Introduction and Significance 

Even if the computer can process the data, such large files are impractical for storage and transfer. 

This is especially important for files that will be transferred over the internet. 

Compression creates a new file that stores data in a format that requires less space. In some cases the compressed file can later be decompressed or expanded. 

How Compression Works 

Compression is achieved through the use of a codec: a compression – decompression algorithm that looks for redundancy in data files. 

For example, XXXYYYY could be reduced to 3X4Y. In this example, the compression is considered “lossless” because the file can be decompressed and restored to the original format without any loss of data. 

Video compression, however, is considered “lossy” because it results in a loss of data. 

When compressing video, codecs look for redundancy in areas where the human eye or ear cannot distinguish between differences.  Since the human eye is less sensitive to color differences than to brightness, the color information (chrominance) is separated from the brightness information (luminance). 

The codec then averages the chrominance data for adjacent pixels, which reduces the volume of data. The luminance data is not changed. 

Normally, individuals cannot perceive the differences caused by the lossy compression.  However, when the image is enlarged, the data loss is obvious, and it will look blocky. 

The algorithm of a codec achieves compression in other ways, such as by recording only the differences between frames. Since most of the information from frame to frame is the same, less data needs to recorded. 

Compression is also achieved by assigning short codes to frequently used characters while less frequently used characters are converted to longer codes. The amount of compression that is achieved is known as the compression ratio. Compression ratios will vary according to the type of images in the video and how much the images change. A busy image with many colors will not compress as much as one with less colors.  

Compression Formats for Video 

MPEG is the acronym for Moving Picture Expert Group, and it is the standard for digital video. The different types of MPEG are used for different data rates and applications. MPEG 1 was developed first to encode video at data rates of less than 1.4 Mbps. This is the data transfer rate of a single-speed CD drive and allows for video CD format. MPEG 2 supports better quality video. It has a data rate of 10 Mbps and is the compression format for DVD, direct broadcast satellite, and digital cable. MPEG 4 has been developed for video transfer over the internet and thus has the lowest data rate at 64 kbits/s. Other types of MPEG for use in multimedia are being developed such as MPEG 7 and MPEG 21

H.264 is the new industry standard for video compression in security DVRs. Prior generations used MPEG-4 and even MJPEG for video recording. H.264 has the advantage of offering the highest compression ratio, while maintaining excellent video quality for security applications. H.264's higher compression ratio (up to two times better than prior-generation technology) effectively increases storage capacity by 100%, producing smaller file sizes and, therefore, longer recording time on a fixed-capacity storage device. In addition, the use of H.264 allows high-quality images to be transmitted over networks at very low bit rates. Security systems that involve multiple cameras can quickly exceed the available network bandwidth without efficient compression. DVR system designs without H.264 often rely on reduced-frame-rate or lower resolution recording techniques that degrade picture quality in order to increase recording time and reduce video bit rates. Older codec formats (MPEG-4 and MJPEG) are often still required for legacy support, but the industry's trend to adopt H.264 as the primary codec in DVR is well under way. 

https://www.maximintegrated.com/en/app-notes/index.mvp/id/4683

Factors to Consider for Video File Size 

1. Frame Rate: 

Raw video runs at 30 frames per second. However, the illusion of motion can still be generated with speeds as slow as 10 frames per second when there is no fast moving items. Cutting the speed to 15 frames per second or less can cut the size of a file in half (or less than half) without sacrificing quality when there is only a moderate amount of motion. 

2. Image Size: 

Full screen resolution is typically 640 x 480 pixels. Like frame rate, reducing the image size can significantly reduce file size. When reducing image size, a 4:3 aspect ratio should be retained. It may be possible to play back a 320 x 240 image sized video at double-size to become a full-screen movie with reasonably good results. A small video size would typically run at 192 x 144 pixels. 

3. Color Depth: 

Normal digital video contains 24-bit color (millions of colors). Pulling down the color range to 16 bit (thousands of colors) will reduce file size by one third. Some codecs allow 8-bit color (256 colors) which might work only for animations. Tests should be done on the video file to determine which color depth is necessary because reducing color depth can greatly distort the image.

http://www.edb.utexas.edu/minliu/multimedia/PDFfolder/CompressingDigitalVideo.pdf