Processing for your content

We are content providers now, as you may recall. In fact, about the only aspect of our content provision that is increasing is the use of streaming audio. For this reason I suggest avoiding grabbing an old processor off of the shelf in the engineering shop (or worse yet, bringing one back from the transmitter) for processing a streaming audio feed. Even though the reasons for processing the streaming audio are basically the same as that for over-the-air signals, the methodology is different in some key respects. For example, clipping is the most basic way to build loudness in an analog system. The reason is that it is easy to get away with, in the sense that distortion products are often either super-audible, or so fleeting in nature that they're almost impossible to notice.

Specialized processors fit specific needs, such as accurate leveling with the Tranlantech Ariane Sequel or low bit-rate processing with the Neural Audio Neustar 4.0.

It's not that simple in a data-reduced system; you do not want a streaming encoder trying to allocate bits to distortion products. That results in fewer bits left over for what you really want to hear. Throw that idea out.

Another way we have built loudness in our old-fashioned analog systems is by the use of HF limiting. The digital systems we are now using are flat — that is, no emphasis — so the HF limiter idea is of no value there either. And finally, we all know that our analog systems don't necessarily have a rigid brick wall limit through which they can't pass — unlike digital systems, which are all burdened with their 0dBFS limits that cannot be exceeded.

Therefore, the three key aspects of audio processing for data-reduced paths (whether it is HD Radio or streaming audio) that remain in common with established analog techniques are wide-band AGC, multi-band AGC and peak limiting. Modern processors often use look-ahead limiters as opposed to the peak limiter (fast attack and fast release) that most of us have grown up with. So what's available?

Translantech offers the Ariane Sequel digital audio leveler. This unit uses a digital signal path all the way through. It includes a sum and difference mode with an independent or stereo linked mode available; output peak limiting; sample rates of 32kHz, 44.1kHz or 48kHz (syncs to input); eight user presets; TCP or RS-232 remote control; and peak and RMS readings on input and output sides.

Thanks to DSP, processors are more compact but pack a great deal of processing power, such as the Orban 6300, the Omnia One and the Vorsis HD-P3.

Neural Audio offers the Neustar 4.0, a single rack unit audio processor built specifically for low bit-rate applications such as HD Radio, streaming audio and podcast encoding. Make adjustments from the front panel or via TCP/IP and a browser interface. The unit features balanced analog inputs and outputs, AES3 (32kHz to 96kHz sample rate) inputs and outputs, and word clock in/out.

Omnia has recently introduced a processor made specifically for data-reduced transmission paths: the Omnia One. This processor features four bands of AGC and four bands of limiting, AES3 inputs and outputs with sample rates from 32kHz to 96kHz; RS-232, GPIO and Ethernet ports for remote control.

Orban's 6300 is a 1RU audio processor that can be used in a number of applications, including HD Radio or streaming audio. It includes a stereo enhancer, wideband AGC, equalizer, a multi-band compressor/limiter and two (stereo) look-ahead limiters available via one stereo analog output and two AES outputs. It also includes a low delay (about 5ms) talent headphone output.

Broadcast Warehouse offers the DSP Xtra, designed for processing audio for FM, HD Radio and streaming audio. It includes the Ariane leveler, multi-band limiting and multi-band look-ahead limiting. Adjustments are carried out from the front panel, or via USB, serial or Ethernet. The AES3 input will sync with any sample rate between 32kHz and 96kHz. The unit features a built-in stereo generator as well.

Wheatstone now offers the Vorsis HDP3, a 1RU audio processor with built-in three-band AGC, parametric EQ, de-esser, downward expander and peak limiter. The unit has a built-in IBOC delay. Adjustments are made via the front panel, via local computer control or via Ethernet.

Linear Acoustic recently introduced the Aeromax HDFM, an all-in-one processor for FM, MPS (HD1) and SPS (HD2). It features an AGC, five-band compression and look-ahead limiting, an HD Radio-delay function built-in, an AES3 input for sample rates between 32kHz and 96kHz that locks to incoming signal or word clock input, and RS-422/485 or Ethernet communication. A stereo generator is optional.

Inovonics manufactures the Omega FM, an FM processor and stereo generator. This 2RU device relies on a Pentium-class microprocessor to do all the processing and multiplexing functions. Sports analog and digital inputs, composite and AES outputs with low-latency making this product suitable for headphone monitoring purposes.

Audio processing is one province of the field of broadcasting that remains solidly in the hands of the engineer at a radio station. It's wise to keep up on the latest trends in that art. I also happen to believe it is imperative that we, as broadcasters, do everything we can to keep our leadership position in the dissemination of content — whether in the methods practiced for 80 years or the brand new methods via HD Radio or the Internet.

Irwin is the chief engineer of WKTU-FM, New York City.