Fullscreen HTML5Med Res (??MB)HTML4

Video Formats: WHY BOTHER? You’ve no doubt noticed that most DVD players have an “S-video” output, as well as the familiar “composite” video output. And many newer models also provide outputs for “component” video . So what’s the reason for this extra complexity when it comes to video signal connections? N OT SO LONG AGO, video was just video – or that’s the way it seemed. Video monitors and TV sets had single RCA sockets for their video inputs, as did VCRs for their video inputs and outputs. With this format, you simply fed the video signal from one unit to another via a single RCA-to-RCA coaxial cable, with other cables only needed for the audio. When Laser disc players came along, most of them used exactly the same arrangement (although they gave much clearer pictures than VCRs). However, when DVD players arrived, even the early models had an extra video output socket (usually a 4-pin mini DIN socket) which was marked “S-video”. At the same time, TV sets also started to appear with an S-video input socket – this in addition to the more familiar RCA-type video input, which was now being called the “composite video” socket. So you now had a choice when it came to connecting a DVD player to the TV – use either a single RCA-RCA cable or one of the new 4-pin DIN to 4-pin DIN “S-video” cables. And the word soon spread that using an S-video cable gave better picture quality. Then things got a little more complicated again. Some of the higher-end 8  Silicon Chip DVD players started to appear with a third kind of video output known as “component video”. This was usually made available via three more RCA sockets marked Y, Cb (or Pb or B-Y) and Cr (or Pr or R-Y). Naturally, component video inputs also began appearing on TV sets and video projectors at about this time, giving the consumer yet another choice when it came to connecting video signals. As before, word soon spread that using component video cables gave the best possible picture quality – even better than S-video. And it wasn’t too long before component video outputs appeared on even low-end DVD players. So what’s it all about? Why have video connections become so complicated and do the fancy, newer formats really deliver better picture quality than good, old composite video? Let’s find out! About composite video First of all, let’s talk about composite video. As the name suggests, this really isn’t just one signal but is a “composite” or a collection of a number of signals (or components). First, there’s the black-and-white or “luminance” (Y) video component, which conveys the basic picture detail by JIM ROWE and contrast information. Then there’s the “chrominance” (C) video component, which conveys the picture’s colour information (the chrominance component is itself actually two components, not one, but we’ll go further into this shortly). Finally, there are the synchronising pulses and the colour subcarrier burst pulses, which collectively form a third component in the composite video signal. Although these components are all lumped together and sent along a single coaxial cable, they really are different video signal components with distinctly different functions. But why were they originally all lumped together to produce composite video signals in the first place? The answer to this is that when TV broadcasting began, engineers needed to pack all of the video components into a single video signal to be modulated onto the TV station’s radio carrier (along with the sound signals, of course). This also meant that when the TV signals were demodulated again in the TV set, they reappeared initially as the same composite video signal. In the TV set itself, the composite video signal then had to be split up into its various components before the pictures could be displayed on the siliconchip.com.au screen. First, the luminance information had to be extracted so that it could be used to vary the three picture tube beam currents together (ie, from the three “guns”), to recreate the picture contrast and details. Second, the chrominance information had to be extracted so that it could be used to control the beam currents individually, to recreate the picture colours. And third, the synchronising pulses and colour burst information had to be extracted so that it could be used to lock the picture scanning oscillators and ensure that the colour information was decoded correctly. Whew! When VCRs subsequently came along, the easiest way to handle the video information that they fed to a TV set was to use this same composite video format. That’s why domestic video connections were originally all made using the now-familiar single coaxial cables with an RCA plug at each end, usually with yellow colour coding. Compromises, compromises Although composite video signals can produce quite good picture quality, there are a few compromises involved in “packing” all of those video components into a single composite signal – and then subsequently processing them in this form. The big problem is that it’s relatively easy for the various signals to interact with each other, in a way that actually degrades the ultimate picture quality. Probably the most serious type of siliconchip.com.au Fig.1: as shown here (top), the luminance and chrominance signals share the frequency spectrum between 3.2MHz and 5.5MHz. The magnified view shows how the two sets of information exist in evenly spaced “clumps”, with the colour clumps neatly slotting between the luminance clumps. interaction that occurs is “cross modulation” between the luminance (Y) and chrominance (C) information. This can happen fairly easily with composite video, because of the way the chrominance information is conveyed as modulation on a separate colour subcarrier, which has a frequency of 4.433MHz for PAL video or 3.58MHz for NTSC. Although the colour subcarrier itself is suppressed in the video signal (and recreated in the receiver using the burst information), the actual colour information “sidebands” share some of the same frequency spectrum as the luminance information and are actually interleaved with it. This is shown in a slightly simplified form in Fig.1. As you can see, the luminance and chrominance signals actually share the frequencies between about 3.2MHz and 5.5MHz. The magnified close-up view shows how the two sets of information exist in evenly spaced “clumps”, with the colour clumps neatly slotting between the luminance clumps. This interleaving was done deliberately, in an effort to minimise the interaction between the two components. However, it doesn’t entirely prevent interaction, which is why you tend to see shimmering “cross-colour” bands on a picture area where there are finely spaced lines, such as a finely checked shirt. This type of visible Y-C interaction was much more pronounced with early colour TV receivers, because they had to use fairly traditional analog filters to separate out the luminance (Y) and chrominance (C) information. The problem here was that the low-pass filter used to extract the Y information had to have a cutoff frequency no August 2004  9 and R-Y). DVD player outputs Fig.2: this diagram shows the different signal processing paths involved for component video, S-video and composite video. Component video has the least amount of processing (and the best picture quality), while composite video has the most processing (and the worst picture quality). higher than about 3.2MHz, in order to filter out all the colour information. Similarly, even if the high-pass or bandpass filter used to extract the C information had a lower cutoff frequency very close to 3.2MHz, there was still be quite a bit of Y information present in the chrominance signal. By the way, notice that with this analog filtering method of Y-C separation, all luminance information above about 3.2MHz must be “thrown away”, to avoid getting colour information mixed in with the luminance. So the picture resolution is degraded as well. To get around this problem, later colour TV receivers (as well as recent monitors and video projectors) use a more sophisticated technique to separate out the Y and C information. This technique is known as “comb filtering” and it involves the use of digital techniques to produce filters which have responses shaped like combs – with “teeth” that can separate the two sets of information clumps. One filter extracts all of the Y information clumps, while the other extracts all of the C information clumps. That way, the Y and C video information can be separated properly, without sacrificing the bandwidth of 10  Silicon Chip either. The result is clearer and sharper pictures, with a minimum of crosscolour interaction. DVD recording format Of course, the only way to completely ensure that there’s no interaction between the luminance and chrominance is to keep them separate in the first place. And that’s why when the standards were being developed for DVD video discs, it was decided that the video would actually be recorded in “separate component” format – with the Y information kept completely separate from the C information. What’s more, even the C information would be split into two separate components, to keep the colour “cleaner”. As you may know, both the video and audio are recorded on DVDs in digitally compressed form (MPEG-2), to allow everything to be squeezed into a maximum bit rate of 9.8Mb/s (megabits per second). But the video is still kept as three separated components, even when it’s digitally compressed. So when a DVD is played back, the initial output from the player’s MPEG decoder section is in component video form: the luminance (Y) signal plus two “colour difference” signals (B-Y When the first DVD players came out, most TV sets were only provided with a composite video input (that’s if they had a video input at all). So, to ensure that people would be able to watch DVDs on their existing sets, the manufacturers fitted their players with additional video processing circuitry, to combine the decoded component video signals into composite video. That was fine but it meant that the video signals had to be passed through extra processing circuitry in the player to produce the composite video. It then had to go through a full Y/C separation and colour separation process in the TV set again, to produce the three component video signals needed for the TV set or projector to display the pictures. These steps are shown in Fig.2. As you can see, this way of playing DVDs via the composite video path involves quite a bit of processing, not only in the player but in the TV set (or projector) as well. The component video signals have to be combined in the player and then separated again in the TV set or projector – all so we can connect the two pieces of equipment using a single video cable! All of this extra video processing inevitably causes signal degradation. And because we deliberately force the various video components through a composite video “tunnel” (ie, the cable at the bottom of Fig.2), it also tends to introduce some Y-C interaction. That’s a pity, because the signals actually coming off the DVD video disc already match the component video format that’s ultimately required inside the TV set to display them. This is also illustrated in Fig.2, which shows that much less processing is involved for component video signals. Obviously, it’s far better not to combine the component video signals at all but to send them to directly to the TV or projector in their “native” form, to drive the display circuitry. S-video input The first big step forward was when TV and projector makers started providing their sets with S-video inputs, which could at least cope with separated luminance (Y) and chrominance (C). This had already started by the time the first DVD players appeared, because S-VHS camcorder makers siliconchip.com.au had got the ball rolling by fitting their products with S-video outputs. This was done so that consumers could take advantage of the improved picture quality possible with S-VHS. By providing their first-generation DVD players with S-video outputs (as well as composite outputs), the DVD makers made it possible for consumers to take advantage of the better picture quality offered by DVDs. As you can see from Fig.2, an S-video link at least bypasses the Y/C combining circuitry in the DVD player, as well as the Y/C separation circuitry in the TV or projector. This removes two signal processing steps and also means that the Y and C components are never combined at all – not even briefly. As a result, Y-C interaction is avoided completely. Users soon found that S-video was well worth the extra hassle of having to use a different video cable. However, the picture quality would be even better again if the pristine component video that came direct from the DVD player’s MPEG decoder could be piped directly to the display circuitry of the TV set or projector. Of course, this couldn’t be done until TV and projector makers started providing their sets with component video inputs. Once such sets began appearing, DVD players with component video outputs began appearing as well. As a result, consumers could finally feed fully separated component video signals from DVD players directly into their TVs and video projectors. Get the idea? Although S-video and component video connections might seem to be more complicated and messier than composite video, they’re actually less complicated for the video signals. That’s because the components are kept separate and go through much less processing. And that means they’re degraded less and so you get clearer pictures. What about RGB? Some TV sets of European origin are provided with inputs for component video in yet another format known as “RGB”, where the three primary colour signals are already separated. This type of component video outputis also provided by some pay-TV and digital set-top boxes. In theory. RGB should offer slightly better picture quality again than Y/BY/R-Y component video, because siliconchip.com.au the display drive circuits in a TV or projector do ultimately need the video signals in this very form. However, in practice, the picture quality is often much the same, because even if your set has direct RGB inputs, the signals still have to be converted into this form (from Y/B-Y/R-Y) in the DVD player or set-top box. The proof is in the picture Perhaps you still don’t quite believe that S-video and component video really deliver better picture quality. Well, the best way to be convinced is to compare them with your own eyes. But since you may not find it easy to do this, we’ve taken close-up shots of part of a standard test pattern image, as reproduced from a PAL DVD test disc on a video projector. The first picture was obtained using a composite video link, the second using an S-video link and the third using component video links. These pictures will give you at least some idea of the improvements that can be achieved. Notice in the composite video image that there are bright multi-coloured fringes in the circular “Fresnel Zone Plate” pattern at centre left. These all consist of “fake colour”, caused by high-frequency Y information getting into the colour information (ie, crosscolour interaction). There are also weak bands of fake colour in the two frequency band squares at top centre of this image. As you can see, the luminance response does extend all the way to 5.5MHz, as shown by the tapering lines on the right of the image. This is presumably because the projector used to display these images uses comb filters to perform the Y/C separation from the composite video, so the upper luminance frequencies are not being “thrown away”. Still, those fake colour artefacts do result in noticeable picture degradation. If you compare the S-video and component video images with this first image, you’ll see that there is much less colour fringing using the S-video signals and virtually none at all using component video. There’s no doubt that the S-video link gives significantly clearer pictures than composite video, while component video gives the cleanest and sharpest pictures of all. Note: Sanity currently stock the disc at www.sanity.com.au or phone 1300 722 121. SC (1). Composite video (2). S-video (3). Component video Fig.3: these three pictures clearly illustrate the improved picture quality delivered by S-video and component video signals. These’s much less colour fringing using S-video compared to composite video, while component video gives the best picture of all. August 2004  11