A frame is a single still image within the sequence of images that comprise the video. Note that, in an interlaced scanning video system, a frame comprises two fields. Each field contains half of the video scan lines that make up the picture, the first field typically containing the odd numbered scan lines and the second field typically containing the even numbered scan lines.

To compress video signals, the MPEG system categorizes video images (frames) into different formats. These formats vary from frame types that only use spatial compression (independently compressed) to frames that use both spatial compression and temporal compression (predicted frames).

MPEG system frame types include independent reference intra frames (I-frames), predicted frames that are based on previous reference frames (P-frames), bi-directionally predicted using preceding and frames that follow (B-Frames) and DC frames (basic block reference levels).