Digital Program Insertion

To generate a positive return on the digital program insertion (DPI) investment, the systems must be standards-based, cost effective and robust. To grow revenue through advanced localization and personalization, the systems must be modular and flexible. To maintain customer satisfaction, the systems must provide the highest quality of both the programs and the advertising.

One of the most important aspects to the video delivery business is the correct and effective placement of advertising, whether at the national or local level. Analog or baseband switching solutions for ad insertion are well established in today’s video architecture. They enable programmers and operators to insert programming and advertising in specific time-slots. Local advertising is a key component of any operator’s business model. As programmers and operators aggressively transition from analog to digital, they must be able to not only maintain, but grow that advertising revenue. To compete for a larger share of the spot-advertising market, the ability to serve ads based on specific zones is a key competitive advantage. For these reasons, digital program insertion (DPI) has quickly become an important part of the landscape. It is vital that operators have the same program insertion functionality on the digital programs that they now employ with analog programs. For the new all-digital programmers, operators and service providers, DPI is a must-have requirement if they hope to capture some of the revenue that flows downstream from Madison Avenue.

As with any new technology, the standardization process is important to ensure successful interoperability and encourage competition. In order to successfully place and track ads, the systems that insert them must be standards-based and fully tested for correct operation. There are three primary standards relevant to DPI. SCTE104, the most recently defined, describes the message format that triggers a digital video encoder to insert a splice-point in the MPEG transport stream indicating where an avail is to occur, how long the avail is to last and any other information relevant to that specific insertion opportunity.

The encoder inserts the splice-point in the stream according to the second standard, SCTE35. The SCTE35 message in the transport stream indicates to the splicer where to insert the program or advertisement. The splicer receives the SCTE35 message and communicates with the content server according to the third standard, SCTE30.

The content server serves the program or ad to the splicer, the splicer inserts the program or ad at the appropriate time and the successfully inserted program or advertisement leaves the splicer for distribution.

There are a number of factors to consider in designing a DPI solution but the two most important are the location of the ad-insertion equipment and the architecture of the MPEG encoding equipment.

Ad insertion can be either centralized or distributed. There are two advantages to a centralized approach: all equipment is within reach (more readily manageable) and the cost is less as the functionality is implemented in one location and used across multiple regions or zones. The disadvantage of centralized architectures is that they are relatively inflexible with regard to serving different content or ads to different regions or zones. It is possible to design a centralized approach and serve different content to different zones, but this approach necessitates duplicating those programs with avails across two or more transmission channels. Although bandwidth intensive, this duplicative approach is a viable option for those operators without significant bandwidth constraints.

Distributed architecture provides operators with full flexibility and control at the regional or zone level without having to duplicate channels. One channel with SCTE35 splice points can be sent to unlimited zones, each zone inserting its own specific program or advertisement. This flexibility allows for a more targeted advertising campaign, a higher CPM rate and subsequently greater total advertising revenue. With the help of new advanced set-top boxes, operators may one day extend the distributed architecture into the living room, targeting advertising to individual consumers. This degree of customization and personalization is not readily achievable in a centralized approach. The disadvantage of the distributed approach is that even though the storage may be centralized, the costs of splicing and insertion are duplicated at each region or zone.

The encoding solution is equally significant in designing the DPI solution. There are two approaches: constant bit rate (CBR) and variable bit rate (VBR)/closed-loop. Each architecture has its own benefits and drawbacks. The considerations for choosing one over the other are larger than those relevant to DPI and include total network efficiency, signal quality, QAM management and system cost. Paramount to the success of a digital programming and advertising operation is the best possible quality for both the programs and the advertisements.

CBR is often considered a conservative approach to digital conversion because key parameters are predetermined. The number of encoders, the bit rate per service, the number of services per QAM and VOD encoding rates are known at the outset and will not change without operator intervention. For this reason, CBR architectures are optimal for systems desiring less complex operations.

CBR architectures are also optimal for switched delivery architectures such as Telco/IPTV or cable switched broadcast. The benefit of the CBR architecture for DPI is that programs or ads to be inserted can be coded and stored at the same rate as the program for which they are bound. It is therefore unnecessary to rate-shape the program or ad to make it fit in the avail slot. This ensures that the quality of the ad is both constant and known a priori. Since the rate does not change, the quality of the viewing experience is consistent from program to ad and back. The drawback of the CBR architecture is bandwidth efficiency. VBR or closed-loop architectures are 20-30% more bandwidth efficient, allowing more services to be delivered in a given amount of bandwidth.

VBR and closed-loop are different from CBR in that each encoder varies its rate in accordance with the complexity of the content being encoded. While closed-loop is more bandwidth efficient than open-loop, or “native” VBR, the two are nearly identical in their relevance to DPI. While these approaches allow you to pack more programs into a given amount of bandwidth, they can add additional complexity to the DPI design. If advertising is inserted in a remote location, there is no guarantee that the ad being inserted will match the bit rate that was left available by the original encode. This “rate mismatch” necessitates the use of rate-shaping in the ad-splicer, introducing a second generation of MPEG encoding and negating the value in the original closed-loop encoding solution.

To avoid rate-shaping, the closed-loop encoding system can be set to generate a multiplex with a total video bit rate that is lower than the total channel bandwidth. Thus, when ads with higher bit rates are inserted into the multiplex stream, the total bit rate does not exceed the total channel bandwidth. This conservative approach supports ad insertion without rate-shaping, but offers reduced capacity since the total channel bandwidth is not utilized when there is no ad-insertion. Overall, this approach results in a general underutilization of bandwidth. Historically, “rate mismatch” and “under-building” have been limitations of closed-loop systems. However, ETG’s new encoders built on programmable architectures are now offering RateLock solutions to manage distributed ad-insertion models for closed-loop applications.

Chris Gordon is Director of Product Development for EGT, Inc. He can be reached at