How video encoding works
Uncompressed video is enormous: a single hour of 1080p runs to hundreds of gigabytes. An encoder shrinks it by a factor of a hundred or more using two ideas. Within a frame, it simplifies areas of similar color. Between frames, it stores only what changed, since most of a picture is the same from one frame to the next.
Periodically it stores a complete frame, called a keyframe, so that playback can start or switch quality at that point.
The settings that matter
- Codec: H.264 for universal compatibility, H.265 or AV1 for smaller files on newer devices
- Resolution and frame rate
- Bitrate, or a quality target that lets bitrate vary
- Keyframe interval, usually two seconds for streaming so segments align
- Preset: slower encoding gives better quality for the same size
- Audio codec and bitrate, normally AAC
Encoding vs transcoding
Strictly, encoding turns uncompressed video into a compressed format, and transcoding converts one compressed format into another. Since the files uploaded to a streaming platform are already compressed, what the platform does is transcoding. In everyday use the two words are interchangeable.
Encoding for streaming
A streaming service does not encode a title once but several times, into a ladder of resolutions and bitrates for adaptive playback. Upload the highest quality source you have: each rendition is made from it, and quality lost in the source cannot be recovered.
Last reviewed