Difference Between Inter and Intra Frame Compression

In the recent era of digital technology, people connected to the network need to send data within a minimum time and aim to save storage space for media files. When large files have to be transmitted simultaneously in a shorter period, then the user can choose any compression method to reduce the data size and upload it. Data compression is a complex task as each frame of the image or video has to be technically processed and it should not compromise the quality of the original data before sending it through the network. Two types of compression for video formats are Inter and Intra frame compression, features and working of them are discussed with its difference.

Inter Frame Compression

It is a technique used across video frames for compression where minor changes are made from frame to frame. Less data is stored as frames are checked using reference frames and only the difference is noted or stored. In the encoding process, one input frame is analyzed and keyframes are chosen to compress and this keyframe is used as a reference for the remaining frames. At decompression, we need to add changes that occur in the frame by comparing the keyframe. These keyframes are updated, whenever there is a major change in video (new scene not related to existing one).

Consider a real−time example: An anchor reading live news on a channel, here movement occurs only with the anchor’s body, and no changes are noted in the background frames. So, the difference to be stored is only the movements of the person and not the background for every frame. The difference from the first frame is taken as the keyframe and used as a reference for the remaining frames.

Frames are divided into three types that are used commonly by video compression algorithms:

• I−frame: It is termed as intra−code frames which are actual keyframes used as reference ones during the decoding process. It consists of all information for decoding. It provides an outline for each scene thereby it reduces the number of bits to encode a frame.

• P−frames: Predictive frames use I−frames as reference frames to predict the difference between the references and the contents of the remaining frames.

• B−frames: It predicts the information available in I−frame and p−frame for reference, called the bi−directional frame.

• S−frames: This switching frame is a predictive frame that uses I−frame as a reference and the rest of the information is obtained from the video stream. It has been invented as a part of AV1 code and allows an Adaptive Bite Rate that streams smaller portions of frames without reducing the quality of the video.

Intra Frame Compression

The compression process is made within the frames. The single image frame is divided into blocks and differences are noted between each pixel and these are encoded. Lossless and lossy compression are made to the information relative to the single or current frame and differences or predication derived is not related or compared with other frames in the input video sequence.

Each picture in the video sequence is encoded separately as an individual frame. Consider we are compressing a video using Motion Joint Photographic Experts Group that compresses frames as individual images. Users can recreate each frame without deriving the information from previous and upcoming frames. It is useful when editing a video, where every frame is encoded within the sequence to reduce the size and lower the bit rates, while not compromising on the quality of the original video.

Basis of difference	Intra Frame compression	Inter Frame Compression.
Definition	The compression process is made across the available frames.	Compression is made for each frame of the video sequence.
Functions	It identifies keyframes on one frame and uses them as a reference for the remaining frames.	It does not rely on previous information for the processing of the current frame.
Redundancy	Temporal redundancy is broken among the successive frames.	Spatially redundancy within the frames.
Processing time	It requires high processing time as the algorithm needs to process and compress several frames for decoding.	Less time than Intra frame, as encoding happens within individual frames.
Dependency	It depends on the keyframes of i−frame or p−frame to know the difference and make computations.	It does not depend on keyframes to know the differences within the pixels.
Types	I − frame, B− frame, P− frame, and S−frame	P−frame and B−frame.
Few popular Codec	MPEG−2,MPEG−4, H.265, H.264 AVC	DVCProHD , Apple ProRes, and Avid DNxHD.

Conclusion

Video compression methods aim to provide reduced size, residing/retaining the same quality as the original input video for faster transmission and saving storage space. Various codec methods are used with lossless and lossy, inter−frame and intra−frame compression. Video editing professional can use inter or intra−frame based on the software he/she uses and how the frames are captured in sequence in the same or different background environments.