FM Overmodulation: Cause & Cure
WITH THE continued improvement in recording and FM broadcasting equipment, the increased use of special equalization effects by recording artists and studios, the differences in microphone equalization and response, and the use of close microphone techniques, the high frequencies fed into FM broadcast transmitters are often of sufficient amplitude (after pre-emphasis) to cause serious overmodulation problems. The increasing number of incidents of FM overmodulation, and the resultant FCC citations, indicate a real need for corrective actions.
Why Does the Problem Exist?
Normally the lower frequency
component of an audio signal is
amplified more than the higher
frequencies. This occurs at every
step between the input equipment
and the transmitter, unavoidably
adding a high frequency noise
to the desired audio signal. As a
result, the signal-to-noise ratio is
low at the high frequencies. Where
the audio spectrum from CO to
15,000 cps is utilized, a situation
such as this cannot be tolerated.
To improve signal-to-noise ratios, various methods of equalization and pre-emphasis are employed. Pre-emphasis—increasing
the amplitude of high frequency fed to the transmitter — reduces
the unfavorable relationship between high frequency audio and
high frequency noise. High frequency program signal is increased, while the high frequency
noise level remains the same, improving the high frequency signalto-noise ratio.
However, as signal-to-noise ratios increase, equalization or preemphasis can actually result in
signal degradation due to distortion brought about by the excessive high frequency signal levels
that are forced through the various amplifiers. The main reason
for the equalization and pre-emphasis network is to make sure
that high frequencies are not
blocked out by the inherent noise
characteristics.
Since there is less energy in the
upper section of the 50 to 15,000
cps region than there is in the
portion below 1,000 cps, a 75-
microsecond pre-emphasis curve
was adopted by the FCC to take
advantage of this distribution.
Fig. 1 illustrates the standard preemphasis characteristic curve. The
solid curve shows a flat response
below 200 cps; at 1,000 cps the
gain rises to +0.9 db, to +8.3
db at 5,000 cps, to +13.8 db at
10,000 cps, and at 15,000 cps the
gain increases to +17 db. In general, a 75-microsecond pre-emphasis curve means that amplifiers
all along the audio line will be
handling a 15,000 cycle signal 17
db higher than the 1,000 cps signal. If the transmission at 1,000
cycles is to be maintained at an
adequate level, there is a definite
probability of overloading the amplifiers at 15,000 cycles when using the 75-microsecond pre-emphasis.
What are the Causes?
The FCC standard pre-emphasis
curve is quite severe. When this
curve was adopted, FM programming was not too competitive and
the FM broadcaster usually operated the transmitter with low values of modulation to allow the
transmission of the full dynamic
range of recordings. The standard
practice for FM broadcasters was
to operate at a low modulation
level without the use of peak limiting amplifier& Although limiters are being used today, more
FM stations are still being given
citations for overmodulation. This
has led many engineers to the conclusion that limiters were not operating correctly. However, tests
showed that they were operating
on a flat response curve, and the
high frequency signal fell below
the threshold of limiting. Overmodulation is caused by pre-emphasis of the audio signal after
it passes through the limiter.
A limiter cannot always handle
the complete job of loudness and
level control. High-amplitude,
high-frequency signals can trigger
the limiter and cause a drop in
overall program level, balance, and
loudness. Where the limiter is inserted after pre-emphasis, it will
be particularly susceptible to this
triggering with a reduction instead of an increase in signal
strength. This method of operation can limit the high frequencies as well as the mid-range and
low frequencies. If the program
level is maintained well below the
threshold of limiting, acceptable
program levels will be produced.
However, some programs will
cause unnatural effects, as shown
in Fig. 2. The results shown are
for a limiter with an attack time
of approximately one millisecond
and a 90% recovery time of about
two seconds. When the pre-emphasized high-frequency peaks exceed the threshold of limiting,
gain at mid-frequencies is reduced. Thus, the de-emphasized
signal sounds like the limiter is
undergoing a blocking-type of oscillation; the gain suddenly drops,then recovers on the normal RC
slope. This type of programming
will produce a large group of complaining listeners. The right side
of Fig. 2 shows the same signal
without pre-emphasis.
If the limiter is placed before
pre-emphasis, and if it is not triggered by high frequency peaks,
the pre-emphasis can -cause transmitter overmodulation with all its
distortion and problems. Live studio programs, music, and other
recorded sounds containing applause, percussion noises, fingersnapping, clinking and tinkling of
glasses and keys, tap dancing, and
other peaked high frequency
sounds (when combined with
program music or speech) will
cause transients that will trigger
the limiter and actually cause an
attendent gain reduction. If the
high-frequency content of music
and other recorded sounds never
exceeded the curve in Fig. 3, the
limiter signal could be fed into
the FM transmitter pre-emphasis
network without causing overmodulation. However, present day microphone techniques and orchestration, combined with RIAA treble pre-emphasis, often create
quite impossible levels at the high est audio frequencies. Fig 3 is
complementa•T to the standard
pre-emphasis curve and can be
used in the FM receiver to deemphasize the signal to restore
the original frequency response.
It can be seen that overmodulation will result if the high frequencies exceed the limits of the
curve shown by the shaded area
in Fig. 3.
FM transmitters are designed
to tolerate extensive overmodulation with a minimum of distortion.
However, considerable overloads
can swing the carrier beyond the
maximum 75 kc limit. A 6 dbprogram peak with low percentages of modulation will produce
a carrier swing to ±.150 kc, which
is well beyond the allowable limit.
Numerous measurement studies,
using spectrum and wave analyzers, have been made.* The data
provides a typical peak power distribution curve, as shown in Fig.
peak power measured from many
types of programs broadcast from
a typical FM station. During
these measurements, the limiter
was set up for a constant midrange level. The very high and
very low frequency peaks occurred
much less frequently than the intermediate high and low frequency
peaks. Yet, these peaks show why
broadcast stations are being given
citations for overmodulation.
From this study the broadcaster
can assume that every FM broadcast station, without preventive
overmodulation systems, will produce a peak power distribution curve that will equal or exceed
Fig. 4.
To reduce the excessive accelerations which occur at high frequencies, many FM stations simply reduce the modulation level to
an average of 50%. This practice
reduces the average program power output to one-quarter of the
maximum allowed (Fig. 5.) The
curve indicates that no overmodulation occurs below 5,000
cycles, but it is possible to overmodulate the transmitter as high as 70% at 15 kc. The 10-kc portion of the curve shows that overmodulation can exist up to 150%.
Thus, even though the mid-frequencies are controlled by the
limiter, the transmitter should be
operated at a lower level of midfrequency modulation to prevent
high frequency overmodulation.
Even with 30% modulation, signals containing sharp transients
could still cause more than 100%
modulation at high frequencies.
With 30% average modulation, the
program power output of the FM
station is approximately one-tenth
of the maximum power allowed.
In general, operation with very
low modulation levels is not very
practical because many of the
peaks which cause the reduction
in modulation are so great thamany FM receivers will not pass
the signal and listeners are usually incapable of hearing it.
What Are the Cures?
The most logical approach would be to eliminate the necessity of pre-emphasizing FM signals in the transmitter. The great hardship in achieving this objective is to have the set manufacturers modify receiver circuitry. Since this isn't immediately practical, corrective measures must be employed. One method is to install a low pass filter which NVill cut off all signals that fall on the slope of the filter curve (Fig. 6). This actually results in signal degradation just to protect the station from a relatively few overmodulation peaks. Thus, the low pass filter is not the cure for the overload problems. Another method of controlling the high frequency peaks, developed primarily for the disc recording industry, pre-emphasizes the program material, acts upon the troublesome high frequency signals, and de-emphasizes in a complementary manner to give an overall flat output. Although this type of operation can eliminate overmodulation in most cases, it can cause gain reduction at all frequencies above 600 to 900 cycles with some resultant signal degradation.