---------------------------------------------------------------------------- The Florida SunFlash Multimedia: Video (5 of 6) SunFLASH Vol 40 #29 April 1992 ---------------------------------------------------------------------------- 5 Video Desktop, full-motion digital video brings a new level of collaboration and communication to corporations. Just as visual communication, through pictures, can provide vastly more information than words alone, moving pictures can offer more and in some cases better, information than still images. Video Applications Full-motion digital video means a continuous sequence of video images integrated in the workstation. The ability to capture, archive, edit, display, and transmit digital, full-motion video provides both new and better internal corporate communication, and innovative, attention-getting, external communication with customers. Collaboration The shared workstation video-window will redefine workgroup computing. The networked video-capable workstation will provide a new level of collaboration among working groups. The current workgroup collaborates by sharing email and documents. The workgroup of the 1990's will also share images, video and audio. Multimedia Mail Just as email can include images and sound, it can also include a sequence of images. A mail application could enable users to place, and possibly edit, digitized video in a standard email message. The video might have been captured from a television broadcast, an in-house talk or trade show demonstration, or synthetic images generated on the workstation. Video Conferencing Video conferencing using special equipment not associated with the workstation has become both commonplace and a necessity for corporate communication. Current implementations are expensive (both for equipment and dedicated high-speed phone links) and typically available in very few locations within a corporation. Video conferencing on the workstation will fundamentally change the role it plays in corporate communications. Individual workstations, with video capture capability and inexpensive cameras, will capture, compress and transmit video (and audio) over local networks or ISDN lines to one or many workstations. You will be able to start up this application in a window as easily as you send email. You will also be able to share data from your workstation with others - for instance, engineering drawings or presentations and spreadsheets. A variation of videoconferencing is broadcast video. One person gives a presentation and broadcasts it to many people, while the audience can interactively ask questions through video conferencing capabilities. Software-only Video Playback Video software playback displays stored video without any additional hardware. Generally, the video is compressed to ease the storage requirements. This type of application is very useful for training and presentations or whenever someone only needs to play back video, rather than capture, edit, or store it. Video Editing Video editing enables you to capture video clips, then manipulate and store them. Different aspects include dropping and duplicating frames, creating special effects, synchronizing audio with the video, splicing clips together, and so on. This capability is very important for high-quality video production, but also for creating presentations, demos, training applications, and so on. Education and Training Full-motion digital video can transform computer-based training from a slide presentation into a live presentation. Opportunities abound for sales training, just-in-time technical training, occupational training, on-line documentation, and computer-based education courses. There are two categories of customers for these applications, each with different requirements. o Users of video authoring software, such as developers of training materials. need the software plus the video capture and compression capability. o Users of training materials may only need the developed materials, with video decompression and display capability. Interactive Presentations and Information Access Multimedia presentations are very similar to the applications described under education and training. Specific opportunities include interactive kiosks in public locations (such as instructions in a train station showing how to purchase and use tickets) and interactive commercial presentations (such as video presentations of a cruise ship in a travel agency where the customer can pick an activity or port-of-call and see a video of the selection). Once again there will be users requiring authoring capability (including video capture) and others requiring only display capability. Professional Video Productions Video productions, for broadcast video (such as television commercials and flying logos) or for distribution on CD-ROM (video games and sales material), can be planned, created, and edited using workstation video capability. Corporate Custom Applications As full-motion video on the workstation becomes more accessible, its use in corporate applications will become more commonplace. An initial application includes remote medical imaging for diagnosis and surgical planning. Key Video Concepts Full-motion digital video is actually a combination of several, related, hardware and software technologies for video capture, video compression and decompression, video transmission over a network, video display in a workstation window, and video archiving on disk or other storage media. Video Capture (input) The video we are all used to seeing on our televisions and VCRs is an analog signal. To capture the video signal, the video data must be digitized and stored in memory or in the frame buffer. The digitization requires special purpose (but commonly available) hardware and produces a digital image similar to a Sun raster file. For the image to be saved it must be transmitted across the available buses to memory. Capturing and storing or transmitting a continuous video sequence at 640 x 480 resolution requires from 9 to 27 Mbytes/second of bandwidth for 8-bit or 24-bit images. This stretches or exceeds many system buses such as SBus, LANs (Ethernet), or WANs (ISDN). Therefore, real-time video capture will generally also include real-time video compression to reduce the bandwidth required. (This is discussed further below.) Compression is also required to save video data. Without compression, one minute of video data requires almost a gigabyte of storage. Video capture is needed for many, but not all, full-motion video applications. It is obviously necessary for every workstation used for video-conferencing. It is also required for many video authoring situations (though the video capture capability can often be a shared resource). Full Motion Digital Video Display Display of full-motion digital video generally means starting with video that has already been digitized and compressed and is received from some network connection (Ethernet or ISDN), from a live video capture device (camera or VCR), or from some storage media (disk or CD-ROM). To display the video must be decompressed (preferably in real-time, e.g. 30 frames per second) and sent to the frame buffer for display in a window. Full Motion Digital Video Compression/Decompression Video compression and decompression is one of the areas of technology currently receiving the greatest amount of attention. As discussed earlier, video compression is a requirement for reducing network and bus bandwidth needs and for reducing space for video archiving. There are many ways to compress and decompress video, with different costs and benefits, both monetary and performance. The many techniques available or under development differ in a variety of characteristics. Characteristics of Compression/Decompression Techniques Some of the characteristics of compressing and decompressing full-motion video include lossless vs. lossy compression, compression ratio, intraframe vs. interframe compression, computation cost, and symmetric vs. non-symmetric compression. o Lossless vs. Lossy Compression Lossless compression means there is no information lost when an image is compressed and then decompressed; that is, the decompressed image is identical to the original image. Lossless decompression is often used for saving space on disk storage text and binary data. Lossless algorithms tend to provide only a small amount of compression - perhaps a 2:1 uncompressed-to-compressed ratio. Certain applications, such as medical imaging or satellite data interpretation, require lossless compression. The video compression algorithms used for full-motion digital video are generally lossy algorithms, that is, not identical to the original image. The amount of information loss, and thus the final image quality, varies considerably with the different techniques and within a single technique according to the parameters chosen. As a general rule, the greater the compression ratio, the greater the information loss. o Compression Ratio Compression ratio describes the change in size, or amount of storage space required, one achieves in compressing the image or video. o Intraframe vs. Interframe Compression Intraframe compression compresses a single image at a time, without regard to the previous or succeeding images. Other algorithms do intraframe compression and then do additional interframe compression. This technique is much more costly to encode but, for a given compression ratio, provides a higher quality sequence of images. This provides the very high compression needed for video conferencing over LANs and WANs. Techniques that are intraframe-only are more appropriate for still images. o Computation Cost Different compression techniques vary considerably in the amount of computation required to compress and decompress images. Some decompression techniques are relatively low cost (measured in time and money) and can be computed in software on standard workstations at reasonable speeds (for example 10-30 fps). Other techniques currently require special purpose hardware to perform compression or decompression at rates close to full-motion video. o Symmetric vs. Non-symmetric Compression A symmetric compression is one in which the computation cost for compression is roughly the same as the computation cost for decompression. A non-symmetric technique is one in which compression typically is significantly more expensive than decompression. Symmetric techniques have the advantage that specific hardware can be developed to provide both the compression and the decompression. Asymmetric techniques have the advantage of providing inexpensive playback without additional hardware. Compression Techniques As mentioned previously, there are a great many compression techniques in existence or currently under development. Some of the more common ones are described in the following paragraphs. o JPEG Joint Photographic Experts Group (JPEG) is close to being endorsed as an industry standard for still picture compression. It is an intraframe compression standard that achieves a ratio of about 20:1 r o MPEG Moving Pictures Experts Group (MPEG) is a video compression standard for full-motion video. It is an interframe method that can compress to a ratio of about 100:1. o Px64 Px64 is also known as H.261. This technique is a standard proposed specifically for video transmission over ISDN lines: that is, for video conferencing and videophone. o DVI Digital Video Interactive (DVI) is a proprietary video software architecture developed by Intel to run on their hardware. o RPZ - RPZ RPZ - RPZ is the Apple" proprietary compression/decompression technique and is part of their QuickTime` software for multimedia applications. Other Techniques There are also several lossless compression algorithms, including Run-Length Encoding and Huffman encoding. These tend to provide much less image compression. They are also often used in conjunction with the above techniques to provide some additional compression. Challenges Full-motion digital video on the workstation is currently possible and no longer a research question, but it is far from commonplace. The greatest challenge is to make video on the workstation inexpensive enough to have on every workstation, and easy enough to use so that it is no more complicated than email. The challenge of low cost is the easier one to meet. As CPUs become more powerful and as workstations are becoming multiprocessors, more video functions can be computed on a standard workstation. Furthermore, special- purpose chips to aid in video compression are becoming available and will advance along the price curve until they are on every system. Additional hardware, including cameras and higher speed networks, are also becoming more accessible. The challenge of ease-of-use is not so predictable. Widespread use is not likely to occur until multimedia presentations are as easy to produce as creating slides or displaying analog video on a VCR. Video conferencing will not be widely used until one can point and click and bring up a video conference window on the workstation. Furthermore, full-motion video must integrate seamlessly with the other aspects of multimedia - documents, fax, and audio.