HomeSoftwareArchiveTop TipsGlossaryOther Stuff







Digital video is coming to a PC near you, but let’s not get carried away; Rick Maybury looks at the hard facts behind multimedia



It’s very easy to get swept along by the multimedia bandwagon and end up believing that computers are going to take over the world; before you know it we’ll be throwing away our VCRs, edit controllers and post-production equipment... But hold on, let’s look at the facts, and what can realistically be achieved now, and explore some of the more fanciful claims being made .


But what about that word Multimedia, what on earth does it mean? The truth is it can mean anything and nothing, but behind it is the coming together of computers, with video and audio technologies. Traditionally computers carried out mundane and repetitive tasks, like number-crunching and word-processing, but within the past five or six years, as computers have become faster and more powerful, they’ve been able to handle more diverse forms of information, including sound and moving pictures. It’s a very powerful combination of technologies, and we have only just begin to explore the possibilities, so now let’s try and make some sense of what is going on.


From the video movie-making point of view there have been a number of key developments. The first one was edit control, a job computers are eminently well suited to, and the necessary control interconnection between suitably equipped camcorders and VCRs, and computers are reasonably easy to achieve. Next came graphics and title generation, again another fairly straightforward task for a computer, though getting video out of a computer and on to tape proved difficult at first on some systems, though nowadays even the most video unfriendly business-oriented PCs -- notably the IBM PC (and compatibles) and Apple Macs -- can be equipped with PAL-compatible video outputs reasonably cheaply.


After that the next logical step was to enable computers to process moving video, and that meant getting a video signal in to the machine, converting it to digital data, manipulating it in various way, mixing it with computer generated effects and other video sources and getting back it out again, on to tape. That took a little longer but once again most computers can now be fitted with video input/output cards, and there’s a huge range of titling and special effects software on the market, particularly for IBM compatibles.



The last and most difficult hurdle was to find a way of  storing moving video on a computer’s hard disc, and read it back out again -- effectively using a computer as a digital video recorder. This is multimedia’s Holy Grail, and a potential revolution for movie-makers as it opens the way for what’s known as non-linear editing. In theory video sequences stored on computer discs can be accessed and read out quickly, in any desired sequence. New scenes can be slotted in or swapped around  in a matter of seconds and  mixed with various effects. The results can be played back on screen almost instantly, and the output committed to tape. No more messing around with VCR source machines, waiting for tapes to shuttle back and forth or synchronising inputs, it’s quick and efficient, but there’s a snag, a very big snag...


The problem is moving video, when translated into a digital format a computer can understand, equates to a vast amount of information, around 30 megabytes per second in fact, far more than most machines can comfortably accommodate; some means had to be found to reduce the amount of data, without compromising picture quality. The solution is data compression, an extension of a technique originally developed to squeeze large amounts of computer data onto floppy discs and down noisy telephone lines. Video compression systems work by analysing each frame picture information as it comes into the computer and then dumping or ignoring redundant data.  For example, most of the time there’s relatively little difference between any two adjacent frames of a video recording, maybe some movement somewhere in the scene, but most of the picture -- the sky and background etc. -- is virtually the same from one frame to the next, so there’s no need to process duplicate that data. Compression ratios of 200 to 1 are now possible using systems developed by the Joint Photographic Experts Group (JPEG) and Motion Picture Experts Group (MPEG), but those levels of compression usually result in tiny, jerky postage-stamp sized images, which are completely useless for serious movie-making applications.


Lower compression ratios, or 15:1 or less are generally accepted as being of ‘VHS quality’, but that’s open to debate. Compression ratios of 10:1 are just about in the VHS ballpark but it’s significant that professional systems work on compression ratios as low as 4:1. Realistically that means decent quality images require fast, powerful computers with massive hard disc storage systems, that can store only relatively short video sequences; or to put it another way, you would need a couple of hundred thousand pounds worth of computer hardware to achieve the same level of picture quality and storage capacity as a domestic video recorder costing a few hundred pounds.


So, for the moment non-linear ‘on-line’ editing, where the PC video output is used to create the final master recording, is a non-starter for any video movie lasting more than a minute or two. Nevertheless there are plenty of applications where that’s sufficient, and the technology has already created a growing market in the production of short AV presentations, corporate videos, commercials and most significantly, multimedia software for CD ROMs, which rely on tiny, miniature moving video ‘windows’. PCs have also developed an important role as ‘off-line’ editing systems, where the computer is used as a creative tool, to assemble a kind of moving storyboard of edits and effects, that generate a template or time-coded edit decision list (EDL) for a conventional tape-based on-line editing system. That, in a nutshell, is the situation right now, but this is a fast-moving, open-ended technology that is advancing all the time. In  another year or two it could be a very different story; maybe then we will be telling you to dump your VCR and edit controller...



The biggest problem with JPEG and MPEG compression systems is that picture information or data is lost. Even seemingly irrelevant and repetitive data, representing an unchanging background or sky for example, will affect the quality of the image in subtle ways, if it is removed. Digital conversion is also unkind to colour, even the best systems cannot cope with more than a few thousand of the infinite variety of shades and hues that exist in nature. Digitally processed images almost always acquire a ‘texture’, it’s barely noticeable on professional and broadcast equipment, but still there, nonetheless.


The most obvious give-aways are ‘artefacts’. Artefacts are block-like shapes and shades that appear inside colours and around objects, where there’s a lot of contrast, detail or movement. They’re caused by the computer not being able to process information quickly enough. Compression systems don’t like sudden changes, they favour smooth, low-contrast images with large objects, and as little detail as possible.


JPEG compression systems have another undesirable habit, they ‘drop’ frames arbitrarily; needless to say this can cause havoc with on-line timecode editing, leaving awkward gaps in the recording. However, this can normally be minimised by selecting higher compression ratios; this will be at the expense of resolution, though this is acceptable for off-line editing purposes.


If you are recording material with a view to PC editing or processing there are a number of techniques that lessen or eliminate some of the worst effects. Keep the subject as large as possible in the shot, against a plain (preferably white or single colour ) background. Big curvy shapes are better than small ‘bitty’ ones. Avoid harsh shadows, small intricate detail and contrasty subjects. Try not to zoom quickly as compression systems sees this as movement and it will produce artefacts. If your camcorder has a portrait mode use it as the narrowed depth of field will help reduce background detail. Compression systems have difficulty differentiating between noise and detail, so only digitize clean, master-quality recordings. 



Ó R. Maybury 1994 2107



[Home][Software][Archive][Top Tips][Glossary][Other Stuff]

Copyright (c) 2005 Rick Maybury Ltd.