DTV: A Manufacturer's Perspective, Part 1 — Sony's Peter Dare on Digital Television's Future

by Peter A. Dare, Sony Corporation
Monday, March 30, 1998

The following is from a presentation given by Sony's Peter Dare late in February at the IBC DTV Conference in Las Vegas. "It is just as relevant today," offers Dare in an Email to us on 3/30/98. This piece sums up for most among the professional product producers their current view regarding DTV.

Dale Cripps
HDTV News Online

__________________________

Since the FCC's decision to endorse the ATSC standard [Sans table 3] there has been more written, discussed, and debated about the implementation of Digital Television, than perhaps any other subject. It's just a pity that well over 50% of what has been presented as fact is flawed in some way, thus creating even more confusions. I am hopeful that this presentation will not be just another review of the standards, or another interlace versus progressive tutorial, rather I plan to deal with issues that have yet to be dealt with, particularly from a manufacturers perspective.

Digital Television [DTV] is not new. In some market segments, such as found in the Post Production communities, have been "digitized" for well over I Years, at least in terms of Standard Definition Television [0]. High Definition Television [HDTV] has been a hybrid of analog and digital technologies.

DTV has now taken on a broader meaning. DTV means different things to different people—digital production, digital contribution/distribution, digital transmission digital SDTV, digital HDTV, digital transmission. It is in the world of digital transmission that seems to have had the effect of requiring the industry to step back and look at the infrastructure that might support such a service. Needless to say that compression technology has played a major role in deployment of many technologies that are being planned, and in fact without this technology the DTV service could not have been even considered.

1. Once we have established that the transmission pipeline is just digital "bits" or digital data, the horizon for applications beyond just television becomes possible. All forms of data can be transmitted, subject, perhaps, to some political influence. Much of the US industry is still searching for a financial model that makes sense—data broadcasting, multi channel SDTV, single channel HDTV. The DTV pipe line is not a single dedicated service. It is Configurarble on a daily, hourly basis. We have to learn to harness and live with this new flexibility. This is not a matter for discussion or debate. The fact is we are going to have DTV transmission What fills the digital pipe is a marketer's dream. If the broadcast industry can not figure it out, others will

Finally, the views being expressed by the broadcast community, and their concerns over financial viability, cannot be the reason for any delay. The DTV service should be viewed as a new start with many new technologies. DTV is not just the digitizing of the Analog plant. If it is viewed this way it will fail.

DTV-What is it ?

In it's simplistic form DTV is a digital pipeline into the home. It is capable of delivering 19.2Mb/s of user payload [In it's terrestrial form] with the TV station deciding how these digital bits will be used. In some cases, the TV station may elect to allocate all the bits for one or two HDTV channels, or allocate the bits to many SDTV channels. Tbe new exciting option is that some or all of these bits can be used for Data Broadcasting. To explore all opportunities there is the capability to have a mix of all of the above applications running simultaneously. Manufacturers of transmission and reception equipment are busy working on software and hardware solutions to fully explore the flexibility offered by this digital bit stream. We are, however, working somewhat in the dark as the broadcasting industry struggles with financial models and the new opportunities that are possible.

The broadcaster may well be in the business of selling bits for a price. Can you imagine that CD quality music can be downloaded in non-real time while normal TV programs are being transmitted? The latest version of computer games can be downloaded during the news. There are these and many other options that have yet to be fully explored.

Having a programmable digital pipeline is a new concept to the broadcaster, who will simply become a digital service provider. In the early days of the DTV service it is likely that the survive will be unidirectional. In the future there will be interactive and transactional services via a return path via the telephone company.

TV set manufacturers and computer manufacturers are working together to creat a plug and play connection (despite what may be reported from time to time) with greater simplicity than exists today in the home theatre environment That is at least the aim

Many new operations will exist within the TV plant. The role of the scheduling function, billing function etc., will need to be expanded to handle the new data services. By NAB 1998 manufacturers will have multiplex management software to address what we think are the needs of the broadcast community, I am not sure that a clear picture exists on the service providers side, perhaps this can be considered a call for user requirements to be generated.

DTV-FORMAT BATTLES

There will be forever discussion over which format is best. Is progressive better than Interlace? Which sampling strategy is best? The internet is full of opinion, much of it uninformed and even when accurate it is just an opinion, as is mine The choice of format will be made by the end user, NOT the manufacturer. It is as simple as that.

In the digital world there is likely to be many formats, bit rates, compression schemes, and all focused on a given application. For example: The bit rates and compression schemes used in production will be optimized for mutigeneration and layering work, where there is a requirement for no or few artifacts. In the world of the small broadcaster there is typically only the need for forward-and-store, or limited picture processing.

Surrounding any digital format there are issues to be dealt with, typically the greater the number of bits the higher the cost to record or store. Having said that, there is an exception and that is through the use of tape streamers. The penalty that you pay is the real-time aspect of television. Low cost tape streamers may not be able to support real-time transfers.

On the other hand, streamers don't know about the TV raster formats. They just record data. There is the evolving market where video is a file that is not raster based as is the case with "traditional TV". This is not a new concept, rather it is becoming more ubiquitous in the production community, and may well find a similar application in broadcasting.

If the production format and it's characteristics are different from the distribution format clearly format-convertors of varying types and quality will be required. Again, this is not entirely new to the post production community. We have had 10-8 bit digital convertors around for some time. Even this conversion was not without it's complications. Many convertors did not do a good job and caused many lost hours of technical personnel investigating the why's and wherefore of the conversion process. 1 would be less than honest to say that we will not go through this process again. We certainly will.

System designers need to take into account in-plant design dealing with formats and conversions along with other issues related to a compressed plant—the most difficult being the differential latency of video and audio signal paths.

DTV-THE REGULATIONS

In recent times there seems to be a difference of opinion that exists between what some broadcasters want to do with their digital bit stream and the politics surrounding the amount of mandated services. In addition to the percentage of "HDTV" that should be transmitted there is also debate over what is "HDTV." In some cases it could be simply up-converted 525 lines. Frankly, it can look HDTV -like. Without getting into a never ending debate it is likely to evolve into a mixture of SDTV and HDTV services driven by marketplace acceptance and implementation costs. What has been mandated is that the structure of the digital pipeline will be the ATSC standard without the defined table of sampling structure recommended by the ATSC. It is expected that the industry will adopt Table 3 as a voluntary standard with it's various Digital sampling options.

This voluntary implementation of the HDTV sampling structure is the one issue that is causing an on-going debate. One network [CBS] announced support for the 192O X 1080i format while others have delayed making a decision as they explore the available options. There is little question that the SDTV format (based on SMPTE 125 sampling) will be the format of choice for the SDTV service. It is not quite as clear if such a universal choice will be made for HDTV. That leaves the industry with a multitude of conversion problems with which to deal. The reality is that the manufacturers will make what the industry wants and will pay for.

The problem is there needs to be interoperability between the possible many HDTV formats, otherwise the results may be no better than today's 525 line service. There also needs to be affordable equipment available, otherwise the implementation costs will be beyond the reach of many stations. The DTV equipment costs are not just related to the decisions within the USA. It is a world wide influence that will determine equipment costs. The SMPTE has as one of it's standards a sampling structure 192O x lO8O. Likewise the ITU and ARIB also have the same common image format, at least for production. DTV introduces the concept that the production format may differ from the transmission format. Although possible, it is not without it's complications. In addition, the receiver may not display the transmitted format. There may be internal conversion within the TV set

Choice of the sampling strategy is a users choice. It is their pocket books that have to pay for this conversion. Manufacturers have the marketing task of promoting the equipment available from that supplier to suit the requirements of the end user, and only if the requirement is known. For a number of years Sony has made available equipment to the 1920 x 1080 [actually 1035 on early models] standard. In recent years we have expanded that range along with drastic price reductions along with increased functionality.

The immediate challenge facing the industry is a mandate from the FCC that requires conversion to DTV over the next 9 years—a Monmouth undertaking considering the very complex infrastructure. Filling the 19.2Mb/s pipeline on a consistent bases will be a further challenge, in particular beginning the business of Data broadcasting. Many issues remain unresolved in the area of data broadcasting. The one of interest to this manufacturer is the interfaces and API's that need to exist in the broadcasters' plant and in the receiver. This manufacturer believes the receiver operating system should be open and NOT tied to a single manufacturer. The application should run on all operating systems and always have forward and reverse compatibility. The idea that you have to buy a new TV set every two or three tears is not a model that we should be targeting.

Having a focus on data broadcasting alone may be too narrow, may be what is required is the soft ware necessary to manage the pipeline with links to various parts of the plant and the generation of the play list. An equally important issue relate to revenues and revenue sharing needs to be defined rather than evolving. Manufacturers need to know the type of software management systems to put in place. It is clear that these systems should view the 19.2Mb/s pipeline as a revenue stream In addition, management of the pipe line needs to address conditional access and the other system management issues defined by ATSC or DVB. It should be noted that the system information issues differ significantly between DVB and ATSC, with ATSC having the greater flexibility.