Like many other television outlets transitioning to digital transmission, KENS-DT was facing the dilemma of keeping audio properly distributed in our transmission feed. The San Antonio, TX, CBS TV affiliate station had to integrate regular stereo content with 5.1 surround material that was being transmitted in the CBS network feed.
The DTV station's technical configuration is fairly simple. The station has a pair of Harris HD IRDs provided by CBS, a pair of Dolby DP-572 Dolby E decoders and a Dolby DP-569 digital audio encoder, all ahead of a Harris Flexicoder and an Evertz 12 × 2 router as a DTV master control switcher. All of the DTV switching is done under Sundance automation control. Doing metadata switching between network and local programming was not so difficult. The problem was how to deal with 2.0 content that CBS itself would intersperse through the 5.1 surround audio in the program.
In order to solve the dissonance in the transition between the stereo and Dolby content, the engineering staff initially attempted to extract external metadata from the incoming signal and apply it to the final audio encoder, a DP-569. We connected a cable from the “metadata out” port on the DP-572 and connected it to the “metadata in” port on the DP-569. That worked for a while, until CBS became more sophisticated in its switching methods. After that update, if we persisted in using the metadata extraction method, the encoder would often fail to switch back to 5.1 surround from 2.0 stereo. That left us with nothing but the front surround channels on the air instead of the full 5.1 surround.
Another alternative to automatic metadata switching left the encoder in 3/2L mode all the time, causing dramatic spatial and level shifts whenever the program content changed between 5.1 surround material and 2.0 stereo commercials and local content. This was unacceptable.
The team also fired a GPI closure during stereo programming, switching the audio encoder into its 2.0 mode. This at least was better than allowing the broadcast audio content to be all over the landscape, so to speak. Although, it introduced another manual operation to an operator's already busy workload.
By mid-2005, the station's engineering team had no idea how to automatically transition between mixed audio formats. The thought of using mixers and surround audio detectors was not appealing and would have been a primitive solution at best. Plus there was no solution to the problem of how to reliably record HD content with 5.1 surround audio for delayed broadcast.
Then, Reed Wilson, an A. H. Belo technology manager, suggested a solution that was successful for WFAA-TV in Dallas. The Neural-THX Surround MultiMerge from Neural Audio is a way to automatically and seamlessly switch between 2.0 and 5.1 audio content.
The install and setup
My initial research on the MultiMerge suggested that it was a plug-and-play device. We would simply switch it on and place it in the digital audio stream between the multichannel output of our master control switcher and the input of the Dolby audio encoder. The unit would do the rest. We set up a range of inputs and outputs on our audio patch panel and awaited delivery of the device.
As with any digital device requiring synchronization between audio and video, it is important to provide AES reference to the Neural-THX. With a stable reference, you can be certain the lip-sync adjustments remain where you put them. The station team used a standard AES silence at 48KHz applied to the word clock input on the back of the unit.
The setup illustration in Figure 1 above shows the MMC-100 (MasterPlus) AES audio outputs as they pass through one jack field onto the audio processor and finally to the DP-569. For EAS events, the DP-569 is switched to 2.0 stereo operation through a GPI closure to pass EAS audio from the TFT 999, and it is switched back to a Neural 5.1 preset (which we created) once the EAS event concludes.
Now we switch multiple HD sources using a 12 × 2 router controlled by GPI from our automation system. And instead of audio from the MMC-100, we now have a final demultiplexer card that separates out three AES pairs for delivery to the unbalanced 75? Neural-THX inputs.
The front panel is clean and uncluttered, with simple status indicators, an LED display showing the type of audio input and a separate audio bar graph for each audio channel. The meters can be switched to show input or output.
Since installation, we have upgraded to the new Neural-THX MultiMerge + Neural Loudness Control (NLC). It has all-unbalanced AES audio inputs and outputs, simplifying cabling to and from our final audio demux and the encoder.
The initial on-air test
When we tested the system for the first time, we were already broadcasting a CBS network show in Dolby 5.1 surround. When we inserted the device into the program stream at the end of a commercial break, the audio processor immediately recognized the Dolby 5.1 audio stream and passed it through without alteration. It doesn't add coloration or change the 5.1 audio. It was completely transparent to the stream.
We eagerly awaited the next commercial break, which was a pod of four national ad spots in stereo delivered by the network and part of the program itself. The MultiMerge reported the input switch to 2.0, but seamlessly continued to deliver simulated 5.1 surround audio in the stream. The viewers did not detect that there had been a change, and surround channels actually reproduced convincing audio from the 2.0 stereo input. This capability was a definite plus in solving our audio problems.
When we transitioned from network programming to local material, the transition was equally seamless. The local stereo audio remained as full and rich as the network surround content had been, and the spatial distribution of all audio channels likewise remained consistent, delivering a pleasing listening experience. The upgraded model with NLC has level control capabilities to deliver audio with consistent quality and at the correct level.
With no particular adjustments to the final audio encoder, the first audio processor delivered an excellent performance just by virtue of its presence in the program stream. But to make the audio sound its best, we created a preset for the audio processor in the Dolby encoder, which we named Neural 5.1. This preset guarantees that the metadata settings remain consistent, which in turn assures seamless, transparent operation at all times. And since adding the version with NLC, the station team can create individual operating profiles based on a number of presets built into the unit.
The user manual is helpful for getting the desired presets loaded and modified to suit local needs. KENS knew it wanted to retain the open sensation of 5.1 surround audio without too much processing. However, the station wanted control over the sometimes aggressive loudness in a few local commercials. Through experimentation, the team reached a compromise that keeps the commercial blasters reined in while holding on to the dynamic range of the surround audio. Once the target dialnorm was set to -22dBFS, most of the -23 preset was very close to where we wanted it.
The station has experienced only minor issues with the MultiMerge and the MultiMerge +NLC upgrade since installation. (See Table 1.) Like most new equipment, it was necessary to upgrade the firmware to the latest version.
With the MultiMerge + NLC, the station has the best-sounding audio of any DTV station in the San Antonio market. Other stations in the market are forced to switch their audio between 5.1 and 2.0 on a regular basis. The spatial shift when listening to them as they switch is exactly what we avoid by using the this system. It is truly a plug-and-play device, requiring little maintenance or adjustment once configured.
Jerry Paonessa is assistant director of technology, broadcast media, for KENS-DT, a CBS affiliate in San Antonio, TX.
Table 1. A summary of the situation experienced by the engineering team at KENS-DT
|The needs ||Properly integrate local stereo content to a network's 5.1 feed |
|Properly deal with stereo content that the network intersperses through the 5.1 audio |
|Solve dissonance between stereo and 5.1 |
|Keep audio properly distributed in the final transmission feed |
|Deliver consistent audio quality to viewers |
|The solution ||Automatically and seamlessly switch between 2.0 and 5.1 audio content |
|Reside in the digital audio stream between the multichannel output of the master control switcher and the input of the Dolby audio encoder |
|Employ the Neural-THX Surround MultiMerge + Neural Loudness Control |
|The results ||Automatic detection of 5.1 encoded audio stream |
|Completely transparent, without any alteration to 5.1 audio content |
|Automatic detection of 2.0 audio stream |
|Once 2.0 is detected, continues to deliver simulated 5.1 surround audio in the stream |
|Viewers do not detect switch from 5.1 to 2.0 |
|Surround channels reproduce convincing audio from 2.0 input |
|Applies equally to the network's and affiliate's stereo content switching |
|Seamless and easy integration within standard infrastructure |
|Easy to use and operate |
|Delivers consistent audio quality and entertainment experience |