Multipath Distortion - Menace to FM Stereo?
February 1963 Radio-Electronics

February 1963 Radio-Electronics [Table of Contents] Wax nostalgic about and learn from the history of early electronics. See articles from Radio-Electronics, published 1930-1988. All copyrights hereby acknowledged.

We have been told that rock music's Buddy Holly perished in an Iowa plane crash in 1959 - "The Day the Music Died." I'm thinking maybe it was a ruse in order to fulfill Mr. Holly's secret ambition to design amplifiers for the Marantz electronics company, under the alias of Richard Sequerra. Marantz, founded in 1953, is still in the business of designing high quality receivers and amplifiers. But I digress... This 1963 article in Radio-Electronics magazine called upon industry leaders to comment on the deleterious effects which multipath can have on the reception of stereo FM ( frequency modulation) radio. Commercial FM stereo broadcasting was still in its infancy at the time. Left-right channel separation was made more difficult when multiple signals are present in the analog decision making circuitry, and acceptance by the public depended on successful operation. Stereo was a big selling point for a new paradigm in music radio, so listeners had to be convinced the increased cost for receivers and speakers needed to be justified. Modern digital signal processing helps to mitigate multipath effects.

Multipath Distortion - Menace to FM Stereo?

Six Industry Authorities Analyze the Problem

Multipath distortion occurs "when an FM signal reaches the receiving antenna by two or more separate transmission paths. This is usually caused by reflection from an obstruction near either the transmitting or the receiving antenna. The signal reflector could be a large building or hill, or a moving object such as an airplane." The phenomenon is well known to television viewers as "ghosts."

Multipath distortion has been noted since the beginning of FM, but has not given most listeners much trouble. With stereo, things may be different. A much wider bandwidth is required, and multipath "ghosting" may even - under certain circumstances - put the left channel on top of the right one, confusing the stereo effect completely.

Nobody seemed to know exactly how serious multi path distortion might be, so Radio-Electronics sent a questionnaire to a number of authorities in the field. We asked whether audible distortion, impairment of separation, or both, are the most important effects of multipath distortion; how the number of listeners affected by multipath distortion on FM stereo would compare with the number now having trouble on monophonic FM, and what is the chief hope for reducing the effects of this type of distortion. We also asked how the listener could recognize multi path, and if there were any other causes of distortion which might be confused with it. The response appears below.

Multipath signals cause both audible distortion and impairment of separation in FM stereo reception. The audible distortion usually shows up as mushy middle tones, and raspy and harsh high tones; the impairment of separation is normally indicated by the loss of stereo effect of the subcarrier "second channel" and the presence of background noise or hiss in this "second channel". A classic example of the latter can sometimes be heard as "airplane flutter."

The degree or seriousness of multi-path distortion definitely varies with locality, but our experience indicates that it is widespread and causes some distortion in a large percentage of all locations, if a top-quality directional antenna is not used.

Many other causes of poor reception and distortion, ranging from poor recordings and faulty station operation to tuners, amplifiers and speakers, of course exist. However, these are frequently blamed when the true cause is multipath distortion. And such multi-path audible distortion is found in monophonic FM, as well as stereo FM, apparently to nearly the same degree and frequency. Monophonic FM is not faced with the problem of separation of the subcarrier second channel that aggravates the stereo FM reception problem.

The FM receiver, whether monophonic or stereo and regardless of its quality, sees its receiving antenna as its source of signal, and the fidelity of its reproduction is limited to the purity and sufficiency of this signal. Receiving antenna gain is of real importance in more remote locations, and rotators may be required if directions vary greatly. High front-to-back ratio, excellence of impedance match, narrow directivity pattern front lobe and elimination of multiple side lobes are the prime factors in most locations if multipath distortions are to be eliminated, and real high fidelity realized.

We have carried on a veritable avalanche of correspondence with the field (individual consumers as well as service personnel), which verifies the conclusions from our own engineering test work that high front-to-back ratio on the receiving antenna is probably the most single important item in the elimination of multipath distortion, which has shown up in a surprisingly large percentage of stereo FM receiving locations. A narrow directivity pattern front lobe and elimination of multiple side lobes seem to run a good second in curing these problems, along with excellent impedance-match characteristics.

Impedance-match characteristics seem to become more important as the installation becomes more remote, since installation is made on a higher mast requiring longer transmission line, thus making line reflections from impedance mismatch more important.

The fact that FM receivers have greater sensitivity than TV receivers and that FM enthusiasts are more critical and discerning seem to have combined to make the elimination of multipath reflections more critical and demanded in high-quality FM reception than in most TV reception. We also find that the FM enthusiast is more apt to study and intelligently compare FM components and antennas than the vast majority of TV viewers.

John Frank

The most important effect of multipath distortion upon stereo reception, beyond that upon monophonic reception, is that it introduces noise and phase-shift effects in the range between 23,000 and 53,000 cycles - that part of the transmission channel carrying the stereo information.

This is heard as whistling, rushing or warbling noises in the background, accompanied by a rasping distortion, which is partly or totally eliminated when the listener's system is switched to monophonic reception.

Few localities are immune to multipath problems, although heavily populated urban areas, with concentrations of structures that reflect radio waves, undoubtedly present more of a problem than suburban or rural locations.

Distortion caused by multipath reception can easily be mistaken for poor alignment or low signal strength, and vice versa. The simplest test is to try another station in a different direction. If a rotatable antenna is being used, and it is sufficiently directional, turning it will usually establish whether the distortion is due to reflected signals.

In monophonic FM listening, the result of multipath interference is not as likely to be heard. It most often consists of ultrasonic noise, which, even if passed by the audio circuitry associated with the tuner, is not likely to be audible. It is the much greater audio bandwidth required by stereo transmission that has enlarged the problem, since the ultrasonic noise is now converted to audible frequencies by the multiplex demodulating circuitry. Moreover, the phase disturbances associated with multipath reception can severely confuse the multiplex circuit, which depends on exact phase information to sort out the left and right channels properly in a stereo program.

So, assuming that no more actual interference exists now than did before multiplexing became common, the stereo listener is more likely to have his listening disturbed. Since multipath reception causes distortion of phase, frequency and amplitude, tuners that exaggerate these effects will be especially susceptible to multipath problems; this is particularly true where limiting is poor, i.f. bandwidth narrow, or capture ratio (the rejection of co-channel interference) is inadequate.

While no simple method of designing tuners presents itself that would totally eliminate such interference, or, rather, its effects, the theory of FM reception supports the hope that this may someday be possible. In the meantime, rotatable directional antenna systems would appear to offer the home listener the best hope of improving reception now disturbed by multipath distortion.

Robert E. Furst

For monophonic FM reception multipath distortion has never been a major problem. FM limiters have been able to eliminate the variations in signal intensity due to cancellation and reinforcement of multipath transmission. A slight phase shift in the higher end of the audio spectrum also caused by multipath transmission is not noticeable to the ear. As a result of this, multipath reception was troublesome only in fringe areas, where transmission cancellations caused the signal at the receiving antenna to drop below the limiting capabilities of the receiver.

With FM stereo reception, the situation changed. Since the faithful reproduction of the stereo FM program requires the maintenance of very carefully controlled phase relationships of the main and subcarriers, any phase distortion caused by multipath interference will be noticeable to the listener. It manifests itself in loss of stereo separation and, in severe cases, may give rise to harmonic distortion.

During this last year, we have had the opportunity to introduce many thousands of FM tuners and receivers to the market, and have been able to study, on a nation-wide basis, the difficulties caused by FM multipath distortion. Fortunately, it appears that a very small percentage of listeners have been plagued by this difficulty. In almost all of these cases, a change in the antenna system eliminated the problem. As a remedy, we have recommended the installation of an antenna with strong directional characteristics, such as a Yagi, and a carefully installed antenna transmission line to link the receiver with the antenna without the introduction of standing waves.

Quite often the difficulties attributed to multipath interference are actually caused by improper adjustment of either the stereo broadcast equipment or the multiplex circuitry within the tuner. The audible results are similar, usually again lack of stereo separation, and audio distortion.

What can be done within the framework to the receiver design to minimize the effects of multipath interference? In my view, the most important pertinent characteristic of the receiver is the quality of the limiter circuitry and the capture ratio of the tuner. Good limiting will permit the receiver to accept the secondary transmission path only when its intensity is in the same order of magnitude as the main path. This is the case only on rare occasions.

Visual indicators have been explored which will show in some form of display the amount of multipath at the antenna. The thought here is to rotate the antenna for optimum results. All schemes proposed so far have been cumbersome and difficult to manipulate by the nontechnical public, and our field experiences have not indicated any need for this type of device.

F. L. Mergner

In FM stereo reception the most important effect of multipath propagation is audible distortion, sometimes causing complete loss of intelligibility. Multipath distortion is accentuated during stereo broadcasts because the method used to transmit stereo information requires, not only a much wider frequency spectrum than monophonic FM, but depends to an even greater extent on amplitude and phase relationships. In addition, multipath propagation can cause an imbalance in the two sidebands produced by modulation of the 38-kc suppressed subcarrier, thus increasing distortion. Although multipath also decreases channel separation, this loss is not nearly as objectionable, because the accompanying distortion is much more disturbing.

Multipath distortion is normally more prominent in large cities due to the greater number of reflecting surfaces, such as tall buildings, especially those with large metal surfaces. However, it can occur in outlying areas as well. Tests using monophonic transmissions show that distortion becomes more serious as the difference in path lengths and the amplitude ratio between reflected and direct signals increases.

Loose connections, especially at one of the antenna terminals, can cause mismatch between the antenna and the input circuit, a condition which will reflect and delay a portion of the signal and cause distortion which could easily be mistaken for multi path. A drifting 38-kc generator in the multiplex section, poor synchronization due to marginal gain of the 19-kc amplifying circuits, especially during weak signal periods, can also cause phase distortion in addition to loss of separation.

In some locations there can be a problem with multi path reception even on monophonic programs. Investigations have shown that for path differences less than about 1 1/4 miles, delayed signals of relatively large amplitude can be tolerated as long as the FM receiver maintains adequate AM suppression. As the path difference increases from 1 1/4 miles, the spectrum of distortion extends to higher order harmonics. Since the ear is very sensitive to such harmonics, the tolerable amplitude of the delayed (reflected) signal falls off very rapidly over the range of path difference from 5 to 12 miles.

For path differences of 18 miles and more, the harmonics spectrum extends beyond the audible range, and the distortion is heard as a hissing noise.

Tuners which maintain a minimum AM suppression of 35 to 40 db under all receiving conditions are able to reduce multipath distortion during monophonic reception to acceptable levels. Therefore, I do not feel that this problem is too serious in most areas at the present time.

I would estimate that in urban areas one out of two installations using an antenna that proved sufficient for monophonic reception could have trouble with multipath distortion during stereo transmissions.

Antennas, highly directional in the vertical as well as horizontal plane, and rotators solve most of the problems. However, good tuner design can also help. For example, having the proper match between antenna and input circuit of the FM tuner will eliminate the possibility of reflections in the lead-in. Well designed tuners with a good capture ratio (low numerical figure in db), as well as a high order of AM suppression, will generally provide much more satisfactory monophonic and stereophonic reception than others under conditions of multipath propagation.

Richard Sequerra

Multipath distortion of FM stereo reception may cause more or less severe high-frequency distortion, inability to recover the 19-kc pilot, phase shift of the upper sideband of L-R signal, reversal of the left and right, and a further decrease in the signal-to-noise ratio. How can the listener recognize it? The only practicable solution we know is our oscilloscope tuning indicator, which presents dynamically the instantaneous deviation plotted against the produced AM components of a multipath signal passing through a non-phase-linear network.

Multipath distortion quite obviously affects FM monaural reception. This can be heard as high-frequency distortion. For a long while, all of us have been guilty of blaming radio stations for the poor quality transmissions, especially at the high frequencies.

As we increase the bandwidth of the transmitted signal for stereo reception, the problems created by multipath distortion rise at an astronomical rate. We feel that most listeners are likely to have multipath distortion of FM stereo.

Frequently, antennas and rotators with the proper lead-in wire such as a 300-ohm balanced shielded cable can make significant improvements in many locations. However, there must be some tuning device which will display the amount of multipath reduction when the antenna is turned. The receiver that has the most phase-linear response will be least subject to multipath distortion in all cases.

Unfortunately, little serious theoretical work has been done on multipath phenomena. We are all quite ignorant of the specific quantities and qualities. This is what we at the Marantz Co. think:

The long time delay between wanted and unwanted signals is the characteristic that produces the most severe multipath distortion, unlike the short-path delay that produces the most significant ghosts on television.

If the field strength of the unwanted signal or signals can be made sufficiently low relative to the wanted signal, we will have very low distortion. With the use of the antenna, rotator, proper balanced lead-in and a linear phase i.f. system, we can achieve in almost all cases a very significant improvement in the sound and insure the maximum of stereo separation for a given receiving situation.

Daniel von Recklinghausen

Multipath distortion is the result of multipath interference. The transmitted signal reaches the antenna simultaneously over direct and reflected paths. These signals add in phase and in amplitude, and cause undesired phase and amplitude modulation.

The most obvious result of multipath interference is audible distortion that is more noticeable with the tuner in the stereo than in the monophonic mode. The amount of phase modulation is approximately proportional to modulating frequencies, and therefore the higher frequencies show more distortion than the lower ones. With an ideal tuner, only phase modulation will be noticed. The tuner's amount of amplitude rejection is a measure of its performance - at least 40 db is required for good performance.

If the tuner has poor AM rejection, distortion will be greater. The more multipath interference is present the more stringent is the job the tuner has to perform. A directional antenna will always reduce multipath distortion, but this may be difficult to do in cities where landlords may not permit outdoor antennas, or in areas where multiple reflections are too numerous to be eliminated by antenna directivity.

Distortion in the tuner is frequently blamed on multipath interference, but often may actually be a distorted signal produced by the station. The FCC permits distortion values between 2.5 and 3.5%, not including the amount of distortion in the original recorded material. Therefore any added distortion due to multipath may be unnoticeable because it is relatively small. For best results, a tuner engineered for multiplex operation should always be used. Some earlier tuners were not engineered for that, and show a fair amount of high-frequency distortion.

Separation may be affected to some slight degree by multipath distortion but generally the amount of distortion is considerably higher in the audible degradation of separation.

Summary

It would appear, then, that while agreement on many of the phases of multipath distortion is not complete, there is a general tendency to agree that the remedy lies in an excellent directional and well matched antenna, and that there may be possibilities of further improving the situation by attention to a number of other factors. Especially, excellence in tuner design may have an important effect.