Video Resolution
Video resolution describes the width and height of a video picture in samples.
| Familiar description | Common picture dimensions | Main-grid positions per frame |
|---|---|---|
| 720p HD | 1280 × 720 | 921,600 |
| 1080p Full HD | 1920 × 1080 | 2,073,600 |
| 2160p Ultra HD | 3840 × 2160 | 8,294,400 |
| DCI 4K full container | 4096 × 2160 | 8,847,360 |
Resolution is the width and height of the picture's sample grid. The table counts that main grid, rather than all component samples or a file's compressed bytes. The “p” in labels such as 1080p describes progressive timing; it is additional information about how the picture is formed.
“4K” is a family of descriptions. Television UHD commonly uses 3,840 columns, while the full DCI cinema container uses 4,096. Cropped cinema images can have other heights. The explicit width and height are the useful values when calculating or comparing files.
The UHD calculation
Doubling the width doubles the number of positions in each row. Doubling the height doubles the number of rows. Together:
3840 × 2160 = 4 × (1920 × 1080).
This is why UHD contains four times as many main-grid positions as Full HD, rather than twice as many. At 30 frames/s, a 1920 × 1080 stream represents 62,208,000 such positions per second; a 3840 × 2160 stream represents 248,832,000.
That count helps explain processing load. Scaling, filters, and picture reconstruction operate on the samples. Frame rate adds another dimension to the work: UHD at 60 fps represents twice as many picture positions per second as UHD at 30 fps.
File size has a different calculation. An encoder can represent similar regions and neighboring pictures efficiently, so four times the positions need not mean four times the compressed data.
Which rectangle is being described?
A media file can contain a 1920 × 1080 frame with a widescreen film occupying only 1920 × 800 positions. The remaining 280 rows are black borders. Its stored resolution is still 1920 × 1080, while the active image contains 1,536,000 positions.
A cropped copy can store only that 1920 × 800 rectangle. These two copies can look the same when fitted to a 16:9 television, yet use different stored grids. Bars supplied by the player do not count as stored picture rows.
The presented rectangle can differ again. A phone clip stored as 1920 × 1080 may display as 1080 × 1920 after its rotation is applied. An anamorphic 1440 × 1080 grid can display as 16:9 because its samples are intended to appear wider than square.
Samples and detail
Two files with the same grid can resolve very different detail. A camera's optics, focus, exposure, processing, and compression determine what information reaches those positions. Scaling a 720p picture to a UHD grid calculates additional output samples from the existing ones; it changes the grid available for display, rather than adding new camera observations.
When comparing library versions, first identify the stored and active rectangles. Then use frame rate and bitrate to understand processing and delivery needs. That keeps a resolution comparison attached to the pictures actually represented.