Pixels with unequal width and height — common in DV and legacy video codecs. Compositor must correct pixel aspect ratio, otherwise image stretches distorted.
You're working with DV or older video codecs and suddenly notice: the image looks like it's been squashed or stretched sideways. The culprit is almost always the non-square pixel—a relic from the analog video era that still lingers in some formats today. While computer graphics are built on square pixels (1:1 pixel aspect ratio), many video standards use rectangular pixels with different widths and heights. PAL DV, for instance, uses 720x576 pixels with a pixel aspect ratio of 64:45, while NTSC DV uses 720x480 with a 10:11 ratio. This leads to distortion—faces appear unnaturally wide or narrow, and circles become ellipses.
You rarely notice this on set because the monitor already compensates for the aspect ratio. It becomes critical in editing and VFX: if you import DV footage into a square pixel timeline without correcting the pixel aspect ratio, you'll get a distorted image. Every professional NLE—Premiere, Final Cut, Avid—has settings for this: you simply define the correct source material format, and the software automatically scales the pixels to 1:1. In compositing software (After Effects, Nuke), the same is done through "Interpret Footage" or similar functions. Without this adjustment, you'll be stuck later.
The practical tip: Save the pixel aspect ratio values for your common formats. HDV needs 16:9, certain MiniDV variants need 4:3—documenting what's going into the project saves hours of debugging. When digitizing older material or working with archive footage, the first thing you check is the pixel aspect ratio. Modern formats (ProRes, DNxHD, H.264) work almost exclusively with square pixels, so the problem is diminishing. However, in documentary archives, post-production, or broadcast projects, you'll still regularly stumble over this invisible pitfall.