Wondering about Widescreens

March 29, 2004

Introduction

Visual presentations adapted for wide screen formats have existed since the days of the synchronized 35mm slide presentation. Systems using thirty-two or more slide projectors with soft-edge slides and paper-tape synchronizers were the rage at one time and many AV technicians spent hours in dark rooms listening to a music track accompanied by the click-clack movement of slide trays.

When video projection became an alternative to slide, wide-screen video presentations could be created using multiple projectors and in some ways, the projection equipment was well suited to the application. CRT projectors of the era possessed adjustability and contrast ratios that far exceed what is possible with most modern projectors. However, many other challenges, such as content creation issues and technical instability made wide-screen video an exotic undertaking.

As projection technology advanced from the dim, but infinitely adjustable CRT technologies to LCD, DLP and other digital imaging technologies, the emphasis shifted from wide format to high-resolution and ease of use. With modern projectors it became easy to throw up a high-resolution computer image, while more complex non-standard formats became more difficult, due to contrast limitations and the lack of adequate multi-screen image processors.

As of 2004, technology has caught up once again and there are now image processing systems that work very well with today's powerful, flexible projectors. In combination, this technology can be used to create presentations in any screen format that can be imagined, with any type of content you are likely to use; including standard video, computer graphics and the myriad High-definition Video formats.

 

The Presentation System

Before we go any farther, an examination of the component parts of the system and what their purposes are should be performed. The presentation system is a chain; it starts with the source device and ends at the display device. In between those two components can be any number of boxes – routers, switchers and effects devices, for example – that are used to distribute and process the source signals before they hit the display device.

Why is this issue of wasted light and resolution important? When using 16:9 format screens with 3:4 projectors it isn’t (unless it’s for digital cinema or other critical uses).  However, when considering really wide screens or other unusual screen formats, the issue becomes visible to all and sundry, as the size of the pixel becomes intrusive and the viewer wonders why the image seems grainy and dull.  Fortunately for the intrepid Show Designer, good solutions, in the form of multi-projector systems, are available.

Employing more than one projector to fill a non-standard screen size helps maximize resolution and brightness. The array of projectors fills the screen, while each individual image matches the display imager shape (or native resolution) and not a single pixel of resolution or lumen of brightness goes to waste by overshooting the screen as would happen with a single projector. Of course, by using multiple projectors we create other challenges, including image blending and overlap processing, that must be overcome and that’s where the subject of this paper becomes important.

Vista Systems Montage

For a system to work properly, a number of changes must be made to the signals output by the source devices. In the case of Montage, the multi-screen display processor developed by Vista Systems, the incoming source signals are digitized and when selected by the system operator, the digitized signals are pushed down a data bus and picked up by an output buffer. This output buffer chooses which portion of the signal will be displayed in a particular area of the pixel map addressed by the Montage processor.

Let’s review this concept by looking at signal flow end to beginning rather than beginning to end. In the three-projector system described each projector would be connected to a specific Montage output. However, ALL three projectors are used to create a single “pixel map” that represents the entire image area. The Montage operator selects the sources that are to be displayed and then manipulates them in the form of “Picture in Picture” (PIP) windows, which can slide back and forth within the overall pixel map.

The Montage operator is also able to change the size of the PIPs and the speed of their movement, add borders to windows, change transparency levels, cut, wipe and cross-fade. The PIP’s can also be layered and revealed using the available effects. All of this happens within a frame and a half of real-time, irrespective of signal type, which makes Montage a truly “live” system. 

In addition to the visible image effects that Montage enables, there’s an important image manipulation function that the system handles in the background – edge-blending. This technique is used to combine multiple projected images in a single, continuous image without visible seams, by overlapping the edges of projected images.

There are two parts of the edge-blending process: image-doubling and soft-edge blending, both of which are accomplished within Montage. Image doubling is used to place shared portions of images on adjoining projectors, while soft-edge blending applies an electronic mask to the vertical edges of the projected images. The electronic masking, in simple terms, darkens the image in a graduated fashion so that where the two overlapped portions are combined, the overlap area is uniform with the areas that are not combined.

Montage allows overlap areas of different sizes, providing flexibility for screen sizes, and it can be set up with no overlaps if screens are to be separated. Separating screens in a system like Montage might seem contradictory, since its main purpose is to create seamless wide screen images. However, the requirements of a particular set design might not accommodate a single screen, for example, so multiple screens may be a necessity. In this application Montage is still an attractive alternative because it allows freedom of image movement across screens, whether they adjoin or are separated by a significant distance.

The kind of magic that Montage performs has been part of the high-end presentation scene for a long time. However, up to now, the magic act has been compromised by a number of factors and that is what makes Montage a revolutionary system. Prior to Montage, systems made up of bits of this and that were wired up to create wide-screen image systems. Employing routers, scalers, video switchers and machine controllers, these systems were difficult to operate and often relied on cardboard and gaffer’s tape to effect the soft-edge blending.

They also suffered from image quality issues such as motion artifacts and excessive frame-delay, which made live camera images appear on the screen after a noticeable delay. Montage makes the magic happen without compromising image quality and with minimal frame delay making it the ideal system for layered wide-screen presentations that must be operated live or in a semi-automated fashion.

While Montage can be operated in fully manual mode (as in, Director wants to display Camera 1, so Montage Operator presses button marked “Camera 1”), the system can be programmed comprehensively so that complicated sequences are automated and initiated based on manual triggers or external control triggers. Typically, a Montage show will be a hybrid of manual and automated cues based on the show requirements.

For example, imagine a show that begins with sequence of tightly synchronized lighting and video moves, where multiple sources must play at a precise moment and PIP windows move in a specific timed pattern. While it is theoretically possible for the operator to punch all these moves with enough speed and accuracy to make the sequence work, the better alternative would automate these moves to some degree. Simple sequences can be built on the spot within minutes while more complex sequences may require thirty minutes or more and a couple of run-throughs to confirm that the programming is correct.

Usually, Montage programming takes place on site during set-up and rehearsals, although it’s not uncommon for shows to be pre-programmed, to some extent, in order to minimize the amount of work on show site. Montage’s architecture supports intensive programming even in the absence of show media, allowing considerable work well in advance of show media being delivered.

Although Montage is a system based on specialized hardware, considerable improvements will be made via future software upgrades. The upgrades will mitigate the few existing limitations of the system, most important of which are the 7,000,000 pixel map limit and an inability to soft-edge blend vertically as well as horizontally. Expansion of the pixel map will allow a single Montage system to address more projectors, higher resolution projectors or a combination of the two, while horizontal and vertical edge-blending open up the possibility of mixed vertical and horizontal screens in a presentation.

In the case of Montage, all this power comes at a considerable price, as systems can easily cost more than $200,000. Fortunately, there are wide-screen alternatives for less ambitious budgets that can also deliver the “goods”, that is, dynamic wide-screen presentations that have startling image quality.

Dataton Watchout

Dataton Watchout is a wide-screen system that is almost the polar opposite of Montage in its approach. It is relatively inexpensive, but is not easily adaptable to live operation.

Watchout is a software product that runs on networked computers and can be used to create seamless multi-projector displays of various sizes and configurations, whether vertically or horizontally oriented. The concept behind it is elegant and ingenious. Figure 4 illustrates the basic architecture, which utilizes WinTel platform computers running Windows XP connected by a standard TCP/IP network. The network is comprised of a Master or Composition Station and a variable number of Client Stations determined by the quantity of displays that are addressed.

The Master Station application is used to configure the system for the parameters of the presentation, including the number of projectors, overlap percentage and other edge-blending parameters. The Client Stations connect to the projectors or other display devices using the PC graphics card output port in either RGBH/V or DVI format.  

A Watchout presentation is an assemblage of various media assets, such as JPEG still images and QuickTime movies that are placed within the image area and then animated, layered and sized to appear as a coherent presentation. This programming takes place on the Master Station, which contains the authoring application as well as the media assets that are used in the presentation.  When the programming is complete or needs to be seen on the actual displays, the program renders a specific show file for each client station and those files are downloaded to the clients across the network. The show is executed when the system runs the individual files in perfect synchronization.

If the show needs to be modified, the changes are made on the Master Station and the revised show is re-rendered and distributed to the Client Stations. Using current high-end PC technology, this process is relatively quick and painless for except the largest Watchout shows. For most shows, the rendering and distribution process takes less than a minute. Once the show is loaded onto the Client Station, the show is controlled (Stop, Start, Pause, Jump) from the Master Station by manipulating a show timeline.  This timeline can also be controlled from an external control system to provide comprehensive automation capability.

However, the fact that the show must be built and rendered limits Watchout to applications that do not require live or real time media displays, In other words, Watchout is suitable for “canned” shows that are totally programmed in advance with all content on-line and loaded. Having said that, we can backtrack a little bit, because the latest version of the program enable some use of live content by adding video capture cards to the Client PCs. The capture cards can then be connected to live video or graphics source devices. Using this architecture an external videotape source or even a PowerPoint show can be displayed as part of the pre-programmed Watchout show.

Watchout is a platform from which the highest image quality presentations can be created and that is what sets it apart from all others. Image reproduction of animated still images are flawless, animations are smooth and with recent advances in High-definition compression techniques, the system can now support HD video within the Watchout framework. With the right creative approach, even a simple Watchout show can be a thing of great beauty.

Summary

Table 1 summarizes and compares the two products, Vista Montage and Dataton Watchout. Both are excellent at what they are designed to do and can even be used together very effectively. Both products require experienced operators or programmers and, perhaps most importantly, will only perform well if the content fed into the system is high quality.

Table 1

Criteria	Dataton Watchout	Vista Montage
Cost Factor	Low	High
Technology basis	Software	Specialized Hardware
Built-in edge-blending	Yes	Yes
Supports horizontal & vertical	Yes	No
Handles live sources	Sort of	Yes
Will control external devices	No	Yes
Can be controlled externally	Yes	Yes
Content authoring	No	No
Can be operated manually	No	Yes