It’s been a slow train coming, but HD Radio is happening. If your station has already invested in this technology, you may be preparing to kick back and enjoy hours of pristine sound.
While HD Radio is capable of delivering unsurpassed audio, you need to understand some of the issues related to cascading algorithms in order to reap all the aural benefits this medium has to offer.
In a perfect world, everyone would use uncompressed digital audio all the time. The downsides to uncompressed audio, however, are the amount of bandwidth required to transmit it and the comparatively large size of the files, which require more hard disk space. Compression of digital audio files is necessary to send audio files around economically in the real world.
Problems can arise when a station receives audio files from different sources, not knowing how those files were recorded, and how many and what types of codecs they have been run through. Coding systems are based on a psychoacoustic model of what people can and cannot hear; they are designed to keep processing artifacts below the threshold of perception. When an audio file is compressed several times, however, there is a good chance that this threshold will be breached, and sound degradation will become audible.
While the perils of cascading algorithms have been known since the early 1990s, the addition of HD Radio to the chain, which itself uses digital algorithms, raises new concerns for those charged with maintaining the station’s audio quality.
For example, Jon McClintock, commercial director for Audio Processing Technology, a codec supplier, notes that some difficulties have been experienced with DAB in Europe.
“While the problems of multiple psycho-acoustic passes have long been known, the reality only hit home when digital stations went live. In the U.K., a number of listeners complained of ‘muddy’ or ‘tinny’ sounding audio from commercial stations with DAB.” McClintock said that, with proper precautions, other networks in Europe were able to deliver CD-quality audio.
“Do you know where your audio has been?” asks Herb Squire, vice president of engineering and operations for DSI RF Systems Inc.
Squire, who is performing real-world tests of various codecs through AM and FM IBOC systems, adds that as much as possible, stations need to know the origins of source material and be aware of what compression has been used on incoming audio files. Issues related to codecs and HD radio are a wide-open area for research.
“We’re in unknown territory, so I’m not sure what I’ll find,” he said. “There will probably be some combinations of codecs that will work well with IBOC, and other that won’t.”
He adds that the problem of cascading algorithms will be especially pressing for all-news stations, who must take audio feeds from many sources in a variety of formats and get them on the air quickly.
Music stations with uncompressed digital libraries are not immune from such problems, however, as some record manufactureres have announced plans to distribute new releases via MP3. Squire plans to present the results of his tests “Dueling Algorithms Meet IBOC” at the NAB show in April.
Robert Reams, chief technology officer and co-founder of Neural Audio, said that the problem of cascading algorithms is troublesome because it can remain hidden for so long.
“Errors in cascading algorithms are an exponential function, and often the problem is hidden until it hits the final codec.” Stations may receive audio files that sound acceptable, but have poor quality when played over the air.
The best way to prevent the degradation of cascading algorithms is to make sure the initial recording is free of hum and noise. The traditional analog culprits, ground loops, flicker or thermal noise, do not exist in digital. But there are still ways that noise can make its way into the digital signal, among them video/computer monitor EMI “spew” captured by live microphones, preamplifier noise and noise or hum in external music or commercial content.
Reams said that while there may be few analog outposts in today’s digital stations, those that remain must be maintained to exacting standards.
McClintock of APT recommends four steps to avoid audio degradation: avoid more than two passes of psycho-acoustic based algorithms for files in the digital plant; use 256 kilobits per second for ISDN links; use 384 kbps for storage and networking; and use at least 256 kbps for distribution networks.
Degraded audio isn’t the only reason to be concerned about hum and noise in the system.
All of the new digital transmission systems allow for the transmission of non-audio content such as ancillary data, text or control signalling. This data is held in a separate buffer and merged into the output bit stream using space in the reservoir buffer when it is not required for audio.
“If the codec is busy dealing with audio hum and noise,” Reams said, “the result is slower updates on text, visual aids, control signalling or other revenue-generating features of the digital broadcast.”
Processing with care
Audio processing in the digital realm is another area where care needs to be exercised.
Frank Foti, president of Omnia Audio, said that devising a processing scheme for HD Radio requires a different mindset than what was applied to analog transmissions.
The most important difference involves dealing with data-reduced audio. In HD Radio, the processor needs to manage the audio spectrum in the most efficient manner.
“The processor needs to be thought of as a partner with the encoder,” Foti said, “much in the same way an FM analog processor works together with the stereo multiplex encoder.”
The goal of processing in the HD Radio world is to improve the intelligibility of the perceived audio. Although processing for effect is still possible, the concept of using the audio processor as a weapon to blow the competition away does not apply to HD Radio. “Once you’ve reached digital full-scale,” Foti said, “there is simply no more to be had.”
Because a station’s HD Radio and analog signals will run in tandem, either two audio processors or a dual HD/analog device will be required. More important, though, is the need for them to work together as a system.
Even though HD has been engineered so the audio from both the digital and analog signals arrive at the same time, audio spectrum and phase relationships must be similar on both the digital and analog transmission paths. If there are significant differences, the blend-to-analog action in the receiver will be abrupt and jarring to listeners.
Another concern is to understand thoroughly the audio routing layout for a digital transmitter installation. Foti said that the HDC system employs a master clock to sync everything to the 44.1 Khz sampling rate the system uses. Transmitter manufacturers have their own methods for audio routing and sync. The two need to work together, or problems will develop.
“Ensuring that your processing is operating in sync with the system, and is set to the correct sampling rate, will positively affect on-air audio quality.”
HD Radio is capable of bringing CD-quality sound to FM, but it is not simple as flipping a switch. This is fresh territory for everyone, and there may be some codecs that work better than others in conjunction with HD Radio gear.
As much as possible, broadcasters need to know the origins and coding schemes for all external audio files that are coming into the system. Don’t neglect the remaining analog gear in your station. Analog standards are higher than they ever have been.
Finally, remember that processing for the HD Radio plant requires not only gear that was designed to work with audio codecs, but properly matching the HD Radio and analog signals so that transitions from analog to digital and back in the listener’s receiver are smooth and seamless.