Even With New Tools, Audio Up and Down Mixing Has Highs and Lows

Increasingly, the sound of sports broadcasts has become as fragmented as the Balkans once looked. Viewers are sometimes on a roller coaster ride that goes from a 5.1 surround experience (discrete or matrixed) to the sudden deceleration of the bump to stereo for commercial breaks or other intermissions. While this pas de deux between surround and stereo contributes to some extent to the loudness issue, perceptually or otherwise, it’s surprising that it happens at all, given the growing number of algorithm-based upmix/downmix propositions out there. But as good as they’ve gotten, they can’t resolve all of the complex issues involved.

The abrupt transition from surround to stereo during broadcasts is like nails on a blackboard to Bob Dixon, director of sound design and communications for the Olympics at NBC Universal. But, he says, millions of dollars are spent by advertisers making commercials in stereo, “And it probably should not be up to the broadcaster to change their carefully crafted stereo to an upmixed version,” he says.

The Beijing Olympics time differential (and even the differential for the Vancouver Olympics) underscored the fact that much of the broadcast had to be time shifted. Edited pieces are usually done in stereo, however, and the energy in the announce portion can affect the steering of the upmix in such a way that it narrows the soundfield, creating artifacts that collapse the mix towards the center when the announcers are talking, crowding that channel with additional audio elements. In addition, says Dixon, “When a complete show is run through an upmixer, there can often be artifacts with the announcers’ voices. For example, bits of the announce mix might come out of one or both of the rear speakers sporadically.”

Thus, Dixon requests that the announce track and stereo effects tracks be delivered as separate mixes. Downmixes used by the editors, on the other hand, are performed manually, inside the console for the most part, by the mixers rather than by an algorithm to assure that they will have the best-sounding and -balanced material to work with. “The reality check happens when we send the discrete 5.1 mix to [NBC in] New York. Then, at the same time, we listen to a down mix done by the [Linear Acoustic] Aero.qc using the same algorithm used by television sets and set-top boxes, so we get to hear exactly what our stereo customers will be getting,” Dixon explains.

Tim Carroll, president of Linear Acoustic, acknowledges that upmixing “made a bad name for itself in the early days.” It’s a reputation that upmixing products no longer warrant, he says, noting that the company’s new UPmax II does not allow the center-channel information to “steer” the rest of the upmix, changing the balance of the announcer with the effects sound. “But we also learned from Beijing and Vancouver that the mixer’s technique also has a lot to do with the outcome, too,” he adds.

Jim Starzynski, principal engineer and audio architect at NBC Universal’s Advanced Engineering, says the network encourages content suppliers to provide their audio deliverables with 5.1 sound, and that more of it is coming in that way, including some commercials. “We encourage advertisers to take advantage of the 5.1 format, but the majority are still provided in stereo,” he says. Starzynski adds that policy calls for upmixing is automatically applied to all stereo programs and commercials on NBCU HD cable channels, including SyFy, USA Network, Bravo, CNBC, MSNBC and Universal HD, but that the NBC network does not apply upmixing to any content at network distribution, the distribution channel for sports broadcasting.

However, the issue for now remains confined to those consumers who can play back a 5.1 soundtrack. “For the stereo-only DTV audience, there is no sonic difference between stereo and 5.1 content with the consideration that for 5.1 deliverables,” Starzynski says. “The stereo audience hears an ATSC downmix of the supplied 5.1.”

The policy is similar at CBS Sports, where every live remote is supplied in 5.1, including legacy audio elements, which are upmixed before they leave the truck or studio, via a Dolby DP654 encoder. “We want to keep the metadata consistent at 5.1,” says Bruce Goldfeder, director of engineering for CBS’ sports division. However, if prerecorded programming is delivered in stereo, it will often stay that way, with channels 3 through 8 blanked to avoid inadvertently triggering an upmix in viewers’ set-top boxes.

At Fox Sports, all location audio is run through a Neural Audio MultiMerge that takes the 5.1 content and passes it directly through. It will detect a two-channel (stereo) source, upmixing that to a 5.1 signal. According to the network’s senior audio mixer and consultant Fred Aldous, Fox no longer sends out an SD or stereo mix. “All programming is HD and 5.1 whether it is upmixed or legacy 5.1,” he says. Steven Silva, Fox’s director of procedures & training, concurs, noting that Fox’s policy calls for commercials and other interstitials to be upmixed at the network level, using the Dolby 564, and with loudness set at 24 LKFS.

The ultimate goal is a completely 5.1 source from broadcasters to viewers, and it’s within sight. Till then, though, the road cannot be completely smooth, “Though present delivery of 5.1 and stereo content is a necessity, this practice does move the focus of the sound as the audio format changes,” Starzynski acknowledges. But someday the only issue will be the last 10 feet – from the speakers to the couch.

Password must contain the following:

A lowercase letter

A capital (uppercase) letter

A number

Minimum 8 characters