I suspect that many in the broadcast business have had the same reaction to the term “automation” that I do. I first think of labor “saving” and cost cutting, but careful reflection shows that automation is much more than that today.
Court TV recently went live from its new digital master control facility in Manhattan, NY, with a fully redundant two-channel Sundance Digital FastBreak automation solution. The system drives three Profile XP video servers: two PVS 1026s for play to air, and one PVS 1044 for ingesting the content. Photo courtesy Court TV. Photo by Andy Washnik.
The complexity of modern broadcast operations exceeds what a person could be reasonably expected to accomplish by manual action alone. Indeed, the automation pond has been entered on tiptoes in many operations in the last three decades, as spot playback moved from film and discrete videotape playbacks to robotic playback systems, such as initially the RCA TCR-100 and Ampex ACR-25. These systems, while primitive by today's standards, allowed 2-inch quad videotape spots to be sequenced into entire breaks without intervention. The labor freed from manually loading videotapes and hand switching between multiple sources was seldom turned to the street, but rather turned to other tasks to permit stations to expand their range of services without adding overhead. As labor costs increased over the years, the net savings to owners allowed margins to remain acceptable. The era of 2-inch robots gave way to a new generation of machines that were essentially the same (Sony Betacart/LMS, Odetics TCS series, Panasonic Mark), and performed the same functions, with some seeing use in news operations as well. However, seldom did stations in all but the largest markets opt for full station automation systems, in part due to the complexity, and cost. It is hard to justify the cost of an automation system (upwards of $250,000 in some cases) on the back of labor savings when the labor released was paid $10-$20 per hour. The math is simple: at $20/hour the hours saved would have to exceed six years of full-time labor.
Today the situation is radically different. The cost of an entry-level automation system is less than $20,000, and its capability is likely to be superior to that of a few years ago, when the cost was much higher. This couples with a radically changed set of requirements in station master control. Breaks are more complex, with effects, voice-over items, graphics, squeezeback, and shorter interstitial lengths feeding a structure that is hard for humans to accurately switch. The “coup de grace” that often forces stations' hand is the increasing number of channels being switched. LMAs, duopoly, centralized broadcast operations and local cable feeds often tax the conventional MCR system and staff beyond what can be reasonably and reliably handled. Quite simply, machines do not tire or make mistakes as often. With modern technology the failures are most often created by the human inputting the data on which the system runs. When traffic leaves errors that the machine intelligence in automation cannot analyze and repair, a human must step back in to figure out just what has gone wrong and how to fix the problem, hopefully before it gets to air.
The architecture of modern systems varies only in increments when viewed from 25,000 feet. Traffic outputs a log, automation ingests the data, parses it into commands for hardware playback devices, and then runs the resultant list on a strict time- line synchronized to a reliable clock. However at the 10-foot level the view can be quite different. Automation systems vary in scale to change operations in the markets for which they're best suited. It might serve well to review the areas where systems vary in order to put structure to the discussion.
Most automation systems offer satellite control and program record. These are the basis of much of the program schedule for many stations and, as such, can separate vendors quite rapidly. Important points to consider include the process of record schedule generation, resolution of conflicts for available hardware (VTRs or server ports, crosspoints, antennas and antenna controllers, and receivers), status monitoring and reporting. A good system will be able to handle multiple record devices and schedule the use of a considerable amount of time in advance if the data is available.
Spot ingest is becoming a moving target. Though for more than a decade now many stations have played spots to air from a server after ingest from a VTR, spot delivery to stations is now increasingly from a service that “drops” the content on a small-scale server along with metadata describing the content (advertiser, agency, air dates, spot ID number, etc) instead of a physical videotape shipped in from an agency. Fully featured ingest must in the future have the ability to query a directory on the agency delivery system and pull content into the system, notifying all concerned of the arrival, or perhaps notifying the appropriate parties of the non-arrival in time to find a solution. Ingest also will need to deal with both file delivery and ingest of HD and SD content.
The architecture of the underlying real-time engine and machine control is often quite different. Many systems, especially long-established ones, operate over RS-422 control for playback devices and switchers. Increasingly the devices are network controllable, and many modern automation systems use TCP/IP control over Ethernet to greatly simplify system architecture. This lends itself well to wide-area connections, which can be quite useful in centralized broadcast operations. At least two vendors have developed local control engines that hold in local memory machine control commands issued by the central system. This can even include time synchronization via NTP over the same Ethernet connection, with time offset capabilities to account for system latencies.
In the future, automation, as well as traffic, will need to be able to issue commands to cover new classes of events that arise from the needs of DTV. The accuracy of events will be pushed to no worse than frame accuracy, which includes the generation of frame-accurate traffic logs. Without such capability the programming of encoders, multiplexers, and data synchronization for interactive services may well require a degree of accuracy not routinely achieved by currently available systems.
John Luff is senior vice president of business development for AZCAR. To reach him, visitwww.azcar.com.
Send questions and comments to:john_luff@primediabusiness.com
