, these tricks can shrink video by a factor of 100 or more compared to uncompressed video, with most viewers unable to see any loss.
The trade-offs look like this:
Every video file you have ever watched is compressed. The format (MP4, MOV, MKV) is the container, and the codec inside it (H.264, H.265, AV1) is what does the compression. The codec and the settings you choose affect everything about how the video looks, plays, and shares.
Compression comes in two basic kinds.
Lossy compression. The codec throws away some information to get smaller files. The discarded data is usually invisible to viewers, like subtle color shifts or very fine detail, but once it is gone, it is gone. After you encode with lossy compression, you can’t recover the original.
Examples: H.264, H.265, AV1, VP9. All the standard delivery codecs are lossy.
Lossless compression. The codec finds the repeating parts without throwing anything away. The compressed file can be decoded back to the exact original. It produces much smaller files than uncompressed video, but much larger files than lossy compression.
Examples: ProRes (technically “visually lossless”), Cineform, FFV1 (truly lossless), and various archival codecs.
For most everyday delivery, lossy is the right choice. The files are dramatically smaller, and the quality loss is mostly invisible. A viewer can’t tell an uncompressed 4K video from a well-encoded H.265 version, even though the files differ by 50 times or more in size.
Lossless matters in specific cases:
For online and commercial video, the workflow is usually: capture in lossy (the camera records H.265 or similar), transcode to visually lossless ProRes for editing, then deliver in lossy H.264 for distribution. Three different codecs, each right for its stage.
Push compression too hard and you start to see flaws that mark the video as low quality.
Blockiness. Square patches in the image, most visible in dark areas or smooth gradients. The compression splits the image into blocks and works on each one, so when it is too aggressive, you start to see the block edges.
Banding. Smooth gradients, like a sunset sky or a softly lit face, break into visible bands instead of smooth color. The compression has cut the number of color values, leaving stepped gradients.
Mosquito noise. Faint, fuzzy distortion around sharp edges. You see it around on-screen text and hard contrast boundaries.
Smearing. Fast motion looks blurry or pixelated. The compression couldn’t keep up with the quick movement.
Color shift. Very saturated colors go muted. Pure red looks a little less red, bright cyan goes dull. The compression has cut the color precision.
Macroblocking. Bigger artifacts where whole sections of the image look wrong, with the wrong colors or details. This usually happens at very low bit rates.
These flaws show up more in certain content:
A higher bit rate reduces these flaws. So does a better codec, since H.265 usually handles compression better than H.264 at the same bit rate. So does filming with compression in mind, like avoiding extreme contrast, very busy backgrounds, and overly saturated colors where you don’t need them.
For most professional video, these flaws are invisible at normal viewing distances when the bit rate and codec are right. They only show up when the bit rate is too low for the content, or the codec is older and less efficient.
Different stages of video work call for different compression choices.
Capture (in-camera). Most cameras record in lossy codecs to keep file sizes manageable on memory cards. Common ones are H.264 (older cameras), H.265 (most modern cameras), and ProRes (some professional cameras). Higher-end cameras let you pick from several codecs at different quality and size trade-offs.
Editing. Lossy delivery codecs are hard to edit smoothly because each frame depends on its neighbors. Most professional editors transcode to an editing codec like ProRes or DNxHR, where each frame stands on its own. Editing becomes far smoother.
Master files. The finished video should be saved as a high-quality master, ready for re-use. ProRes 422 HQ or ProRes 4444 are common choices. The master is much bigger than the delivery files, but it keeps the quality for future re-encoding.
Delivery. The compressed file you send to viewers. Different platforms want different things:
Archival. When you want a video to last for future re-use, store the highest-quality version you have. That is either the original camera files, if disk space allows, or a ProRes 4444 / ProRes RAW master.
The standard pattern: capture in lossy, transcode to an editing codec for performance, edit at that level, export to a lossy delivery codec, and archive a high-quality master for later.
Video compression is one of those invisible things that quietly shapes everything: file size, streaming speed, visible quality, editing performance, and how the video holds up over time. The right choice depends entirely on where the video is going. At Clipmasters, your editor matches the compression to the destination, so the codec that is right for editing, the one that is right for YouTube, and the one that is right for an archive all get used where they belong. It is handled as a normal part of delivery.
Compression is the process of reducing file size. The codec is the specific technology that does it. H.264 is a codec, and H.264 compression is the result of using that codec. People use the terms interchangeably, but technically the codec is the tool and compression is what it does.
Lossy compression always discards some information, but the loss is often invisible at normal viewing. Modern codecs like H.265 and AV1 compress so efficiently that videos at reasonable bit rates look nearly identical to uncompressed versions. Heavy compression at low bit rates does cause visible flaws like blockiness, banding, and smearing. The right amount of compression for the job looks fine to viewers.
Lossy compression discards some information to get smaller files. The discarded data is usually invisible but is permanently gone. Lossless compression finds the repeating parts without throwing anything away, so the file can be decoded back to the exact original. Lossless produces larger files than lossy but smaller than uncompressed.
Social platforms like TikTok, Instagram, and Facebook re-compress your video heavily. They use their own codecs and bit rates, often much lower than your original file. Even a perfectly encoded source looks worse after their re-compression. The fix is to upload at the platform's recommended specs and accept that the playback quality won't quite match your source.
H.264 in an MP4 container is the safe default. YouTube also accepts H.265, AV1, and other codecs. For 1080p, aim for 8 to 12 Mbps. For 4K, aim for 35 to 45 Mbps. YouTube re-encodes everything to its own formats anyway, so your upload codec and bit rate are just a starting point. The final playback is YouTube's compression layered on top of yours.
Skip to content
