Skip to main content

Local-Into-Digital Ad Splicing, Part Two

Last time, we examined the environment for splicing local advertisements into digital channels. This time, more about the mechanics of how local-into-digital ad splicing works.

A quick refresher: Opening digital networks to local-ad insertion is hugely important to the people in this industry who sell such spots, because finding new ways to grow advertising revenues matters. And, in large part, it just seems like too much of a natural digital progression to dismiss.

With every new digital-cable customer, local ads on digital-cable channels become increasingly attractive. Last year, cable's base of digital video customers crested past the 20 million mark. People continue to sign up. Many digital networks are niche-oriented; hobbyists tend to like to hear about new tools or services pertaining to their particular niche.

Brand new market

Put another way — if City Floral wants the drought-aware citizens of Denver to know that they can catch and store the rain from afternoon thundershowers in its new line of rain barrels, it should be able to buy itself a local ad slot on Home & Garden Television.

The market for local advertisements on digital cable channels is just now getting started. Technical trials are largely done. The equipment is ready, and it works. Now, it's a matter of waiting for digital program networks to start tagging shows with the marker bits that do what cue tones do in the analog advertising landscape.

For cable operators, there's a blessedly short to-do list when it comes to getting ready for local-into-digital advertising.

It starts with a phone call to incumbent ad-insertion vendors to make sure their servers support the Society of Cable & Telecommunications Engineers standards around "DPI," which stands for "Digital Program Insertion." They'll know what you're talking about.

Need a splicer

Next, it's time to start thinking about equipment. Most implementations — and especially those in which digital "channel-grooming" gear is already in place to customize the digital-cable lineup — needs but one thing: A "splicer," made by companies like BigBand Networks Inc., Terayon Communication Systems Inc. or others.

A splicer works by watching the river of MPEG-2 video bits that blast in from the satellite. When it sees a "digital cue tone," it starts its business of splicing in the local ad, housed on the video server.

The "cue tone," which is actually a short set of packets that identify the splice point for a local ad, comes directly from digital programming networks. In some cases, cue messages get inserted by digital-channel aggregators, like the HITS (Headend In The Sky) platform, although that effort is still congealing.

The rivers of digital-video bits watched by the splicer contain three reference points, according to the MPEG-2 video compression method. Roughly two "initialization frames," or "i-frames," happen every second. Around them are "predictive" and "bidirectional" frames — "p-frames" and "b-frames," in the lingo. Combined, they carry the information that display digital pictures on the TV screen.

Ad server role

So there sits the splicer, watching the bits. In time — usually a few times per hour — it sees a cue message, indicating an upcoming splice point. The splice point is usually coincident with an upcoming i-frame in the stream.

The ad server is a big part of the brains in the mix. It always knows what batch of ads is next, for all involved channels. It keeps the splicer updated on which "PIDs," or "packet identifiers," correspond with which ads to play, on which digital channels, each time a local ad avail occurs.

(The term "PID" also comes from MPEG-2 lingo. It's like a nametag for each piece of digital content. Everything that moves over MPEG-2's downstream, headend-to-home transportation mechanism has a PID — even packets from cable modems.)

When the time comes for the local ad to be inserted into the digital bit stream, the splicer cuts in, and links the pre-established PIDs from the ad server into the ongoing digital bit stream for that channel.

The hardest part about splicing, aficionados say, is finding the "splice out" point. I-frames only happen every 15 or so frames of digitized video. When one isn't around — or when it's already happened — splicers have to get creative.

Smoothing it out

Plus, consider what happens when the fourth ad of four comes in at 29.7 seconds, instead of 30 seconds. This is where the secret sauce of the various splicer makers comes into play. Sometimes it means adding a few frames of black, to compensate for the missing time. Other times, it means fabricating an i-frame, just to keep things moving.

The people who nurture the relationships between local advertisers and cable systems put DPI and "digital-into-digital" in the "pretty big deal" category. So does the Cabletelevision Advertising Bureau, which two weeks ago dedicated the first afternoon of its annual gathering to the subject. So, as bright, sparkly new things go, this one looks pretty bright, sparkly and new.

Questions? Suggestions? Contact Leslie Ellis at