JPEG Compression
JPEG or Joint Photographic Experts Group defines a means of digitally compressing still pictures (or a series of still pictures in motion JPEG). This compression is performed by breaking up the picture to be compressed into blocks of 8 x 8 pixels. Each block is then mathematically operated upon by a process known as Discrete Cosine Transform or DCT. This transform is a fixed mathematical process, which can be implemented in hardware, although it is relatively complex. Rather than compress the data, a DCT transform actually increases it (typically 8 bits of input produce 11 bits of output), but much of the resulting data consists of strings of zeros. (This is especially true if the picture to be compressed has areas of all the same colour and density, whereas sharp edges will cause non zero DCT coefficients). These strings of zeros can be reduced using conventional run - length compression techniques (typically Huffmann coding) as is found in the ZIP type of software compression. Uncompressing at the receiving end can regenerate the original DCT data, which is then subjected to a reverse transform process recreating the original picture information. JPEG is near lossless, but is not completely so the DCT transform generates more data than the original picture, not less, and if the original image is very complex, original picture data will inevitably be lost. JPEG also has a contrast enhancing effect subtle variations in shades of colour are lost, as the DCT transform coefficients whilst not zero (no change at all) are very near zero, and will be treated as zero for the purposes of compression. The result is that the received picture will have less subtle colour and image transitions, giving a subjectively crisper and sharper image. It is possible to vary the amount of compression of a JPEG image by altering the Q (quality) factor.
Motion JPEG (the standard used by the VideoBOX series) simply treats each field or frame of a video sequence as a separate image, and performs a completely new and separate compression on each picture. The JPEG Q-Factor has a significant impact on the amount of data that is transmitted and also the quality of the resulting picture. Q-Factor 20 is widely perceived to offer a TV quality image. This will produce data rates of approximately 18 Mbps for a typical full size, full frame rate PAL or NTSC interlaced video signal. For the same typical PAL or NTSC signal a Q-Factor of 200 will require approximately 8 Mbps. It is possible to use far higher Q-Factors e.g. 1024 with a corresponding reduction in the data rate (in the case of Q-1024 to approximately 5 Mbps). However, it is unlikely that the quality of image produced would be condsidered suitable for human viewing but rather used as input to a digital image processing algorithm.
MPEG or Moving Picture Experts Group defines a means of digitally coding moving pictures and associated audio. MPEG achieves higher compression rates (for typical image sequences) than JPEG due to its use of inter-frame coding. The MPEG algorithm however is considerably more complex than Motion-JPEG and introduces extra latency at both the encode and decode phases. There are several MPEG standards:
- MPEG-1: Coding of Moving Pictures and Associated Audio for Digital Storage Media at up to about 1.5 MBit/s
- MPEG-2: Generic Coding of Moving Pictures and Associated Audio
- MPEG-3: No longer exists (merged into MPEG-2)
- MPEG-4: Very Low Bitrate Audio-Visual Coding
For more information on video compression the