This is one in a series of articles about trends in audio processing.
Mike Pappas is senior vice president of sales for Orban.
Radio World: Mike, how can an engineer possibly choose the best processing settings to achieve both volume and quality, when there are so many ways a listener can consume audio, including phones, Bluetooth, car dashboards, desktop PCs and so forth?
Mike Pappas: I keep coming back to “listening.” As an engineer, you have to listen to all of the services your station is on and make processing decisions based upon what you hear.
Also, don’t just trust your own ears. Solicit feedback from people who can help you get the sound you need. An industry friend told me her ears were frequently “enlisted” to evaluate processing by an engineer in the area because she better represented the key demographic for the station than he did.
I know it takes time but it’s vitally important.
RW: How have the cloud, virtualization and “software as a service” affected the processing marketplace?
Pappas: We have been delivering those variants for several years now; it’s a natural outgrowth of the industry. They all have “plusses and minuses.” None is infallible or the magic bullet in terms of cost savings or performance.
For example, AWS has experienced recent outages. Running a “roll your own” virtualized server can be a significant challenge without a robust IT department. And SaaS is not without its own set of challenges. There is no “free lunch” with any technology.
RW: What about the impact of artificial intelligence?
Pappas: Cloud-based AI playout is a reality, and getting audio out of the cloud along with the associated metadata is a major issue.
We implemented HLS content delivery at the request of our largest customers. HLS has many advantages, with the biggest one being that the metadata is encapsulated with the audio segments so all the content is in sync.
We have one customer running 3 minutes of buffering (to allow Starlink firmware updates) using time-stamped HLS segments. We can play out those time-stamped HLS segments against the internal NTP referenced clock so the 8 a.m. segment plays out exactly at 8 a.m.
Additionally, we have developed a custom “First In First Out” 24-hour rolling backup system that stores the HLS segments and plays them out against the clock in the event of a major network failure. The AI-cloud is sending the next day’s playout in real time and storing it in the processor. And all of the associated metadata for RDS and Artist Experience will be correct and in sync. Listeners won’t know if it’s live or playing out of the backup.
RW: What should an engineer know in 2025 to make a smart selection of an on-air FM processor?
Pappas: If your processing is old enough to have a driver’s license, you probably need to think about replacing it. Seriously, what was acceptable processing in the 20th century isn’t going to work in the 21st.
Ninety percent of listening is in cars, and their sound systems are significantly better, with better FM receivers, than they were even five years ago. For example, my latest vehicle’s AM and FM radio is outstanding, with diversity FM and full HD including Artist Experience on a good-sized display. The AM section is sensitive, supports HD and simply works.
As more listeners purchase new cars, your processing has to keep up with their improved listening experience. Look for a processing company that can help you achieve this.
RW: Can a station use the same processing on its streams that it uses on the air?
Pappas: Analog stereo FM has some significant technological challenges that streaming doesn’t have, such as 75 uS pre-emphasis and its 15 kHz frequency response limitation. Streaming doesn’t have those technical limitations, but it has other challenges including low-bitrate codec issues. Streaming content needs to be processed independently of your FM.
RW: What should we know about dynamic range control for streaming and podcasting?
Pappas: Remember that podcasters are “content creators,” not audio engineers. A quick listen to a number of podcasts reveals all kinds of significant audio issues, some related to bitrate, others involve varying audio levels.
So work on ways to manage issues with the content creator. If you need to “fix it in the mix,” there are several software audio processors that can batch process podcast audio files faster than real time.
Also, a large majority of streaming and podcast listeners are in vehicles or “on the go” using a myriad of headphones, earbuds or you-name-it. So consistent loudness and limiting the amount of dynamic range are important.
Understand the limitations of the streaming codec and process for it accordingly. For example, you might find that reducing the HF limit to less than 20 kHz will improve the sound quality with some low-bitrate codecs.
And again, actually listen to it, and make processing adjustments based upon what you are hearing. There is no “one size fits all” from a processing standpoint.
RW: Should audio processing on multicast HD Radio channels be different from the main?
Pappas: HD-1 processing has to be set up so that the transition from analog FM to HD-1 is reasonably seamless. Loudness matching is a must. You can’t have a major shift in loudness during the transition. And almost 20 years into HD Radio, we are still battling audio diversity delay problems, which is breathtaking.
HD-2, -3 and -4 need to be at relatively the same perceived loudness as the HD-1. Those channels will be running low bitrates and have to be processed to keep the codec from misbehaving.
RW: What recently introduced features or capabilities are most notable?
Pappas: Our shift from audio processing to “integrated content delivery” has been a major change for us. And when you can control the processor and its operation using the HLS metadata, it’s a game-changer.
We owe the folks at Super HiFi, EMF/K-Love and Audacy a big thanks for working with us to make HLS content delivery a reality and on the air.
RW: What are processing designers thinking about the future?
Pappas: We think there is a very bright future for the content delivery business. Listening to our customers and looking at what’s going on in the industry are driving our next round of product development. Bob Orban may have said it best: “It’s always been about music, math and imagination in the service of broadcast audiences.”
Read more about trends in audio processing for radio in a free Radio World ebook
