GV - October 1997
The Present and Future of
Scan Conversion
by Sheldon Liebman
Now that we finally have standards for Advanced Television, the relationship between computer generated displays and video output is bound to undergo some significant changes. With that in mind, this is probably a good time to review the concept of scan conversion as it exists today and is likely to exist tomorrow.
If computers and television used the same display format, there would be no need for scan converters, which take a computer display signal in and put a standard video signal out. With the new standards, it is possible that in certain instances, this will become true in the future. However, given the large number of formats that have been defined, it is unlikely that any computer display will be designed to output all of them, so some degree of scan conversion will remain necessary in order to provide general purpose conversion from computers to video.
Why Bother?
The first question many people ask is why consider scan
conversion at all? Depending on the application, they may be
absolutely right. However, given that this is Government VIDEO
magazine, the assumption is that most of us display, record, edit
and/or distribute video as a part of our jobs. When images from a
computer need to be integrated into this environment, a scan
converter must be used.
In the simplest case, a scan converter accepts a single format of computer display creates a single format of video output. The best examples of this are the sub-$200 devices advertised in many of the computer catalogs. Plug in your 640x480 60 Hz VGA signal and get a full screen representation of the screen in a composite NTSC format. In fact, some laptop computers and VGA cards already have this type of low-end scan conversion built in.
However, if your requirements don't fall into this very narrow area, you need to look at the other features that higher end scan conversion products can supply.
Choice of Input
Resolutions
The first thing to consider is the input resolution that
will be fed to the scan converter. Even laptop computers have
screens today that are larger than 640x480, so the ability to
input a higher resolution is often important. Generally, scan
converters accept a maximum resolution that corresponds to one of
the standard VGA formats - 800x600, 1024x768, 1280x1024 and
1600x1200. With the last two resolutions, a scan converter can
also be used with workstations like those from Silicon Graphics
and Sun Microsystems.
As the input resolution goes up, however, the quality of the output is affected. Even if there is a way to sample the input signal often enough to capture every bit on information on a single line, the number of lines that come into a scan converter at high resolution exceed the number of lines that can go out. As a result, the resolution of the image must be lowered to more closely match the output. The more of a change you need to make, the more original information can be lost. For example, a 1280x1024 image contains over 1.3 million pixels while a 640x480 image has a little more than 300,000. To convert from one size to the other, you need to represent the same image using less than 25% of the original information.
The way scan converters deal with this is to filter the image as it is being reduced. The better the filtering technique, the more detail will appear to be in the lower resolution output. Even the best filters, though, can't perform miracles. If you have small text on a high-resolution screen and convert it to video, it is highly unlikely that you'll still be able to make out the words. This is probably the single biggest problem associated with using scan converters.
Choice of Output Formats
Even before the FCC made its ruling earlier this year,
there were many different ways to represent a video signal.
First, there are multiple standards already in use around the
world - NTSC, PAL and SECAM, plus some variations on those in
certain countries. Second, there are different ways to store
information in each format. The standard signal that we use to
distribute video programming is composite video, where the entire
video signal is present on a single wire.
Component formats start to separate the signal further and can improve quality, although they can't be used for transmission and are often impractical for distribution. For many people, the biggest improvement comes in the first component step, where the composite signal is broken down into two components, the Luminance (Y) and the Chrominance (C). This is the basic description of the S-VHS and Hi-8 formats. Further separation results in three signals and forms the basic for the original BetaCam format and all of the digital formats from D1 on up.
The more signal separation you can achieve, the higher the quality of the output. As you compare scan converters to one another, it's important to look at what signals that actually support. Last year, the first products to support D1 appeared and now multiple products are available for this high quality format.
Basic Picture Adjustments
Just as every television is different, every graphics
card is too. So, it's important to have the ability to adjust the
relationship between the input and output of a scan converter. On
all but the most basic units, adjustments are available for size
and position in both the horizontal and vertical directions.
Another basic capability that exists on most units today is a
switch for underscan and overscan.
When you work in a video environment, it's easy to take into account the "safe title area," which varies from TV to TV but it usually 5-10% of the screen in every direction. Computer displays, on the other hand, like to place information all the way up to the edges. Sometimes, you need to adjust the size of the output so that you can see all of the information. If there is enough play in the horizontal and vertical size adjustments, you may be able to accomplish this. A much easier way, however, is to flip a single switch that squeezes in the output image in all four directions at once. Production video monitors have this ability, but you can't always count on that when you distribute a signal for use on consumer displays.
Advanced Picture
Adjustments
Sometimes, you don't really want to convert all the
information that's on the computer screen. In this case, you need
to be able to zoom in on a specific area and have that area fill
the entire video output. Pan/Zoom/Scroll adjustments allow this
to happen. Depending on the product you choose, there may be
limits to how much you can adjust this or whether you can change
the aspect ratio of the images. However, if the information you
want to show is only on 1/4 of an SGI screen, for example, this
acts as a sort of selectable video window for output.
Integration with Live
Video
Another issue that often arises for video professionals
is the ability to merge the output of a scan converter with other
video sources. If the video signal is recorded onto tape and then
mixed in a production studio, there should not be any problems
mixing it with other taped or live material. If you need to mix
in a live environment, though, you need to make sure the scan
converter supports genlock.
Using genlock, you can adjust the timing of multiple video sources to coincide exactly. The two most obvious areas this affects are position and color fidelity. If two signals are fed into a mixer, using either one independently is usually not a problem. But, when you try to overlay one on another or fade between them, they must be timed correctly. If the position is off, the images will move one location to another on the screen as soon as they are mixed. If the phase is incorrect, the colors will change. Neither of these is appropriate.
Other Considerations
Finally, there are a few other considerations that can
enter into a decision about which scan converter to purchase. One
is the interface for adjusting the unit. Does it use switches and
pots or is there a menu system? Is there support for remote
control? Can you send commands from a computer? Another issue is
the packaging itself. How large is the unit and does it fit into
a 19" rack? Can it easily move between computer sources? Is
it available as a board to fit into a PC, Mac or other computer?
Depending on your requirements, one or more of these issues may
be critical.
Where Do You Want to Go
Today?
The scan converter market continues to evolve. As
mentioned above, support for digital video formats (usually D1)
is now available from companies like Folsom Research (9700XL with
D1 option), RGB Spectrum (VideoLink 1700D-1) and The V.A.S. Group (RTC D1).
Extron has expanded its product offerings at both the top and bottom ends. For the first time, the company introduced a product capable of converting 1280x1024 source images to video. This product, the Emotia Xtreme, is available in both standard and MX models. The MX adds some additional output ports and is rack mountable. At the low end, Extron has added the Emotia Jr. 800 to their product line. This is a low cost ($995) unit supporting up to 800x600 resolution.
There are no new products at this point from Communications Specialties and PC Video Conversion. Their Scan Do (Communications Specialties) and HyperConverter (PC Video) products are still very popular for many scan conversion tasks.
Where Do You Need to Be
Tomorrow?
There's another new product from The V.A.S. group that
addresses some of what we'll need tomorrow as we all try to deal
with the shift from traditional NTSC and PAL video into any or
all of the 18 Advanced Television Standards that have now been
set. The RTC HD 3:2 is designed to interface the world of
high-resolution video production with current video standards.
This very high end product accepts any of the 18 ATV formats and
converts them to high quality NTSC or PAL video. This allows
high-resolution masters to be made while still providing a method
for distributing the finished production in the current video
environment.
This product is one part of where the video production world will inevitably end up over the next few years. With multiple video formats to choose from as well as a variety of computer formats, it's going to be more and more important to be able to choose production and distribution formats independently.
Some companies want to do everything at high res and distribute it in all formats from there. Others will want to work at lower resolutions and have a way to "bump it up" when necessary. The final solution can only be a product that takes anything (or everything) in and puts anything out. Today, trying to accomplish this goal requires a combination of scan converters, line doublers, transcoders and other products. In a few years, though, this may all be available in a single product.
Imagine recording high-resolution computer output as high-resolution video output. Or seamlessly mixing high res and lower res material so that it can be broadcast in multiple formats. The first company to introduce this type of product will definitely be someone to watch.
Back to Recent L&S Marketing Articles
lsmweb