The issues of pre-emphasis, de-emphasis, clipping and repeater audio quality

Up one level
Back to Home The issue of pre-emphasis, de-emphasis, clipping and repeater audio quality.
By Paul Sexauer K3VIX

Preface and Introduction: The issue of pre-emphasis, de-emphasis, clipping and repeater audio quality have been debated in the ham arena since two meter FM became popular. The following are my thoughts based on 26 years in engineering with Motorola Inc.

I still think that the MICOR is probably the best repeater for ham use since you can do just about anything with it using wire cutters and diodes (no software needed!). I did work at Motorola for 26 years before taking early retirement in 1998. The first half of my career there was primarily in RF design and the last half was involved primarily in systems engineering. I'm presently the RF Systems Engineering manager at TX RX Systems out here in New York (near Buffalo). It was kind of nice to escape the hustle and bustle of the Chicago area plus now I get to work on the "other" end of the business.

I did maintain a MICOR ham repeater for about 10 years out in Aurora, IL and it had a reputation for excellent audio. I ran that one "by the book" as far as audio adjustments and that is described later on in this article along with background info on FM, PM, pre-emphasis, de-emphasis and other fun stuff.

The goods: The first question which needs to be addressed is why pre-emphasis and de-emphasis is used. The answer can be found when the properties of FM demodulators are analyzed. If the output of an FM demodulator (discriminator, quadrature detector or whatever) is monitored on an audio spectrum analyzer with no RF carrier input, it will be noted that the response displayed is one which rises at a 6 dB per octave slope. (The mathematics for the FM demodulation process predict this effect.) In other words, with a flat RF noise spectrum entering the demodulator, the output will exhibit a noise characteristic which rises with frequency. If an FM signal were applied to the input of the demodulator and the modulating frequency were swept from low to high frequencies, maintaining constant deviation, it would be noted that the signal to noise ratio at the output of the demodulator would degrade as the modulating frequency increased.

To compensate for this, a de-emphasis circuit is used. In its simplest form, this would consist of an R-C network which would roll off at a rate of 6 dB per octave, canceling out the rising noise characteristic of the demodulator. This eliminates the problem of S/N ratio degrading as the modulating frequency rises, but now results in a rolling off of the audio response if a "flat" FM signal is received. To have a net "transparency" in audio response, it becomes necessary to pre-emphasize the transmit audio at a corresponding 6 dB per octave rate. From this, you can see that de-emphasis came first and created the requirement for pre-emphasis at the transmitting end.

Going to the transmitting side, we have two seemingly different modulation schemes available: FM and PM. The major difference between these two schemes is that a phase modulator has a 6 dB per octave rising audio response (i.e. pre-emphasis is inherent in PM systems with no added circuitry). In a phase modulator, the total deviation is a function of both the modulating signal amplitude as well as frequency. In a direct FM modulator (i.e. in one where the modulating signal is applied across a varactor to vary the oscillator frequency) the deviation produced is a function of the modulating signal amplitude only. To achieve pre-emphasis, a series R-C network needs to be inserted into the audio path.

A sidelight here is that if you look at the mathematical representations of a PM signal versus an FM signal, the difference is that the modulation component in a PM signal is the mathematical derivative of the modulating signal in the FM signal. If you add the series R-C circuit ahead of the FM modulator, the signal at the output of the R-C network is the mathematical derivative of the applied signal. In other words, by adding the series R-C (pre-emphasis) circuit, the output of the FM modulator is now identical to the output of the phase modulator. The output of a direct FM transmitter with pre-emphasis can not be distinguished from the output of a PM transmitter with no added pre-emphasis circuit. The pre-emphasis circuit effectively "makes" an FM transmitter into a PM transmitter.

In the early days of FM, there were no varactor diodes so it was difficult, at best, to produce a direct FM modulator. PM modulation was, however, easy to achieve which is the reason that it was the "standard".

The rising audio response of the phase modulator did cause some problems for manufacturers. Since the audio response fell off at the low end, phase modulators had a hard time when it came to modulating them with PL tones. Direct FM modulators, being flat, had no such problem. With the advent of data communications, direct FM was the only way to modulate a carrier with baseband data.

Now, since the FCC (and the similar authorities in other countries) mandate that FM transmitters must not exceed their assigned bandwidth allocations, some means of limiting the deviation was needed. Along came the clipper circuit. The purpose here was to ensure that the modulating signal amplitude never exceeded a certain specific value (that which produces rated deviation). Of course, clipping an audio signal produces major distortion in the signal and this must be minimized. Therefore a "splatter filter" is inserted after the clipper. This splatter filter is simply a low pass filter and is designed to roll off the harmonics created during the clipping process. Note that the splatter filter does not "undo" the effect of the pre-emphasis circuit. The splatter filter does not "kick in" until the upper end of the audio range.

At this point, let's put some numbers on the pre-emphasis circuit along with the splatter filter. For normal voice communications, the FCC specifies (in commercial radios) that the pre-emphasis be a 6 dB per octave rising response beginning at 300 Hz. In actuality, most commercial transmitters will begin the rising response between 200 and 300 Hz. The FCC further specifies that the pre-emphasis be within +1 and -2 dB of the 6 dB per octave response. This effectively defines a "box" which the TX audio must fit in. Above about 3000 Hz (I don't have the spec in front of me) the FCC requires the audio response roll off at a minimum of 12 dB per octave. This is the response of the splatter filter. Generally, the splatter filter begins to have an effect at around 2500 Hz. At that point, the rising audio response will begin to flatten out and then begin the 12 dB per octave roll-off at around 3000 Hz.

The way the transmitter audio response is measured is to apply a fixed modulating signal level which produces 60% of rated system deviation (3 kHz in our case) at 1000 Hz. Holding this level constant, the frequency is then swept from 300 Hz to about 10 kHz and the output of a calibrated receiver discriminator is plotted. The pre-emphasis and splatter filter responses are then clearly visible.

On a side note, the received audio quality is largely controlled by how "hard" the clipping in the transmitter is. If the clipping is "soft" i.e., it only occasionally goes into significant clipping, the audio will sound very clean. If the clipping is "hard", the intelligibility suffers greatly. Some manufacturers allow for this by providing a deviation set pot (IDC in Motorola) which is set to the rated deviation with the modulating signal in full clip, and then also provide a mic level or line level control to adjust how hard the modulating signal hits the clipper. Proper setting of this control is imperative if good audio is to be maintained. We will see the importance of this in setting repeater levels shortly.

In the case of most amateur transceivers I've seen, the clipping circuits are quite poor which means that deviation may not be limited as much as might be desired, but it does result in very clean sounding audio. The ham market is not subject to the same technical requirements as the commercial folks. Clean audio is, however, achievable with good commercial clipping circuits.

Now we get to the question of how to obtain good quality repeater audio. To do that really requires that we know what produces "bad" audio. As was discussed above, in a mobile or base it is important that we control how hard we go into clip on audio peaks. Harder clipping yields degraded intelligibility. With a repeater, the problem is compounded. This is because we have a clipper in the mobile and another in the repeater. It all comes down to properly setting repeat audio levels. I'm going to stick with the MICOR here since I know that one really well.

In the MICOR, the first step in setting up the audio path is to set the IDC for 5 kHz system deviation. This is done by injecting a 1 volt rms. 1000 Hz tone into the Exciter audio input. This runs the audio into full clipping. The IDC pot is then adjusted for 5 kHz (this should be 5 kHz including any PL which may be generated). The alignment procedure then asks that you reduce the audio level from the generator until the deviation drops to 3 kHz. Write down the audio level into the exciter at this point. Don't rely on the factory value since that was probably done a long time ago. This is referred to as the modulation sensitivity "mod sense" level and is written on the exciter at the factory. At 3 kHz deviation, you are out of clip. At this point, the exciter audio adjustments are done and should not be changed.

The next step is to adjust the Repeat Level (this applies whether you are using a standard MICOR or have an external controller supplying audio to the exciter). Here, you inject a signal into the repeater receiver with 1000 Hz modulation and 5 kHz deviation. You measure the voltage at the exciter audio input point and adjust the Repeat Level pot for a reading equal to twice the mod sense value. Since the mod sense level is the level to produce 3 kHz deviation, setting the input to twice that you should get 6 kHz deviation out with 5 kHz deviation input. Now, the clipper will not allow you to exceed 5 kHz deviation, so this adjustment puts you into a "soft" clip situation. It actually provides a bit of an audio boost for signals which are coming in at less than 5 kHz deviation which is probably the norm. Adjusting repeat audio in this way will produce "near simplex" audio quality. I maintained a repeater out in Aurora, IL for about 10 years and used this procedure exclusively. We had a reputation for one of the best sounding repeater audio systems in the area. I am not a believer in pulling receiver audio off the discriminator and directly into a "flat" transmitter. That originated in the ham circles and I don't believe that it will produce audio any better than "doing it right the first time"! Let the pre-emphasis and de-emphasis circuits do their job (it really results in a fairly flat response) and pay attention to the clipping levels. You'll have great audio and you will keep your deviation within spec, which is mandatory in today's FM bands. Actually, if we could get everyone to adjust their repeater deviation to 4 kHz, a lot of adjacent channel problems would be reduced to acceptable levels. Just a thought.

There are two other web pages on this topic:

An Explanation of "Flat Audio", "Pre-Emphasis" and "De-Emphasis"
Pre- and De-Emphasis, Explanation and Assistance: Running the Numbers

Back to the top of the page
Up one level
Back to Home

This page created May 5, 2001 by Kevin K. Custer W3KKC

This web page, this web site, the information presented in and on its pages and in these modifications and conversions is © Copyrighted 1995 and date of last update) by Kevin Custer W3KKC and multiple originating authors. All Rights Reserved, including that of paper and web publication elsewhere.