Engage your viewers with new types of content delivery.
There are many ways to share video online, and Flash has become a dominant player over the past few years. Its ubiquity in the market and flexibility in allowing content creators to customize the viewing experience make it a popular, easy option.
Silverlight, Microsoft's response to Flash, appeared almost two years ago to challenge the incumbent. Lucky for us, this has led to a flurry of innovation in a short period of time that we may not have been privy to otherwise. It is important to understand and discuss the difference between these two rich, interactive applications so that end users can deliver their content in an engaging experience.
Flash vs. Silverlight
Silverlight 2.0, currently in beta with a scheduled release of October, delivers VC-1 video as did Silverlight 1.0, but adds a great deal of interactive elements. These include Dynamic Language Runtime (DLR), a way of running compiled scripts such as Java, Python and Ruby from Silverlight (further increasing the options available to developers). It also features DeepZoom, a technology for panning and zooming around highly detailed online images. For example, an artist recently published a site containing a 10,000 × 10,000 pixel image of Barack Obama that viewers can interact with.
Meanwhile, Flash 9.0.115 and later versions offer the ability to playback H.264 video in addition to the already supported On2 VP6 and Sorenson Spark. This is significant because it is seen as heralding both a dramatic improvement in quality and the adoption of a popular new codec.
See Figure 1 for a head-to-head comparison. As it stands now, both platforms deliver live and on-demand video to the end user in a rich, interactive interface. However, they both have different applications when it comes to emerging technologies. Much of the functionality and high quality that Flash touts is tied to its ability to deliver on-demand assets. Sites using live video for Flash currently use the older codecs. Live H.264 encoders designed to integrate with Flash are only now becoming available. In this respect, Silverlight has an edge in being able to provide higher quality live streams, which it achieves by taking advantage of the VC-1 and Windows Media Server infrastructure — something that has made it a popular choice with sports venues such as the Beijing Olympics.
So what do these two platforms mean for live streaming on the Web? How important is quality, and what other features do end users expect? Perhaps most importantly, what does a broadcaster/distributor need in order to deliver streaming content?
Live online video
While the majority of video consumed online is on-demand, much of the money is made with live, real-time video transmission. Live streaming video has been on the Web now for more than a decade, but the quality and technology was lacking to the point of being unwatchable for all but the most frivolous types of content. Today, due to the convergence of better hardware and software, faster Internet connections and, perhaps most importantly, a more Web-savvy audience, live Web video actually stands a chance of becoming a serious draw to viewers.
This mood is reflected in traditional broadcasters' experimentation with online delivery. In the download-to-watch realm, the iTunes Music Store offers music, television shows and movies (the movies available both as purchase and rental options), while Amazon's Unbox similarly offers TV and movies for sale and rent. ABC, CBS, FOX and NBC's main Web sites all offer full-length, ad-supported versions of their prime-time shows. Hulu, a joint endeavor of News Corp (FOX) and Universal (NBC), offers both current shows and a healthy back catalog with limited commercial interruption.
The great part about the Web is that not all video has to be live. The majority of the traditional television viewing experience is delivered in advance to the networks that then transmit to local broadcasters. Content can be distributed via IP the same way, and users choose whether to cache and watch later or watch it immediately as it downloads.
What's been lacking is live Internet video — communal events such as sports, breaking news and other live entertainment — that connects people as easily and effectively as does today's television. Beyond merely matching the television viewing experience, the real promise of Internet TV is to extend passive viewership into actual interactivity. This means we must not just duplicate today's broadcast on the Web, but rather broadcasters must do more to be considered successful. The reliability of live Internet video must be on par with broadcast, but the experience of viewing online has to be richer in order for large numbers of viewers to embrace this new model.
Secrets about your viewers
Want to know a dirty little secret about viewers? They really don't know or care about things like format, resolutions or infrastructure. Just as most people don't care how their car works as long as it does work, it's the same with TV today. People just want to push the buttons on the remote and pick shows. Streaming video has to be just as easy to use.
Want to know another secret about viewers? Quality isn't that big of a deal. Don't get me wrong — if the picture is garbage, viewers will notice, but even cable and satellite today aren't perfect. Cable providers, for example, have repeatedly come under fire for over-compressing HD channels to save bandwidth. Furthermore, it's not the quality in a snapshot that's the issue; it's the reliability of the service over time. Content that looks mediocre, but works 100 percent of the time, is probably held in higher esteem by viewers than perfect content that only works 60 percent of the time.
Choosing the solution for broadcast television
Traditional broadcasters recognize that the way content is delivered has changed significantly, and they are not standing still. They see the writing on the wall and are actively pursuing ways to move their content online to match viewers' expectations. But scaling a massive infrastructure isn't easy. In addition to simply moving the content online, picking the right formats and technologies can be daunting. How helpful is it to move your content online only to realize that the format you chose is virtually unsupported, which then leaves your audience frustrated?
A broadcaster needs encoding equipment that:
will easily merge with an existing infrastructure instead of replacing it;
is adaptable, so changing resolutions and formats is easy;
is hardened and reliable 24/7; and
supports data insertion and injection of additional material into the stream.
That may not seem like a long list, but to a broadcaster with hundreds of hours of content per week to manage, it's a nontrivial workflow issue. The wrong decision can cut productivity, or worse, orphan content.
These are the kinds of challenges that Major League Baseball Advanced Media (MLBAM) faced late last year when upgrading its infrastructure. Providing a better quality video stream was a key goal of the upgrade. Improving internal workflows to provide faster VOD assets and extending the interactive capabilities of the site were also important objectives. The infrastructure upgrade resulted in streams increasing from 700kb/s to 1.2Mb/s for premium broadband subscribers. It also yielded functionality like the Major League Baseball mosaic so viewers can watch multiple games simultaneously.
Although it was a complex upgrade and massive content launch, the few bumps early in the season were quickly resolved, and MLB has successfully provided thousands of hours of live content to Web and mobile subscribers in multiple resolutions. It has also offered a host of game-relevant statistics and data in the viewer's player.
All this was achieved while managing rights and blackout restrictions and generating game highlights in near real time. MLBAM took advantage of Silverlight as its interactive platform, which allowed it to add support to Mac subscribers left out in the past by PC-specific digital rights management (DRM).
What does all of this mean for modern streaming equipment to support these two platforms? It means having the right ingest options to support the formats that traditional media uses to move content around. The only way to plug into the plant is to be part of the plant. This means SDI (both HD and SD) for some places, but it also means having ASI and IP options. Decoding to baseband video just isn't practical in a plant handling hundreds of programs daily. It's essential that solutions match the infrastructure.
It means being able to scale easily with robust, automated management. Broadcasters don't have just one or two feeds to set up and monitor. Most will have a minimum of 10 to 20 feeds, and a cable headend might have 200 or more. Such an infrastructure requires management software and monitoring mechanisms, automated systems that help keep the video reliably streaming without a lot of human intervention over each channel.
There is another side to scaling that broadcasters must not lose sight of — the viewers. This is directly related to reliability. Keeping feeds up and working is vital to ensuring that viewers are happy and coming back for more. Detecting issues and resolving them rapidly is required.
In the past, live video on the Web has been reliable to a certain point, but once a threshold of viewers has connected, service and reliability plummet. Such scaling issues cannot bottleneck delivery in a future-looking IP broadcast facility. While this is less the province of the encoder and more of the servers and delivery network, no vendor or service provider has the luxury of pointing fingers. The infrastructure must provide a reliable, stable solution from beginning to end. It should allow for a quality of service that's acceptable to both the broadcasters and viewers. Only then will adoption move forward.
Finally, flexibility of the encoder to support an evolving range of formats and resolutions is key. Unlike today's broadcast, where these are fixed and known values, in modern IP broadcasting, formats and resolutions are in near constant flux as the bandwidth and computing power increase. There may be a time in the future when these again become fairly fixed, but for now, they are a moving target as the system is optimized to improve both the quality of the image and the reliability. Having an encoding infrastructure that can change easily is essential in such a fluid environment. It should support multiple modern codecs, output options and resolutions so broadcasters can experiment and improve their architecture without the dreaded forklift.
A great deal has been made about Silverlight and Flash as competitors for online eyeballs. In the world of modern IP streaming, both are worthy solutions for delivering an engaging live or on-demand experience. Feeding video to these players takes a balance of new media savvy and traditional media durability. When it comes to the role of the encoders, the successful solutions will strike the right balance between format resolutions, manageability and reliability.
Andy Beach is director of product management for Inlet Technologies.
| ||Flash ||Silverlight |
|Compatibility ||PC, Mac, Linux and Solaris ||PC, Mac and Linux |
|Ubiquity ||High ||New, but growing quickly |
|Codecs ||Spark, VP6 and H.264 ||VC-1 |
Figure 1. A comparison of Flash and Silverlight