High-performance audio

We are all aware that the economy has slowed down, and that advertising revenue is down for radio companies across the country. With capital budget time coming up, it is time to reconsider all the widgets, gadgets and gimmicks needed for the radio station next year. Take stock of what the station already has, and focus on making all of it work well together.

For example, if the plan was to propose a budget for a digital upgrade of the plant, go ahead. Just don't be surprised if it doesn't get too far next year. I'm not suggesting that you sit on your hands; in fact, there are many things that can be done in a typical analog plant that can go along way toward improving the product. Perhaps a new competitor has shown up on the dial; one that may make future capital improvements even more distant.

Listening objectively

If an engineer thinks the sound at the station sounds perfect, then he must be listening all the time, right?

Pretending to be a typical listener is impossible, but you can be a super listener. Be more sensitive to problems heard over a radio than the average listener.

As a super listener, an engineer should notice problems before listeners do, and he should be able to analyze their causes.

Sometimes, as I am stuck in traffic on the way in to work, or on the way home from work, I let the radio run in scan mode, just to see what stations it stops on. I then try to honestly gauge my first impressions when it does stop. How do the station's competitors sound? How does the station sound? Can I always recognize the sound of my own stations, and my competitors? Does my station always sound better than my competitor? How about half the time? Two times out of 10?

Identifying the problems

If you could ride along in the average listener's car, and could ask them why they preferred one station to another, they would give you various answers, but the one you need to be concerned about is this: “Station X seems stronger. I can get it more easily. It sounds clearer.”


Knowing how the various components in a facility function, both independently and as a system, is the first step. Photo by Gary Kline.

“What do you mean it seems stronger?” you ask. “We're transmitting from the same location. We have the same power they do.”

“I don't know … it just is clearer. It sounds like they have more power.” You, as the station's engineer, have your work cut out for you. Remember that, to at least some degree, the station's fortunes are tied to your own.

Try to analyze what a lay person would mean by “clearer.” I would assert it means more intelligible. I would further assert that it means less noise, less distortion and more realistic high frequencies. Lyrics are more intelligible. Announcer voices sound more real. The other station may sound louder, even with what seems to be the exact modulation level as your station.

Identify the differences between your station and your competitor (or competitors), and then get to work.

Examining the system

The engineering element of a broadcast engineer's job is the knowledge, ability and experience necessary to assemble a bunch of disparate components designed by a bunch of anonymous engineers, into a system known as a radio station. There is some art to it, as well.

To maximize the performance of the system, one of the most important things to know is how the components of the system work together. Also, make sure the human components of the system are working correctly with respect to the electronic components.

Continue listening carefully. After listening to the station carefully, you have come to the conclusion that there is room for improvement. However, don't stop listening. Continue listening carefully. If you work at a station with a music format, familiarize yourself with the most commonly played songs by listening over and over again. Get to know the sound of jocks and other voice talent. If you don't, you won't be a super listener, and you won't be able to tell if other work you are doing on the system is having an effect.

Production elements. There are many radio formats but, I have yet to hear one that didn't use some sort of recording media, for the purpose of delayed playback. By this I mean anything from spots, to music, to long-format programs.


Ensure that the equipment operators are using the equipment properly and maintaining good audio levels. Photo by Ben Weiss.

Before diving into the electronics of a recording device, make sure that the users are recording the material correctly. If material is recorded too loudly, for example, there will be obvious distortion during playback. If the material is recorded at too low of a level, your broadband AGC may have to pick up the slack; when it does you may notice the effects of quantization. If material is recorded with left and right channels out of phase from one another, little will be heard on a mono FM radio. (Remember, at least half of the radios in the field are mono; many of the stereo ones will be in blend mode at any given time.) Inconsistencies in the way material is recorded can generate inconsistencies in system performance. Develop the correct recording procedures, and make sure your production people follow them. New staff should become acquainted with these procedures when they begin working.

On-air playback. Obviously, material that is recorded in-house is only part of what gets played over the radio station. Even the best material can be affected if on-air personnel do not handle it correctly. That's where console metering plays an important role.

Many consoles sold today use LED indicators and refer to something other than a VU scale; they may refer to the peak level the console electronics can handle, or something in-between. If your console metering is by VU meter, then most jocks will know how to use them. If LED meters are used, determine the correct operating indication, and make sure all on-air personnel know it as well.

Reference level and headroom. What is the correct level? In any audio system, there will be a maximum level of signal that can be passed from input to output without distortion. Generally speaking, this level is determined by the power supply voltage available to the amplifiers, and the amount of load connected to their outputs. Most every console made within the last 25 years uses a bipolar power supply that will run outputs of at least +15V and -15V. Most integrated circuits used in output circuits (such as 5532s) will have an absolute peak output capability of about 1V less than the supply voltage; so, two 5532s set up in a balanced line driver will be able to have a maximum peak output of about 28V peak-to-peak, assuming the power supply rails are at +15V and -15V. This translates to +22dBu. The nominal level of most systems (especially one with VU meters) is most likely +4dBu, and perhaps +8dBu. The difference in decibels between the nominal level and the peak level is known as headroom.

With music, a peak-to-average ratio of 12dB is to be expected. With a nominal level of +4dBu, and considering a peak-to-average ratio of 16dB, I would never expect my peak levels to exceed +20dBu. With a +22dBu limit, though, there is only a 2dB cushion prior to clipping the output amps. In theory that is tight, but in practice it is adequate. Make sure that jocks know that 0VU is the goal when setting levels during their show. Pinned VU meters could mean distortion is generated in the console, or perhaps downstream in the system.

If the mixing console uses LED metering, determine if the scale is referred to as a nominal level or a peak level. If it is referred to a peak level, but a few of the jocks think it is a nominal level (like VU), then there will be problems with the sound of the station.

Measuring system gain and frequency response. The mixing console is probably only one component making up a chain of equipment at the studio location and through the STL system. Be familiar with the performance of the entire system. Only then can you be sure that all the components are operating together correctly.


Check the audio system’s headroom by observing the point at which the audio is clipped. Photos by Ben Weiss.

The first step is to establish the nominal system level, and then to make sure that it remains constant from one end of the system to the other. Start by running an audio generator into one of the line inputs on the console. A 400Hz +4dBu sine wave fed into one of the line inputs, with a net gain of zero, should generate a +4dBu sine wave output, and the console metering should indicate that the nominal level is being met (i.e., 0VU on the console metering). Calibrate the meter amps if necessary. If there is a distribution amplifier attached to the output of the console, its output channels that feed the system should be at +4dBu as well. Other fixed-gain devices, such as an equalizer, should have an output level that is equal to the system level. If there is a variable gain device (such as an AGC) in the system, attach the measuring device at its input.

After establishing that the nominal level is constant through this part of the system, it's time to check the headroom. The receive measuring device should have a monitoring port that can feed an oscilloscope without unbalancing, or otherwise affecting, the point being measured. Once the scope is connected, look at the +4dBu sine wave. It should look clean. Increase the input to the console by 10dB; the VU meter will peg, but the sine wave should look the same only bigger. Increase the sine wave by another 10dB. The sine wave will have a clipped top and bottom now. Reduce the input level in 1dB steps until the clipping disappears. At that point, you have determined the headroom that the system really has in practice (at least up to the input of any variable-gain devices). The system should have at least 16dB of headroom above the nominal level.

After establishing this peak level, mute the audio generator. The resulting level should be around -80dBu to -90dBu. Calculate the dynamic range by taking the peak level and subtracting the noise floor (Example: +22dBu-80dB=102dB). Make sure that the measurement device is not high-pass filtered. You want to know about any low frequency noise that may exist in the system.

Another important facet of the headroom measurement is level vs. frequency. Once the headroom has been established at 400Hz, sweep down in frequency to 100Hz, then 50Hz and then 20Hz. The result should be close to the level at 400Hz (within a couple tenths of a decibel). Now go the opposite direction — up in frequency. Can the system pass 15kHz at the same level as 400Hz?

On-air processor located at studio site. Some or all of the on-air processing may be located at the studio site. Make sure the input gain of the audio processor matches the system's nominal level. Many audio processors (young and old) have different gain settings so that the user can match the gain structure of the processor with the levels expected at the input of the device. Make sure the processor does not expect a -10dBu input, when +4dBu is supplied. Doing so could generate distortion in the input amps of the processor.

On-air processor located at the transmitter site. If the on-air processor is located at the transmitter site, then the system measurements need to be extended all the way through to the transmitter site. The STL is often the weakest link in the chain. In fact, the performance of the STL will usually be the limiting factor in the overall performance of the system. Move the measuring device to the outputs of the STL, and take the same set of level vs. frequency measurements. Adjust the gain of each element in the system, from console gain, to VU meter calibration, to directional antenna gain, to STL input gain and output level, until constant levels are established all the way through. Remeasure level vs. frequency, and see how the high end (15kHz) compares to the low end (20Hz). Remeasure the noise floor and recalculate the system's dynamic range.

Stereo generators. If the station uses an analog stereo generator, it can cause different types of problems. Newer stereo generators (less than 10 years old) are going to have an output amplifier that can drive a long piece of coax (as much as 100 feet of RG-58). In most transmitter site installations, the length of coax on the output of the generator will be a non-issue. However, if the station owns an older stereo generator, refer to its manual to see what the recommended maximum amount of cable is that can be hung on its output. Attaching too much cable on the output of an amplifier not designed to drive a capacitive load, such as a long coax, could roll-off the frequency response of the composite system, causing a loss in stereo separation.

Analog composite STLs have similar issues. Typically, analog radio STLs are calibrated so that 3.5V peak-to-peak generates the full deviation of the system (otherwise known as 100 percent modulation). Input metering on the transmitter side, and output metering on the receiver site, will indicate 100 percent modulation with this voltage level on the transmitter input. While program material is running through the system, the output of the stereo generator should be set so that the STL metering is indicating peaks of 100 percent. If the peak level is too low, the overall dynamic range of the system is compromised. If the level is set too high, there may be subtle distortion generated in the demodulation process of the receiver. In either case, the on-air sound is affected in a negative fashion. Cable length is an issue with older analog systems. Keep the composite output cables as short as possible, locating the composite source as close to the exciter input as possible.

Optimizing the system

Lack of headroom. If the system cannot maintain at least 16dB of headroom, (down to 20Hz and up to 15kHz), see which device or link in the chain is specifically causing the problem. Do this by setting up a headroom test, and working backwards (or upstream, from the last device in the system, backwards to the console output). A headroom problem across the entire audio band could be attributed to something that has incorrect input gain vs. output gain. An example is if there is a distribution amp (DA) in the program chain that has an input gain control and an output gain control. Perhaps, by accident, someone turned up the input gain by 10dB, and then turned the output gain down by 10dB to compensate. The throughput gain is zero. However, the input amps get clipped early resulting in only 6dB of headroom. Remove the DA from the chain anyway (set it up to bridge the console output; don't run the air chain through it). When it comes to audio quality, less (equipment) is definitely more (quality). If you insist on leaving it, readjust the input gain so that the input amps can handle +22dBu, and then adjust the output amps afterwards so that the throughput gain is zero.


To maintain stereo separation, keep the length of the composite audio coaxial cable to a minimum. Photo by Ben Weiss.

What if the system lacks high-frequency headroom? Are there equalizers in the system? If so, how are they set up? A graphic equalizer, for example, set up in the smile configuration, may lack headroom at low and high frequencies. At low frequencies, the effect would be pronounced clipping. At high frequencies, the effect would be intermodulation of high frequency components (because harmonic distortion will be rolled-off farther downstream in the system). If you insist on leaving this configuration, readjust the input and output levels to maintain the 16dB of headroom at the opposite ends of the audio band. Make a note that the throughput gain at 400Hz will be something other than 0dB.

It is possible for the system to suffer from noise, especially when measuring the system's dynamic range. Low frequency noise related to the power line frequency can be easy or tough. A 60Hz hum can usually be attributed to inductive pickup. Move high-gain devices (such as mic pre-amps) away from big power transformers. A 120Hz buzz usually comes from dc power supplies. Check to make sure the filter caps aren't dried out. This type of noise can be introduced into a system by ground loops.

Emphasize balanced audio feeds — not only in the input and output amps, but also in the cabling itself. Analog radio STLs are sometimes noisy. Make sure the transmit and receive antennas are pointed correctly. Check that the transmission lines on the transmit end and the receive end are in good shape, and that the connectors are on correctly. Be sure the transmitter and receiver are as close to the same frequency as possible. If the receiver and transmitter are off, the result can be noise and subtle distortion from the demodulation process in the receiver. If the system periodically fades, consider getting bigger dishes on one or both ends, and shorten transmission line paths.

Transmitter tuning. The effects of multipath distortion on FM transmission are obvious. There isn't much that can be done about receiver multipath, but there is something that can be done to ensure that a station isn't transmitting multipath. The symptom of non-optimized transmission is synchronous AM noise transmitted along on the FM carrier. The basic component of the measurement setup is a sample element that produces a dc output that drives a meter. If a wattmeter sample is available in the output transmission line, use it for the dc sample and then minimize the synchronous AM in one of the following ways: Connect a scope across it (use a low vertical bandwidth, like 20MHz) and tune the transmitter for the least amount of noise riding on top of the dc; or, connect a sensitive audio amplifier across it and use your ears to minimize the amount of noise riding on top of the dc sample. (Do either of these as a touch-up after tuning the transmitter by the manufacturer's recommended means.)

AM transmitters: modulation level. Unlike its FM counterpart, the amount of power transmitted in an AM system is proportional to the amount of modulation. The more modulation, the louder and stronger the station sounds. However, because of the nature of AM, the transmitter is not forgiving of the overall peak level. Too much negative modulation creates splatter, and makes the station sound bad. Most current AM processors have tight level control, so it is possible to achieve consistent results by carefully adjusting the modulation level with the use of an oscilloscope.

Advancing to the next level

When it comes to audio processing, the old adage garbage in equals garbage out is true. Even without changing any of the settings on a processor, it's likely that the station sounds better after you have studied and fixed the on-air chain. Upgrading to the next level of quality involves making changes to the electronics of your systems components, the efficacy of which is subjective. Before going that far, continue listening to your radio station and see to it that you've conquered all the easy problems first.


Irwin is director of engineering for Clear Channel, San Francisco.


Today in Radio History

Milestones From Radio's Past

The history of radio broadcasting extends beyond the work of a few famous inventors.

EAS Information More on EAS

NWS XML/Atom Feed for CAP Messages

The feed provides feeds for all US states and territories.

Wallpaper Calendar

Radio 2014 Calendar Wallpaper

Need a calendar for your computer desktop? Use one of ours.

The Wire

A virtual press conference

Information from manufacturers and associations about industry news, products, technology and business announcements.

Join Us Facebook Twitter YouTube LinkedIn
Radio magazine cover

Current Issue

88Nine Radio Milwaukee

A New Home Brings Community Together

Browse Back Issues

[an error occurred while processing this directive]