|Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting
Departments | Features | ADs | Equipment | Music/Recordings | History
by Leonard Feldman
IF YOU ARE a regular subscriber to Audio, you may I remember that in the January 1973 issue I wrote an article in which I pointed out how obsolete the tuner measurements we use in describing FM products had become since their original issuance by the Institute of High Fidelity way back in 1958. In that article, I also suggested a number of additional measurements that I thought should be published by manufacturers, as well as the revision of a few existing specifications and their method of measurement.
Since then, the IHF has been engaged in the preparation of new proposals for measurement standards, and the first draft of Tuner Measurement Standards is currently being analyzed by the Board of Directors of that organization, as well as by members of the sub-committee that had volunteered to come up with such new standards. As is true of many industry organizations, trying to agree upon standards is a difficult task. Comments on the first draft of the proposed tuner standards have ranged all the way from "It's not tough enough" to "It's much too complicated and there are too many measurements called for." There is, however, unanimity of agreement that new standards are needed and that new emphasis must be given to stereo performance measurements. The old 1958 standards were confined to monophonic performance only--simply because stereophonic FM did not become a commercial reality until 1961--three years after the standards were issued! Here, then, is a summary of the measurements, old and new, that would be required, if a manufacturer were to go along with the new proposal for Tuner Measurement Standards.
In the old standards, this spec combined the effects of residual noise and residual harmonic distortion in one meaningful number, as illustrated in Fig. 1. With 100 per cent modulation applied at an audio frequency of 400 Hz, a suitable 400 Hz null filter was inserted between the output and the output meter and signal input was reduced until a difference of 30 dB was observed between readings with and without the filter in place. The number of microvolts applied to the antenna terminals to fulfill this requirement was then called IHF sensitivity, or "least usable IHF sensitivity." In Fig. 1 the value for this measurement is 2.0 microvolts.
Today, variations in this reading are minimal. While manufacturers continue to quibble over whether their product achieves an IHF sensitivity of 2.0 µV, 1.9 µV or 1.8 µV, from the consumer's point of view there is little audible difference between the performance of tuners at such low input signal levels. No high fidelity enthusiast is likely to be content with a signal-to-noise ratio of 30 dB today, nor is harmonic distortion of 3 percent a satisfactory figure by today's standards. The newly proposed measurement standard for tuners introduces two new measurements to supplement this all but meaningless number. A "50 dB quieting sensitivity" specification has been proposed in recognition of the fact that a 50 dB signal-to-noise ratio constitutes reasonably good listening quality by today's standards. As seen in Fig. 1, the reading for this new measurement, in this example, turns out to be about 5.0 microvolts. The new measurement also provides the prospective purchaser with an indication of how rapidly, or steeply, the quieting curve approaches its maximum value, which in this case turns out to be about 65 dB, a reading identified as the ultimate S/N ratio in both the old and newly proposed measurement specifications.
Since the new measurement does not combine residual distortion with residual noise, a separate measurement should be made of total harmonic distortion when the signal applied is that required for the 50 dB quieting result. A plot of THD versus signal input is shown in Fig. 2 and at 5 microvolts the hypothetical tuner produces about 1.2 percent distortion, still measured with full modulation applied to the r.f. signal.
In the interest of uniformity, it has been suggested that 1000 Hz be used as the modulating frequency for all single tone measurements instead of the 400 Hz previously used. At this frequency, the de-emphasis characteristic having a time constant of 75 microseconds will influence the reading by about 1 dB, but this is considered negligible, and since it has become standard practice to quote stereo separation and other related measurements referenced to 1000 Hz, the switch to this frequency for a majority of required measurements would simplify procedures somewhat.
Most manufacturers have, up to now, listed distortion for a single mid-band audio frequency. It is now suggested that frequencies of 100 Hz, 1000 Hz, and 7.5 kHz be used to measure THD and that at least the results of these three measurements be listed in the specifications. The curve of Fig. 3 is a complete plot of THD in monophonic mode versus frequency and the three required points are designated, both for the THD at 50 dB quieting sensitivity and for the "ultimate" THD readings, one of which was required to be published in the old standards.
While measurement of capture ratio was a requirement of the older standards, it has been found that capture ratio varies considerably with input signal strength on many tuner products. The newly proposed standards therefore require that the results of this measurement be published for input signal strengths of 100 microvolts as well as 1000 microvolts.
A completely new measurement applicable to the monophonic portion of the tuner is "adjacent channel selectivity." The older Standards outlined the means for making this measurement but required only that alternate channel selectivity figures be published. Today it is not at all uncommon to detect interference of the desired channel by an adjacent channel displaced in frequency by only 200 kHz rather than 400 kHz, and so both measurements of selectivity should be stated, and each should be properly identified.
There has been a tendency on the part of manufacturers to abbreviate the "alternate channel selectivity" measurement and to simply call it "selectivity." AM suppression, or the ability of an FM tuner or receiver to reject AM modulation of the received r.f. signal, was a required measurement in the case of the older Standards.
Only one measurement was required, with an r.f. input signal strength of 100 microvolts. Again, recognizing that this parameter varies considerably with input signal strength on most tuners, it has been proposed that the results of this measurement be presented or published for input signal strengths of 100 microvolts and the 50 dB quieting sensitivity value of microvolts determined in the earlier measurement.
It should be noted that the AM suppression capability of a product is directly related to its response to multipath signals and the distortion and other adverse effects noted in the presence of such reflected signals-particularly in the case of stereo reception.
As noted earlier, the old tuner standards did not concern themselves with measurements of stereophonic performance. The newly proposed standards attempt to rectify this situation by adopting and standardizing many of the measurements that some manufacturers have voluntarily published in their advertising literature over the past few years. In the high fidelity component field at least, much of the FM listening done is in the stereophonic mode. Signal-to-noise ratios in this listening mode are considerably poorer than when listening to monophonic transmission because of the greater bandwidth employed by the stereophonic composite signal and because of the AM modulation used to modulate the suppressed 38 kHz sub-carrier used.
The newly proposed standards require that "least usable sensitivity," measured in much the same manner as it was in monophonic units, be measured for stereophonic reception a well. In order that residual carrier products not be included in the residual noise and distortion reading, this measurement (and several related stereophonic measurements) should be made with a suitable band-pass filter inserted between the outputs of the stereophonic tuner and the metering instruments and null filter or distortion analyzer. The two curves plotted in Fig. 4 show a typical comparison between monophonic least usable sensitivity and stereophonic least usable sensitivity.
For the reasons stated earlier, the new "50 dB quieting sensitivity" measurements should be repeated for stereo reception and the total harmonic distortion observed at that quieting sensitivity should also be measured and published.
In general, harmonic distortion at almost every signal input level is higher when operating the tuner in the stereo mode than when it is operated monophonically. A typical pair of curves comparing distortion for monophonic and stereophonic operation of our mythical tuner is plotted in Fig. 5 for frequencies of 100 Hz and 7.5 kHz (in the case of monophonic), or 5 kHz in the case of stereophonic. The lower 5 kHz figure is chosen for the highest frequency to he measured in stereo because many tuners with less than perfect multiplex decoding circuitry often produce sizable "heats" between high-frequency modulating frequencies and internally generated 19 kHz and 38 kHz pilot and subcarrier signals. These beats, while not truly in the realm of "harmonic distortion" would, nevertheless, be summed and read on the meter of a typical distortion analyzer. If readings of harmonic distortion at higher frequencies than the 5 kHz recommended for the stereophonic mode of a tuner are desired, the only practical way to do this is with a spectrum analyzer with which actual harmonic contributions related to the fundamental desired audio modulation can be individually measured and summed up to yield a meaningful total harmonic distortion figure.
In any case, by providing distortion figures at the three frequencies suggested for mono and stereo, the customer will be provided with a better basis of comparison of this quality than if THD is quoted for mono only-and at a mid-frequency at that. Again, it is suggested that distortion figures for stereo operation be provided at the 50 dB quieting sensitivity signal input and at 1000 microvolt input.
Although a statement of frequency response was required by the old standards, a separate stereophonic frequency response measurement and specification should be recorded and published for stereo operation, since the results will often differ from those obtained and recorded for monophonic operation.
While most manufacturers of stereophonic tuners and receivers have been publishing the separation capability of their stereo circuitry, statements of this parameter have been largely confined to the dB figure measured at a midband modulating frequency of 1000 Hz, where separation is apt to he greatest. The newly proposed standard requires that separation capability be stated for frequencies of 100 Hz, 1000 Hz, and 10,000 Hz or that a complete plot of separation, such as that shown in Fig. 6, he made of this characteristic.
The three required figures in this illustration would be 2S, 40, and 22 dB, for 100 Hz. 1 kHz, and 10 kHz, respectively.
When a tuner is operated in the stereo mode, its ability to reject high frequency carrier signals is important for at least two reasons: A large amount of residual 19 kHz signal observed at the output of the tuner may be amplified by the high fidelity component amplifier with which it is used and may cause harm to high-frequency drivers in loudspeaker systems even though the listener may not be bothered by the presence of such a signal. Multiples of 19 kHz or residual products of the 38 kHz restored subcarrier may also "beat" with bias frequencies used in tape recorders and may therefore severely affect tape recordings of FM broadcasts made by the user. Accordingly, the new standards offer procedures for measuring rejection of such subcarrier signals by the product and define the manner in which such rejection capability is to be stated in published specifications.
Finally, standard means of measuring SCA rejection are proposed and a statement of this rejection capability, in dB, is made mandatory in published specifications.
The old standards required that certain specifications be published for a "minimum" description of product performance and that additional specifications of somewhat lesser importance be included if the specifications were to be considered a "complete" statement of product performance.
The newly proposed Tuner Measurement Standards have followed this same procedure and there are 10 specifications required to be published for "minimum" product description, as are shown in Table I. In the case of monophonic performance, some specifications, as noted in the Table, remain essentially unchanged from the old standard. Those listed as "revised" are specifications in which the method of measurement or the published requirements have been altered or increased, while the dash notation means that the particular specification is not applicable to monophonic performance. Totally new measurements, not previously called for in either the stereo or monophonic categories, are so noted.
Secondary specifications required for a "complete" presentation of product performance are shown in Table II. To fully specify the performance of a stereo tuner product, it would now be necessary to publish 14 specifications relating to monophonic performance and ten specifications relating to stereophonic performance.
It should also be noted that the old standards described measurements for AM tuners, or the AM section of combination tuner-receivers. The new standards make no new proposals with regard to these AM measurements and required published specifications.
As mentioned at the outset, comments have ranged from one extreme to the other. It has been suggested by some that recent findings in the field of psychoacoustics show that a simple measurement of harmonic distortion does not indicate how good (or how poor) a tuner will sound and that it is more important to specify what the harmonic contribution consists of. It is well known that high order harmonics are more disturbing than, say, simple third harmonic contribution which is at least musically related to the program material.
On the other hand, if one takes into account the de-emphasis characteristic of FM tuners and receivers, the 7th harmonic of 400 Hz (12,800 Hz), for example, will be so far down that even if present, its contribution is not likely to alter the perceived program significantly.
It has also been suggested by some that, "A few meaningful numbers will be easier to understand than the many measurements suggested by this first draft." No doubt true, but how much further can the specifications be "boiled down" and still provide the knowledgeable customer with a basis of comparison between competing products? And if some measurements are to be eliminated from the published specifications, which ones should be dropped? On the whole, however, comments thus far received have been helpful and the second draft, now in preparation, will reflect many of these suggestions, most of which have to do with the actual method of measurement rather than with the substantive information to be derived from those measurements. It is to be hoped that the second draft will be granted rapid acceptance by the membership of the IHF. Upon its acceptance, the next job will be to formulate new amplifier standards--a task which non-industry agencies such as the Federal Trade Commission and local consumer affairs bureaus have assumed and one which, in my opinion, the high fidelity component segment of the home entertainment industry is eminently more qualified to accomplish.
(Editor's Note: Much of the material covered by Len Feldman in this article was presented by him as a paper at the 46th Annual Convention of the Audio Engineering Society in September, 1973. In this article, Len has included some additional thoughts on the subject of tuner specifications, in the light of comments he a received following that original presentation.)
(Audio magazine; Jan. 1974)
= = = =
Prev. | Next