someone tells me that the size(both width and height) of picture/preview should be divided by 16 with no remainder. In daily scenario, we also always see this kind of resolution, e.g. 1920x1088, 2688x1520. Could anybody show me who makes this rule/limit? hardware? all the hardwares following this or just some of them? Thanks very much!
Asked
Active
Viewed 166 times
1 Answers
1
Image sizes that are multiple of 16 pixels are preferred because:
1. Most of the pictures taken are saved on the phone/SD card with JPEG compression. Also, the compression for videos is based on JPEG (MPEG-4).
The JPEG compression algorithm splits the image into blocks of 8x8 pixels, so 8 pixels is actually the lowest limit, also preferred. If the image resolution isn't a multiple of 8 pixels (on both width and height) the encoding algorithm pads the image until this is achieved. This is expensive memory-wise and can complicate the encoding process for certain images/videos.
Sometimes you need to have a 16 pixel multiple due to sub-sampling (considering one of every two pixels) for video compression. [JPEG codec example]
2. Alterations(rotations/flips) to a JPEG image can be done without loss of quality as long as the image size is multiple of 1 MCU (minimum coded unit). [JPEG lossless editing]
ionela.voinescu
- 384
- 1
- 7