Figure 1. In EGT’s distributed closed-loop architecture, the exchange of complexity/bit-rate messages and the output media traffic can be transported over a network. Click here to see an enlarged diagram.
The primary goal of digital conversion is to improve the customer experience by increasing the quantity and quality of available programming. A secondary but important objective is to improve the flexibility and cost-effectiveness of the network.
Although a full digital conversion of the last mile is perhaps years away, the core networks connecting regional headends and hubs are transitioning to digital IP networks with centralized operation centers. In a centralized network, nearly all processing is performed centrally and distributed over IP networks with affordable edge decoding supporting the analog tier. Although this architecture has the advantage of centralizing control at a single location, there remains a requirement for distributed encoding and signal processing.
Distributed encoding
Not all content is available at centralized locations. Digital tier program line-ups are comprised of national, local broadcast and local public channels. While national channels are carried by all systems, local broadcast channels originate in different metropolitan areas. Local public channels have an even narrower viewing area and originate from multiple locations within a single system. This distributed nature of content availability requires that the encoding be distributed too.
Distributed encoding can be more bit-efficient. Encoding efficiency, characterized in terms of video quality, depends on many factors. Primary factors are the number of channels in the multiplex and the mix of channels chosen to form the multiplex. Distributed encoding offers complete flexibility for mixing local channels with national channels with different priorities to achieve optimum quality.
Digital ad insertion can also have an impact. Downstream or distributed ad insertion introduces an additional generation of re-encoding that can deteriorate video quality. By using local channels to reduce the number of ad-insertion channels per multiplex, distributed encoding offers additional flexibility for DPI load-balancing. However, the re-encoding issue still exists.
To fully optimize the network for distributed ad insertion, EGT has introduced DPI RateLock. DPI RateLock is a solution to the re-encoding problem that establishes a pre-defined bit rate for the advertising available and incorporates it into the encoding algorithm. Operators are then free to match the bit rate of the stored ad to the bit rate available, eliminating the re-encoding requirement and improving the quality and efficiency of the multiplex.
Open-loop vs. closed-loop architectures
There are two primary architectures for distributed encoding: open loop and closed loop. In distributed open-loop architectures, encoders are located in disparate locations and operate in an open-loop CBR or VBR mode. The outputs are transported to the aggregation location, and a statistically multiplexer via rate-shaping.
In open-loop architectures, there are at least two generations of MPEG re-encoding: the first one at the encoder and the second during the rate-shaping operation. This leads to a generational loss in video quality. In this architecture, it is necessary to ensure that the encoders and multiplexers are configured appropriately to minimize this loss.
This EGT encoder supports up to four channels of high-quality MPEG-2 encoding in 1RU and is configurable with either analog or digital interfaces.
Closed-loop architectures offer the most efficient compression performance and thus maximize the number of channels in a given bandwidth. In the past, closed-loop encoding has been synonymous with collocated encoding. However, EGT has determined that there is no underlying requirement that the encoders and the controller be collocated. The exchange of complexity/bit-rate messages and the output media traffic can be transported over a high-speed IP network, enabling distributed closed-loop architectures. (See Figure 1.)
EGT has integrated the closed-loop rate control function in the encoder, rather than in a separate piece of hardware. Apart from the obvious cost advantages, this architecture provides flexibility. And the software closed-loop control feature enables simplified network designs, reduced sparing inventory and improved cost-effectiveness of the video network.
Declining costs and increasing capabilities provide new opportunities for centralization. However the need for distributed encoding remains as a prerequisite for any digital conversion project because of the basic geographic availability of sources, the need for multiplex efficiency and the desire to optimize multiplexes for ad insertion.
Conclusion
Embedded closed-loop control and DPI RateLock are important tools for improving the cost-effectiveness of the video network. However, as networks migrate toward IP distribution, careful consideration must also be paid to the delay, jitter and loss characteristics of the IP network.
By combining intelligent system design with robust IP distribution infrastructures, operators will improve the efficiency and quality of their programming. They will also dramatically impact the operational efficiency and revenue potential of their systems.
Chris Gordon is director of product management and Santhana Krishnamachari is vice president of engineering for EGT.
