Sound and fury
Flip through a typical digital TV lineup and you will find plenty of evidence of a vexing technical problem cable operators face—on one channel you have to punch up the volume just to hear dialogue, and when you switch to another channel, you are greeted by an ear-splitting din.
The central problem is inconsistent sound level encoding for digital video among programmers. While there are standards available and new audio tools that could tame the noise, getting the industry to adopt them is another matter.
Scattershot audio levels have been a perennial issue for cable programming, but recently, the problems have been growing with the advent of new video services such as high-definition television (HDTV) and video-on-demand (VOD), according to Jeff Riedmiller, broadcast product manager for Dolby Digital.
"A lot of people think if it comes from Hollywood, it's perfect. We right now are halfway through the analysis, and we are seeing a 15 dB to 16 dB spread in loudness among those programs," Riedmiller says. "And a lot of that stuff ends up on VOD systems and things like that."
There are guidelines. Many programmers in their carriage deals with cable operators provide audio delivery requirements for overall loudness and dynamic range, and they can also look to standards set by the Advanced Television Systems Committee (ATSC). In particular, ATSC has a longstanding specification for dialogue normalization, providing a common benchmark for how loud audio conversations should be in relation to background music and other sounds.
But once the digital video reaches the headend, cable operators have surprisingly few tools to make any changes, and some of that has to do with their digital aggregation strategy, says JC Morizur, director of product marketing at Harmonic Inc. With the advent of digital, cable operators opted for a transcoding system, which simply takes the incoming streams and fits them on to any combination of QAM channels.
"On the one hand it is very price effective because you don't use a decoder and an encoder," he notes. "On the downside, they no longer have the ability to adjust the audio level."
DBS providers don't have the same problem because they re-encode programming channels at the aggregation point before passing them over the satellite link, Morizur notes.Ad volume
It's not just the programming. Commercials have always been a source of customer complaints regarding audio, and the culprits often are advertisers that intentionally set the noise level high to literally shout out their pitches. Riedmiller notes the problem is compounded by the fact that a spot will play differently on two networks, depending on their individual sound levels.
"You end up chasing your tail when creating a specific spot," Riedmiller says. "You shouldn't have to encode that spot three or four times and make sure it gets routed to the right networks."
To that end, Dolby is now working with the Society of Cable Telecommunications Engineers' (SCTE's) digital program insertion committee to come up with more uniform sound levels.
Cable operators are aware that not all sound is the same when it comes to their channel lineup. Premium cable networks tend to produce programs with high-end audio that has a wider dynamic range, while some news outlets opt for the narrower range to keep talking voices consistent, notes Bob Blackburn, director of advanced video engineering at Adelphia Communications. But knowing the problem doesn't necessarily lead to solving it.
"There are things you can do in the headend by using audio power meters to try to equalize that and try to come up with average power level settings for the system. But it's still very difficult to deal with," Blackburn says, noting that the MSO has implemented controls in the set-top that users can select depending on their environments and preferences.Boxes play a role
The set-top itself plays a key role in audio quality for digital video. For example, in Los Angeles, where Adelphia is rolling out boxes powered by Digeo Inc.'s Moxi multimedia software, Blackburn has been hearing about differences when customers switch to the new set-tops compared to their original Motorola boxes.
"The way that different vendors implemented the dialogue normalization levels through Dolby—because it is all Dolby down at the basic level—is different among the vendors," Blackburn says. "I've been hearing people complaining that the Moxi product is too soft. So we're actually trying to address that with Digeo folks right now to see if we can get a little bit more commonality between the Motorola product and the Moxi product."
Riedmiller also sees the set-top as a key factor that is often overlooked.
"If you are going to go and set all of your analog levels for your analog tier of services, you have to understand what level the set-top expects in order to get the digital services to play nicely or even play. And a lot of them (MSOs) don't know that," he says.Taming the cacophony
There are other tools available or due for release that may bring some order to the sound chaos. Real-time content such as news, regular TV programming and events must rely on the programmer to set the proper audio levels, but diagnostic tools like Dolby's LM100 can analyze the incoming content and at least flag it when it is outside sound level boundaries.
Since its debut about two years ago, Dolby's LM100 analysis unit has found a home in about 110 broadcast facilities worldwide, ranging from the BBC in the U.K., to MSOs such as Time Warner, Charter and Comcast Cable in the U.S.
For on-demand content delivered to servers, Dolby is working on tools that can go in and determine if the audio levels are correct, and potentially correct the files.
"So it always plays with the right volume—it always plays with the right metadata," Riedmiller notes. "We actually have some tools we have been developing and working very, very hard at testing right now with some of the MSOs and some of the VOD equipment manufacturers."
But such products could make programmers nervous because it could alter the audio and, therefore, the artistic impact of the program. That's a major reason Adelphia passes through digital video untouched, Blackburn notes.
But Riedmiller stresses the Dolby tools would conform to the ATSC standards and would not change the aesthetics of the audio—only the volume.
"We're really not changing the content—we're really just changing the metadata that represents the content to the decoder. So we really are not changing any artistic portion itself," he says. "We're just changing first and foremost the label that's carried as part of the metadata in the bitstream to the decoder in the set-top—we are changing the label to make it more appropriately represent the loudness of the stream. So the decoder can adjust the volume directly."
There also may be improvements as cable operators delve into simulcasting analog channels in digital. Morizur points out this will require operators to move away from their pass-through transcoding system and instead re-encode the content anyway.
"So we expect that moving forward they will have the ability to adjust the audio level across the board, and consistently, and get rid of the problem," he says.New codecs
In terms of overall sound quality, help may be on the horizon there as well through newer audio codecs including Dolby Digital Plus.
A proposed ATSC standard, Dolby Digital Plus offers improvements including extended bitrates available to the end user in the downstream and upstream, the latter of which may come into play down the road for applications including HD DVD and blue-ray DVD. It also widens the number of discreet output channels that can be used for more sophisticated surround-sound speaker systems.
The new Dolby Digital Plus encoder has just been put out for general availability, and it will be available soon, most likely in HD-DVD systems later this year. Dolby also is developing a real-time encoder using the new audio scheme that will likely hit the market this fall, and it is developing a transcoder that can convert older audio schemes into Dolby Digital Plus, Riedmiller says.
Dolby Digital Plus will probably pop up first in satellite set-top-box silicon that also uses the new advanced video codecs such as H. 264.
Cable technology players including set-top box makers are also looking at it, but "not much [has] really moved in that space," Riedmiller says. "I think they are waiting for SCTE and CableLabs."
Until more consistent encoding rules are adopted, however, it looks like digital cable customers themselves will play a big role in audio quality.
"It is something that customers have unfortunately gotten used to—some channels sound louder than others, and they adjust their controls," Blackburn says.