The motion picture projectionist (Nov 1929-Oct 1930)

20 Motion Picture Projectionist October, 1930 ki i 9 HO "O o A a ft / \ **■ Pf V A \ ~s / W- y^\ •\ k < ^s V ^ f \\ -10 ■^ w^ V ^^ \ ' 3 0 s 9 / 30 300 300 ',ooo 30 90 6000 '0.o< FREQUENCY IN CYCLES PER SECOND. Fig. 2. Showing the relative performance of two loud speakers of the cone type. The superiority of the one with the flatter characteristic is obvious. tion may be effected by means of networks to be placed in the transmission lines connecting the various elements of the system — but most certainly the desired effect cannot be attained through the use of loud speakers having a response deviating from the truly "flat." Function of the Speaker In Fig. 2 we show the response curve of an excellent loud speaker in comparison with one having prominent resonant effects. By "resonance" we have reference to any tendency of the speaker elements to pick up and sustain vibrations of a particular frequency. More correctly, a speaker is not a reproducer, but is a transducer, as it is employed in the change-over of electrical energy to mechanical energy. The rather close interlocking of the electrical and mechanical characteristics of the device give rise to minor resonant effects which account for the "hill and dale" formation of the curve. Just how and why these inter-relationships occur is a matter of interest only to the speaker designer. Returning to the opening paragraph of this paper we might mention the fact that certain of the effects apparent in the curve of the poor speaker might be corrected by a careful design of the amplifier characteristics. With speakers of good quality now available it is of little use to employ poor ones with still poorer amplifiers so as to effect an overall curve in which the deficiencies of one are compensated for by opposite deficiencies in the other. These curves have been plotted in decibels or TUs. In order to get a more accurate picture of just what is occurring it is sometimes better to plot a curve in "per cent response," the response at some mid-range frequency being taken as 100 per cent. This gives a more distorted curve with alarming aspects where there are truly marked deviations from the normal. Mr. Reichmann takes exception to the use of the "decibel" in plotting curves, but inasmuch as the unit is based on a system quite closely allied with the actual conditions governing the effect of sound wave pres sure upon the human ear, the present writer can find no fault with its acceptance as a standard. The decibel represents a power ratio between two quantities and we may readily convert sound, mechanical or electrical power relations into decibels without going through complex mathematical calculation. It is well to remember that the human ear can barely notice a deviation in sound intensity amounting to two dbs. A falling off of two such units corresponds to a loss of 20 per cent when viewed in the light of the other system. In order that reproduction should be perfect with regard to relative power between frequencies it is essential that a sound reproducing system should be flat within two dbs over the entire range — or within 20 per cent over the range where the use of the other system is deemed preferable. Relative Current Phases There is another type of distortion in which the term "flatness" is employed. This has to do with the relative "phase" of the currents. It is a well known fact that dynamic reproducers not of the "horn" type require baffle boards if the lower frequencies are to be reproduced. This is because of the fact that the sound waves from the front of the diaphragm have their greatest relative pressure at the same instant at which those from the rear have their minimum pressure. If the two waves are in such relation one to the other, they are said to be out phase by 180 degrees, and if superimposed would cancel each other. This is exactly what happens unless the air path between the front and rear of the diaphragm is lengthened by the use of a large panel or baffle board. If the panel is not used the lower frequency waves from the two sides of the diaphragm meet in opposite phase and cancel out, killing the low frequency response of the speaker. In horn type speakers where more than one unit is coupled to each horn, they must be connected in phase or the sound will be completely cancelled out by the conflict between the waves from two units which have their dia phragms vibrating in opposition. — even though the vibration of each is violent. Years ago Hemholtz, whose work in the field of sound is well known to all who have studied elementary physics, stated that the human ear could not differentiate the relative phase of sound waves. Recent developments have proven this to be untrue, and inasmuch as Mr. Reichmann must be willing to admit the rather self-evident truth of the things we have already noted, this last premise must be that which he terms "definition." When electric currents traverse long transmission lines or complex electrical networks some frequencies are propagated at a greater speed than others. This effect does not become apparent so long as the curve of "delay" or phase displacement as plotted against frequency is a straight line. Should these deviations fail to be a linear function of the frequency a blurring or masking of the speech or music will result. In other words, if there is to be any difference in the time of propagation for waves of varying frequency this difference must assume a practically straight line when plotted against frequency. It is easy to see that waves might readily be overtaken by others traveling at a faster rate and that the complex wave form of speech or music would be altered in the process. It is not usual that trouble due to this effect occurs in apparatus such as we are considering. The major difficulties of this nature occur in long distance program circuits utilized in national radio broadcasting. The effect may be readily measured for long transmission lines, compact networks, and for devices which represent a complex electrical structure such as dynamic reproducers. Correcting measures take the form of networks or equalizers designed to fit the particular case or a re-design may be necessary in order to retain desirable characteristics while compensating for the delay effect. I merely mention this effect in passing as it is of far too complex a character for the lay mind — -indeed, few engineers outside the larger organizations are at all familiar with the effect or its cures. Introduction of Resonance When a circuit is resonant at any particular frequency — and I refer to mechanical as well as electrical circuits— there is a tendency to sustain that particular frequency long after the true current has been removed. This is particularly apparent in circuits where the low frequency response has been aided by resonating the primaries of the transformers in the interstage circuits. This amounts to the same type of phase distortion noted above, as currents of the resonant frequency are carried over after the true tone has passed, thus distorting the entire output. If you will listen carefully to the output