As we approach the end of analog broadcasting, a scant 22 months away, it is useful to look at what it takes to build a workflow and facility that seamlessly integrates HD into an SD world. Broadcasters are increasingly building news operations around the inevitability of HD local production, with SD slowly becoming the legacy format.
HD represents a totally new medium in addition to the obvious technical differences. It requires careful attention to sound and picture quality, both of which were not as urgent in the past.
Changes in the studio
HD cameras provide less depth of field given identical focal length, shooting distance and f-stop. This can greatly affect planning a new set and the production process. Items out of focus on an SD set would still be visible on an HD set. In addition to the increase in resolution, this effect puts significant pressure on keeping sets in pristine condition and managing the contrast in the set design to achieve the intended look.
Lighting also contributes significantly to managing the contrast and apparent detail. HD studio cameras have about the same sensitivity as SD cameras, but to achieve depth of field similar to SD shots, it may be necessary to lower the lighting levels.
Even when HD cameras are used to create downconverted images for SD, the sharpness of the SD pictures increases. Don't assume that because the output is eventually SD, the issues, including depth of field differences, no longer exist.
Remember that sharpness and resolution are not the same. Apparent picture sharpness is related to the square of the area under the MTF curve, while resolution is a limiting value for detail that can be seen. (See Figure 1.)
Talent will have to learn new makeup techniques that can stand up to higher resolution. Some traditional makeup approaches used for SD look almost theatrical when shot with HD cameras. Lighting can be important in softening what might be perceived as excess resolution that makes facial features, shall we say, more distinct. This is an important issue to programmers and production professionals, as well as on-air talent, and it cannot be minimized.
Aspect ratio
In addition to the factors related to the physics of optics and HDTV scanning, there are other critical issues. The most obvious issue is the wide aspect ratio of HD images. Studio sets may need to mimic the aspect ratio to look natural. Dual SD and HD outputs will be around for some time into the future, so whether at the transmitter or at the set-top converter, it is important to protect the image for later formatting to SD. This technique is often called center safe.
The need to cut content from regular programming has caused most broadcasters to opt for a center cut from the HD image for delivering the SD copy, keeping the top and bottom of the frame aligned, and cutting off one-eighth of the frame from each side. (See Figure 2 below and Figure 3 on the next page.) This results in an over-sampled NTSC picture, with excellent picture quality and a reasonable compromise between the best resolution possible and the most appropriate rendition of the scene.
Tapeless workflow
Other parts of the transition are not quite as settled as aspect ratio and staging issues. News operations are in the throws of conversion to so-called tapeless workflows. Unfortunately, just as momentum is building and stations are making that transition more smoothly, an immature HD news production infrastructure creates new barriers that deter progress.
HD news requires about three times the storage bandwidth of SD. HDV recorders eat up about 25Mb/s, and Panasonic DVCPRO-HD, generally on P2 cards, requires four times the bandwidth. Sony's XDCAM-HD records at 18Mb/s to 35Mb/s on Blu-ray disks.
This is happening at a confusing time in the industry. Hard disk storage is now as cheap as videotape. According to Screen Digest, disk storage of HD content is more than 20 percent cheaper than popular tape formats. By 2008, it is expected to be less than 30 percent of the cost of storage on tape. This offsets the cost of switching to nonlinear workflow and will reverse the adverse economics that HD news faces this year.
Infrastructure
In addition to the cost issues, there is simply less infrastructure available for HD nonlinear news, and what is available is still new in the marketplace. The cost and training barriers are exacerbated by the need to integrate SD legacy and field footage into HD news programs. Even the national news providers that have announced plans to transition to HD acknowledge that field acquisition will remain 4:3 SD, while converting to 16:9 SD and HD over an extended period of time.
At the network level, the investment required to fully enable HD acquisition and production for news is a huge barrier, one without proven sources of new revenue to justify the investment. It's possible that HD news conversion will proceed for the same reasons that helicopters and other large-ticket purchases for news are made, which is of course competitive pressure.
HDV
One factor that helps significantly in ramping up the conversion of field acquisition to HD is the availability of HDV as an acquisition format. Although it was initially conceived as a format for consumers and prosumers, the incredible picture quality and availability of professional features on cameras costing less than $7500 make this an attractive option for newsgathering at all levels.
Network conversion strategies announced by more than one company include the use of HDV as a primary newsgathering system. Editing HDV in some systems has been available for more than a year. Now it is quite possible to integrate HDV into a professional workflow.
In most cases, a nonlinear workflow can accommodate the changes in aspect ratio and resolution in the editing system. This leaves a burden on the editor, but allows the freedom to compose a story using both SD and HD content on one timeline. Render times will suffer, of course, which might make external format conversion before editing more desirable.
Control rooms
The control room is in a state of conversion. For the last decade, it has been considerably more expensive and complex to design a production system for HD origination. As HD equipment has become less expensive, and more available in all product categories, the incentive to build less expensive SD control rooms has eroded.
Now the premium for HD facilities is less than 20 percent, and compared with SD costs a few years ago, it is now about the same price. The complication of building a facility that integrates both formats is that many converters might be necessary to convert all media to a house format, for example 1080i, before feeds are recorded and sent to a control room. One factor that makes this a bit easier is that most HD VTRs have downconverted SD outputs for use as needed, and some format agile SD-capable machines have upconverted outputs as well.
That leaves the most likely choice an HD control room from which a center-cut SD program is derived. Several manufacturers have developed production switchers that integrate both SD and HD internally, lessening or removing the need to convert signals before feeding to the production switcher.
Some of these switchers can process both SD and HD simultaneously. This is a powerful workflow feature that removes the need to build a single format facility and leaves flexibility for the production to seamlessly adapt to content that arrives unexpectedly in the wrong format.
In sync
Studio, control room editing can integrate HD and SD. There is no free lunch, even, or perhaps especially, with audio. Digital production switchers, monitor wall multiviewer processors and flat-screen displays all have latency that is not present in audio. This requires considerable care when designing a control room to ensure that the production team sees and hears precisely the right signals, and that sync is carefully maintained before it leaves the control room. Lip sync is the biggest complaint from consumers, and if facilities don't treat it with caution, it will be difficult to address effectively later. BE
John Luff is a broadcast technology consultant.
