Intraframe vs. interframe


Video compression formats use various techniques to reduce bit rate or overall size of a file. One effecting factor is the use of either interframe or intraframe compression. The two techniques differ in the way one frame relates to the next.

Put simply, interframe compression compiles each frame discreetly – imagine a single JPEG (or similar) image representing each frame – perhaps 15, 25 or even 30 per second.
 
A more bit-friendly technique is to employ intraframe compression, which compares the differences from one frame to the next, and saves bandwidth by only processing significant changes in particular pixels.

For instance, an intraframe compression technique applied to a talking head against a static background will redraw the region of video that contains the head, but will not redraw the static background.

In some formats, keyframes are employed, where, at specified intervals, the complete frame will be redrawn, and the following frames are compared to the previous keyframe – this technique is a combined inter- and intraframe technique, and is a good compromise between the low bandwidth of intraframe and the higher definition of interframe.

It is important to note that compression codecs that utilize intraframe compression, such as H.264, will commonly require higher levels of processing power to calculate the data which specifies differences in specific pixels from one frame to the next, rather than simply redrawing an entire frame each time.

  1. #1 by Sean Lancastle on December 12, 2011 - 11:46 pm

    Hi. Some of my students came across your site whilst doing some research and I think you have some of this information the wrong way around. ‘Intra’ means ‘within’ and refers to compression that happens within each frame. JPEGs are intraframe coded, for example. ‘Inter’ means ‘between’ and refers to the compression that happens due to the redundancy between successive frames. Interframe comression relies on previous (and sometimes future) frames to recreate the original pictures. All of the MPEG familily, including H264, rely on both methods. The first frame in a sequence is intra coded, folllwed by a number of inter coded frames for, typically, 16 frames.
    Great site, by the way, so don’t think I am being critical. I hope this helps.

(will not be published)