Multiformat Encoding

With the explosion in the technological ca- pabilities of streaming media, today’s service providers are certainly pitching their video capabilities to the home over high-performance DSL and cable lines. In addition, many public outlets offer stable WiFi access that can also transport streaming media. The cur- rent generation of cellular networks based on the 3GPP standard is geared to this market as well. Just as in the early years of Web video, device tech- nology has evolved to allow some- thing more than postage-stamp sized images playing back jerky motion, as demonstrated by Apple’s iPhone in- troduced last January.

Today, the myriad array of formats, networks and devices used in the de- livery of video and audio streams cre- ates a true challenge for broadcasters that want to reach these platforms. As advertising dollars shift to take ad- vantage of these mobile audiences, it makes sense for broadcasters to devel- op their infrastructure to reach them as well. Technically, there are several stages in the process that need to be addressed to maintain quality and compatibility with the current fi eld of delivery destinations. (See Figure 1.)

Playback devices

Sometimes it’s best to set an end point before you begin the process. So let’s fi rst explore the platforms that exist for online and mobile device playback of streaming media.

Desktop playback of video and audio still include Windows Media, Real and Quicktime, but Flash video has become the format of choice for many producers because of its prev- alence across computer platforms. Each has its own set of codecs, which players and browsers can decode on the receiving end. If not preinstalled on a system, one-step downloads and automatic updates make them acces- sible to the end user as well.

In the mobile domain, there is an even wider selection of platforms as this market is only in its early stages. The largest market share probably goes to Apple’s iPods and iPhones, which pushed the technology curve well into the future with its introduction last year. While many portable media players (PMPs) are designed for audio-one, Apple’s success is sure to see some competition in the coming years. Microsoft has already attempted to enter this market with its Zune player, which is also capable of playing back H.264 video and AAC audio as well as Windows Media (WMV and WMA) files. Other contenders include cell phone manufacturers, who are always pairing up with wireless carriers to provide new services and content.

In order to reach these platforms, we need to look at their network connectivity. Broadband access has grown tremendously, with several viable options for high-performance delivery over cable, phone and more recently fiber. Many consumers also want the flexibility to receive streaming content on their laptops via wireless LAN and WAN networks. This is best accomplished using one of the more recent incarnations of the 802.11 standard, which have good capacity compared with broadband wired networks, as indicated in Table 1. The most stringent bandwidth restrictions enter in the cellular domain, which generally relies on the 3GPP standard, a subset of MPEG-4, for delivery. As indicated in the table, these networks are less robust and highly variable given the mobile nature of the end user, so target bit rates are generally lower.

Encoding parameters

So, given these playback devices and platforms with various incarnations of IP network connectivity, what are the possibilities for encoding parameters to reach the multitudes? Today’s technology allows the portable media players like the iPhone and Zune to decode 320x240 frames at a solid 15fps and even 30fps. Most computer and laptop displays are much higher resolution, making this frame size much less than one-quarter of a screen. Larger content providers, more typically with longer form programs, are attempting to create 480x360 frames, or using built-in DSP technology to create a full-size video image on the desktop. Today’s CPUs and bus structures can handle that even at 30fps, allowing broadcasters to captivate even their discriminating audiences with broadcast quality on their computers. While this requires increasing the bit rate, most of the broadband networks can sustain well over 500Kb/s to offer a viable experience to the end user. On the other end of the spectrum, the cell phone devices are more likely to playback QCIF resolution images at 176x144 at 10fps to 15fps, reminding us again how early in the game this is for mobile video delivery.

Codec technology has also evolved to the point that many platforms for encoding can achieve this wide range of parameters sets. The lastest versions of Windows Media, Quicktime, Real and Flash can certainly accomplish this range of encoding requirements. They will also reach into the realms of DVD and HD encoding, as Windows Media’s VC-1 and the preferred codec in Flash, ON2’s VP7, have demonstrated recently. Of course, a single standard, such as H.264, most often used in Quicktime, would simplify the process for broadcasters, but technology developments have allowed many device manufacturers the flexibility to offer more than one format for delivery.

Choosing a format

Choosing one format over another for encoding may simplify the process, but most broadcasters will want to reach the largest audience by developing an infrastructure that can manage more than one platform and many different bit rates to accommodate varying network circumstances. Banks of dedicated CPUs, either dual-dual or full quad, will be required to process the target delivery formats in parallel.

Many vendors offering transcoder tools to run on these machines also offer a server/management application that can automate and balance the load across them when networked on a LAN. All of these devices will also be capable of ingesting a multitude of input formats, but the highest quality, progressively scanned source would be the most desirable for the IP delivery discussed in this article.

CATEGORIES