Full Text
Advances in the Transmission of Images over a Distance1
F. Schröter.
The essence and possibilities of applying telegraphic transmission of images are assumed here to be known.2 Since, even with the fullest utilization of all cable lines and of the entire range of ether waves, for each transmission in principle only one wire and one carrier frequency are available, all future telegraphic transmission of images is connected with the following scheme: transmitter and receiver possess parts moving synchronously and with identical phases; for the most part these are drums rotating and moving along an axis, for scanning and reproducing the transmitted image point by point. The conversion of a series of spatial dark and light surface elements into temporal oscillations of the transmitter current corresponds, in the receiver, to the reverse conversion of changes in current into light tones. The latter, owing to synchronism, acquire coordinates corresponding to the conjugate elements of the image being transmitted.
Various methods.
Three methods of scanning are essentially distinguished: 1) contact, 2) relief, and 3) illumination (phototelegraphy). The first method is suitable only for black-and-white images and is based on converting the light oscillations of the original image into current oscillations, which is obtained in the simplest way by copying with an insulating mass on a conducting surface (Bakewell, Caselli, Korn, Dieckmann, and others). This surface is gradually traversed by a contact stylus sliding over it.
The other two methods are applicable both to black-and-white and to colored images. In the second method a relief is used, corresponding to the transitions of tones in the original image, over which (the relief) a microphone lever contact is passed, transfor—
causing changes in pressure to be transformed into oscillations of current. In the third method, developed in 1904 by Korn, in the form of his “selenium transmitter,” the least inertia is obtained; in this method, over an image reproduced on a transparent positive slide, the sharp end of a cone of light is gradually passed. The rays penetrating through the drawing, depending on the degree of its blackening, fall upon a selenium element whose resistance changes according to the degree of illumination. This method of exciting the current of the transmitter was subsequently improved (Rosenthal) by replacing the selenium with a photoelectric cell (Elster and Geitel). It was thereby shown that the circuits used to compensate for the inertia of selenium (Korn) are unsuitable at recording speeds above a certain limit.
Photoelectric cells filled with gas1, used for obtaining a discharge current of sufficient strength, are also not entirely free from inertia, in the physical sense of the word. However, within the limits of frequencies encountered in practice (\(>100\,000\) periods—the upper limit of telegraphic image transmission), they may be regarded as energy converters in which changes of current follow changes in light intensity without appreciable delay. The newest methods of transmission (Bell Telephone Company, Telefunken-Carolus, Jenkins, Freund, and others) therefore all make use of photoelectric cells. For the receiver, almost without exception, photographic reproduction of the images has been introduced by means of a moving spot of light, the brightness of which is regulated, with the aid of a relay, by the oscillations of the incoming current.
During the last two years, telegraphic image transmission over pupinized wires (telephone cables) has been introduced in practice. For example, in the United States of America, over a line more than 1000 kilometers long, the Bell Telephone Company method is used (scanning by means of a photoelectric cell, variation of the light in the receiver by means of a special type of string galvanometer); then between Paris and Strasbourg, and also between Lyon and Marseille, by Belin’s relay method (the receiver recording the light is an electrodynamically rotating mirror). The most expedient method of transmission proved, in cable communication, with the Bell Telephone Company method, to be the use of a carrier wave of 1400 periods, upon which modulation of the brightness-distribution current is superimposed. A surface of \(13 \times 18\) cm, with a scanning mesh of \(1/3\) mm, corresponding to an image element of \(1/2\) mm\(^2\), is transmitted in 5–7 minutes. Belin’s method uses a finer mesh, corresponding approximately to \(1/25\) mm\(^2\), and therefore requires more time. The sharpness of his images is outstanding.
ADVANCES IN THE TRANSMISSION OF IMAGES OVER DISTANCE
In such telegraphic transmission of images over wires, with apparatus practically free from inertia, the transmission speed is determined and limited by the degree of loading, which for the most part corresponds to the needs of satisfactory telephone transmission. If the width of the image to be transmitted, in mm, is denoted by \(b\), its length by \(l\), the width of the image raster by \(r\), and the highest frequency well conducted by the wire by \(n\), then, under favorable conditions, the time required for transmission is expressed as:
\[ T = a \frac{bl}{120\, n r^2}\ \text{minutes}, \]
where the coefficient \(a\) may vary from 1.5 to 3, depending on the type of transmitter used and the transmission circuit. It is the larger, the higher the requirements imposed on the sharpness of the stroke and the fineness of the image. The scanning speed and the dimensions of the light spot on the surface of the transmitted image are selected in accordance with the values of \(T\) and \(r\).1
Studying the practical results obtained so far, we find that the transmission times given by Korn, the Bell Telephone Company, and others come quite close to the limits attainable in telephony over normally loaded wires. The highest modulation speeds occurring here, arising from the requirement that abrupt changes in light tones should not be transmitted as gradual transitions from one image element to another—that is, so that fine strokes can be reproduced distinctly—are satisfactorily achieved by the above-mentioned phototelegraphic instruments (string galvanometer or oscillograph). The need to seek a new light relay for transmitting modulation frequencies higher than 1000–1500 periods arose in the telegraphic transmission of images over telephone cables only insofar as there were lightly loaded sections that passed well \(>3000\) periods. Sometimes a similar case occurs in so-called four-core cables.
It always appears necessary, quite independently of the question of dynamic inertia, to eliminate the sensitivity to the influence of shocks inherent in electromechanical light recorders, and also to make imperceptible their own oscillations and changes in the zero position. From this point of view, the form of Kerr elements introduced by Karolus for regulating the brightness of light in the receiver represents a step forward for every kind of image transmission, regardless of the measure of time and the method of transmission: cable, overhead wire,
ether. Their chief advantages are explained by other considerations as well.
For economic reasons, telegraphic transmission of images cannot consist only of sending photographs and other images for personal or public purposes (illustrations, weather maps, signatures, checks, fingerprints, etc.). The very insignificant traffic so far on the French lines mentioned supports this consideration. The economic basis, and even the exclusive importance, of the exchange of information over long distances will become apparent if the method of pictorial (image) transmission of writing appears as a competitor to high-speed telegraphy (high-speed Morse, high-speed printing apparatus) in productivity and accuracy.
But this is, chiefly, a question of increasing the attainable speed, and from this point of view the Telefunken firm, together with Karolus, has developed wireless telegraphic transmission of images as a kind of high-speed telegraphy. To satisfy these conditions, telegraphic transmission of images must be capable of transmitting, in the highest case, about 1000 letters—200 words—per minute. To reproduce by telegraphic means such a number of characters quite distinctly on a given surface, it is necessary to transmit and then reproduce again in the receiver at least 200,000–250,000 picture elements, independently of their absolute size1. In this case it may happen that, in the successive scanning of the image, black and white alternate from element to element—for example, in a stroke one element wide. Hence it is calculated that the highest frequency of image dots transmitted by the transmitter, passed by the given wire, and reproduced by the receiving light relay, is, at 200 words per minute, 2000–2500 periods per second. With an increase in transmission speed, which is possible only with wireless or high-frequency telegraphy over wires, the highest modulation frequencies of the transmitter correspondingly increase.
Transmission of modulation frequencies of the specified order is carried out in wireless telephony (broadcasting) owing to the fact that speech or musical oscillations attain, in their upper ranges, oscillations of several thousand periods, which are so essential for the timbre of sound. The energy converters used—the microphone and telephone—must be so free from inertia as to satisfy the requirement of converting frequencies while preserving amplitude. As the successes of recent years show, the use of vibrating parts free from mass or burdened with only insignificant mass (ribbon microphone, condenser microphone, katodophone, ribbon
…(loudspeaker), offers great advantages over devices with massive membranes. It is quite natural that also in the telegraphic transmission of images at high modulation frequencies, in view of the further increase in speed, a change was made to an inertialess transmitter-and-receiver system. In this method, developed by Karolus together with the Telefunken firm, the principle of an energy converter free of mass is consistently applied.
Photoelements.
As was mentioned above, photoelements have already long since taken the place of selenium cells. The task in developing the light method of scanning for high-speed telegraphic transmission of images is to create especially sensitive photoelements and to make it possible to transmit an image by means of reflected light. Application up to now of the method of illuminating a transparent image has proved in practice too difficult. However, an attempt to use for transmission directly a black-and-white or halftone image, making use of the different intensity in the reflected light, succeeded for the first time when an entirely new system of photoelement was created. The latter is shown in Fig. 1 in the form of a ring; it has an axial aperture through which passes a sharp cone of light excited by a strong light source. Falling at the exit upon the image being transmitted, wound on a drum, it is reflected from its light areas more brightly and from its dark areas more weakly, after which it falls upon the photoactive potassium surface. The cone of the scattered light thrown back is shown shaded in Fig. 1. It is almost completely utilized by the photoelement, the front side of which must come quite close to the drum. In this way it is possible to attain the same sensitivity as by illumination through the image, and no amplification of the photocurrent is required to excite the transmitter. The appearance of the ring-shaped photoelement, developed in detail by Shriver and Richter in the laboratory of the Telefunken firm, is shown in Fig. 2. Its characte-
Fig. 1. Path of rays when using a ring-shaped element. a. Photoelement. b. Drum. c. Arc lamp. d. Lens.
Fig. 2. Ring-shaped photoelement.
characteristics taken for various applied voltages clearly reveal the relation between exposure and the current of the element (see Fig. 3). At insignificant applied voltages no ionization is yet detected; the curves in this case show a rectilinear dependence between the quantity of incident light and the photoelectric current. The positive ionization that arises at higher applied voltages influences the form of the curves. At the same time there is the danger of the occurrence of a silent discharge not controlled by the light.
Fig. 3. Characteristics of the photocell.
In order, while making the greatest use of the sensitivity of the element by increasing the voltage accelerating the electrons, to prevent the appearance of a silent discharge, grid collecting electrodes are selected of such dimensions that they extend along the entire potassium surface. The field then will be very uniform, and the spatial density of the positive particles arising in the region of the anode will be very small. Thus, in the transmission of black-and-white images, with oscillation of the high-frequency amplitude only between zero and the maximum value—where no transmission of intermediate brightness is required—the possibility was achieved of operating with element currents of the order of 1 mA. The latter are excited by means of light and are interrupted without noticeable lag. The limits of inertia of the Telefunken photocell lie far above 100,000 periods, despite the relatively high gas pressure.
Fig. 4. Image transmitter of the Telefunken-Karolus system, first experimental apparatus.
Figs. 4 and 5 show the experimental model of the telegraphic image transmitter used for the transmissions carried out later.
One can distinguish the feed drum, driven by a synchronous motor, and, in the foreground, the optical apparatus for recording. The cross-section of the light cone used for this purpose reaches only \(1/25\ \mathrm{mm}^2\) at the place where its point touches the image. Such fineness is required for the clear transmission of small
Fig. 5. Transmitter of telegraphic images, Telefunken-Karolus system, second model.
writing or ordinary magazine type. The connection of the transmitter for wireless telegraphic transmission of images is shown in Fig. 6. The photocell current is amplified by means of a resistance-coupled cathode amplifier and is fed to the modulation tube of a generator with independent excitation. A similar circuit is used by the Telefunken firm very successfully for radio transmission. The wireless telegraphic image transmitter is a kind of radiotelephone transmitter, as is revealed by the comparison given above, and contains nothing new from the standpoint of radio engineering.
The receiver circuit is shown in Fig. 7. In this, the amplification of high frequencies is omitted as an inessential element. The high-frequency oscillations received by the receiver antenna are demodulated by a detector. The low-frequency oscillations are amplified by a resistance-coupled amplifier until the voltage amplitudes obtained prove
Fig. 6. Schematic circuit diagram of a transmitter for radio-telegraphic images.
sufficient for exciting the Karolus elements, which serve as light relays. A photographic recording apparatus and a distinguishable Karolus element in the foreground are shown in Figs. 8 and 9. Synchronous motors, similar to the motors of the transmitter, and the corresponding drum for the photographic film are visible. The latter usually runs in a dark space, protected from interfering light; only rays of variable brightness, excited by the receiver currents, fall upon it.
Fig. 7. Schematic diagram of a receiver of radiotelegraphic images. Э — Kerr element.
The Karolus Element.
Here we are concerned with the most essential link of the general transmission chain, to which, above all, the successes in increasing the speed of telegraphic transmission of images are due. The Karolus element makes it possible to dispense with electromechanical relays, whether these be string galvanometers, oscillographic installations, mirrors set in motion electromagnetically, etc. The relays mentioned at the beginning, with massive parts set in motion by electricity, do not manage to follow the high frequencies used in the wireless transmission of handwritten characters in facsimile form. Korn’s light-excitation arrangement, based on the principle of Einthoven’s string galvanometer, has a limiting frequency of approximately 2000 periods.
Although Korn indicated that, with its aid, considerably higher frequencies could be attained, the practical applicability of such relays is unreliable.
Fig. 8. Image receiver, first experimental apparatus.
In any case, operability at frequencies above 2000 periods is obtained at the cost of a correspondingly higher, unusual in amplifier technique, expenditure of current. The light relay of the oscillograph (Petersen, von Mihaly) can operate at the frequencies mentioned, according to the experience of measuring technique, likewise only with a high expenditure of current. Also disadvantageous is the slight utiliza-
… of the light source, owing to the required small dimensions of the mirror and the necessity of using lens sets with large focal lengths; moreover, all these devices exhibit the instability already mentioned. Free from it are gas discharges excited without the motion of massive parts, which can be used in the form of lamps with a silent discharge or tubes with a positive column as the light source. Up to now it has not been possible to create a gas lamp with great luminous intensity, operating with a sufficiently small consumption, at a film speed of approximately \(1.5\)—\(3\ \mathrm{m/sec}\), as required in high-speed telegraphy. The ultra-frequency lamp
Fig. 9. Receiver of telegraphic images, Telefunken–Carolus system, second model.
(Fott Engl, Massolle), using negative cathode light, at a speed of \(0.5\ \mathrm{m/sec}\), takes approximately \(10\)—\(15\ \mathrm{W}\) and several hundred volts of voltage to blacken the film. This leads to inconvenient conditions when receiving at different speeds. In view of all these shortcomings of modern light relays, Carolus systematically investigated magneto-optical and electro-optical phenomena in polarized light, which in themselves are free from inertia owing to the absence of mechanical accelerations. In doing so he found that the electrostatic Kerr effect—the double refraction in the field of a condenser—is most suitable for making a light relay.
Using a substance having electric double refraction, for example carbon disulfide or, better, nitrobenzene, as a transparent dielectric between the plates of a condenser, he found that a ray polarized at an angle of \(45^\circ\) with respect to the direction of the field is split into two equal components—perpendicular and parallel to the field. The latter emerge with a phase difference depending on the applied voltage and on the path traversed between the plates, and therefore can interfere in the plane of transmission of the polarizing prism. If both prisms (Nicols) are crossed at an angle of \(90^\circ\), then the field of view without voltage will be dark,
but it becomes illuminated as the potential of the plates is increased. It is possible, of course, to work also with the nicols in a parallel position, and then the electric field, increasing, extinguishes the light.
As early as 1890 Setton advised applying this effect, discovered in 1875, to telegraphic vision over wires. However, this proved impracticable until the beginning of Karolus’s work, since the Kerr effect required high voltages. Since 1911 the technique of amplifier tubes has made it possible to attain extremely high amplitudes starting from the small voltages that occur in wireless reception or in long-distance wire reception. Nevertheless, under this condition alone it is impossible to regulate the brightness of light on the basis of the Kerr effect. Only Karolus found the working conditions under which an extraordinarily strong action on the brightness is obtained, by connecting a voltage amplifier with a Kerr cell. The constant voltage used by Karolus, applied to the electrodes of the condenser, keeps nitrobenzene, which is notable for its high Kerr constant, in a state of such high insulating capacity and such insignificant dielectric losses even at very high frequencies that, in practice, excitation is produced without expenditure of work. In this way it was possible to reduce the distance between the electrodes (without optical detriment) to such an extent that the required high field strengths were obtained by means of relatively weak voltages.
For the telegraphic transmission of images at a film speed of about \(3\ \mathrm{m/sec}\), operation proceeds, according to Karolus, with a distance between the Kerr electrodes of approximately \(0.2\ \mathrm{mm}\), with a constant preliminary voltage of \(200\text{--}400\ V\) and a superposed excitation amplitude of \(100\text{--}200\ V\). Since the Kerr effect, according to the equation
\[ \delta l = BF^2 l, \]
(where \(\delta l\) is the path difference of the two component rays, expressed in \(\mathrm{mm}\), \(B\) is the Kerr constant, \(F\) is the strength of the electric field in \(V/\mathrm{cm}\), and \(l\) is the length of the path of the light in the field, expressed in \(\mathrm{cm}\)), increases proportionally to the square of the field strength, a weak increase of \(F\) can compensate for a considerable shortening of \(l\) and, in this way, considerably reduce the light-absorbing layer of nitrobenzene. In the Karolus cells of the new system, operation proceeds with a light-path length of only a few \(\mathrm{mm}\), and the absorption losses amount to only a few hundredths of the incident light. In this respect Karolus’s method differs extremely favorably from methods employing the Faraday effect (rotation of the plane of polarization in a magnetic field1).
ADVANCES IN THE TRANSMISSION OF IMAGES OVER A DISTANCE
Since the Kerr effect increases proportionally to \(F^2\), it is possible to raise the sensitivity of the Karolus element to a high value by means of an additional direct-current voltage. Here, too, lies a very substantial difference as compared with the Faraday effect. In the latter, the angle of rotation of the plane of polarization is proportional to the strength of the magnetic field \(H\). Preliminary magnetization does not change the magnitude of the rotation caused by a definite \(\Delta H\), corresponding to an increase of the exciting current \(\Delta I\). On the contrary, in the Kerr effect, an increase by \(\Delta V\) of the voltage \(V\) exciting the field gives an increase of polarization proportional to:
\[ (V+\Delta V)^2 - V^2 \approx 2V \cdot \Delta V . \]
Consequently, the action of the excitation is proportional to the product of \(\Delta V\) and the additional direct-current voltage \(V\). Strictly speaking, this is of significance for monochromatic light. In reality the phenomenon is more complicated, since the excitation corresponding to a given \(V\), with the white light ordinarily used, is different for waves of different length. The phase differences that intensify or weaken the brightness, corresponding to half a wavelength, are obtained first for light with short waves, and then for light with long ones. If the light is to act on the photosensitive layer, then the excitation is obtained chiefly by blue and violet rays (chromatic excitation).
Fig. 10 and 11. Kerr element for transmitting images according to Karolus.
Figs. 10 and 11 show the Kerr element for the telegraphic transmission of images, according to Karolus. In a metal reservoir with a glass window, nitrobenzene is placed as a transparent dielectric between the parallel plates of a condenser, forming an optical slit. The upper electrode is inserted with the aid of an insulating rod of elephant ivory; the lower one is connected with the metal of the reservoir. Before entering, the light passes through a polarizer; on leaving, through an analyzer. Both of them are fastened directly in front of and behind the metal reservoir.
Fig. 12 shows the excitation curve of the Karolus cell under white light. The light intensity is determined by a photocell placed behind it, the current of which, with the applied voltage kept constant, is measured by a galvanometer. In halftone images it is necessary to calculate the preliminary voltage in such a way as to maintain the excitation on the middle, almost rectilinear part of the curve. With overexcitation, of course, extinction of the light occurs as soon as the phase differences exceed half a wavelength. By using monochromatic light, one can in this way establish several
Fig. 12. Current of the photocell as a function of the regulating voltage of the Kerr cell. (Arbitrary units).
Fig. 13. Characteristic of the Karolus cell for monochromatic light.
orders of complete extinction and renewed illumination corresponding respectively to even and odd multiples of a half-wavelength. Only the attainment of the breakdown limit stops any further increase of the voltage. Since the Kerr effect depends on the square of the field strength, the excitation line becomes ever steeper, and the highest values move ever closer together. This may also be applicable in known cases for a highly sensitive Kerr relay (Fig. 13)¹.
The high speed of the film, possible with the Karolus cell, requires especially precise synchronism during image transmission. Previously this problem was extremely difficult; in particular, in wireless communication atmospheric disturbances in reception often shifted and interrupted the auxiliary pulses transmitted for the purpose of maintaining synchronism. In view of these shortcomings of the auxiliary-pulse method, in the system described here efforts were made toward purely local regulation of the rotational speed of the transmitting and receiving apparatus. This problem was solved impeccably by applying tone exciters as invariable frequency exciters. The synchronism of motion became independent of the distance between stations and of the possibility of transmitting auxiliary pulses with the desired accuracy. The local beat indicators were aligned with one another before transmission by means of a simple stroboscopic method and remained constant for a long time with an accuracy of 1 : 100,000.
¹ These curves were recorded and calculated by Dr. W. Ilberg in the Karolus laboratory at the Physical Institute of Leipzig University.
Results of the Experiments
With the aid of the “Telefunken” photocell and Karolus’s light-exciter element, it is now possible, in the Berlin–Leipzig transmission, to obtain very considerable speeds. Over an overhead line of 3 mm bronze wire an image of area \(10 \times 10 \text{ cm}^2\), with an image-element size of \(^{1}/_{25}\text{ mm}^2\), was transmitted in 1.5 minutes. By wireless transmission, for images of the same dimensions and fineness, a time of 20 seconds was attained, at a wavelength of 850 meters. But this value is in no way a limiting one; by means of appropriate mechanical arrangements, when short waves are used, it can be reduced still further.
The times cited show that it has been possible to surpass the speeds used in high-speed telegraphy. A square decimeter contains, approximately, 200 words or 1,000 letters in small handwriting or ordinary newspaper type. At present, with ordinary high-speed telegraphy (high-speed Morse and printing apparatus), this number can in the best case be transmitted in 1–2 minutes. By photo-telegraphic transmission, the above-mentioned speed is at least tripled and, moreover, further practical advantages are associated with it. First, the effect of atmospheric disturbances of reception on the legibility of the characters is insignificant. This circumstance is based on the fact that incorrect records caused on the film by rapidly passing disturbances appear only in the form of very short and thin, for the most part dot-like, strokes. The width of the latter is very small compared with the thickness of the lines of the character. Each character is formed from a large number of photographic dot-like marks having large intervals in comparison with the disturbances, and the distribution of the latter never occurs in so periodically compact a manner as to distort completely the transmitted images of the character. In the usual method of transmission, the individual elements of a character are transmitted close one after another and therefore easily become intermixed and distorted by atmospheric disturbances. As a result, inquiries and secondary transmission are required, whereby the actual throughput is greatly reduced; this drawback, as indicated, is eliminated in the method of photo-telegraphic transmission.
This method receives still greater recognition in the rapid transmission of handwritten characters, when black-and-white images are in question: the antenna current in the darkened places always reaches from zero to its highest value, and therefore the range of action of telegraphic transmission is fully utilized. On the contrary, the transmission of images with gradual tone transitions requires the same operating conditions as telephony for the transmission of different amplitudes, and therefore their transmission range
does not exceed that of the telephone. In itself, the reproduction of the finest shades in the transmission of photographs, halftone images, etc., by means of the Karolus element is very perfect. Figs. 14–17 serve as proof of this. On the left the original image is always placed, on the right—the reproduced one.
Fig. 14. Transmission of a photographic portrait; left—the original image, right—the reproduced one.
In such high-speed transmission, special demands are made of the amplifiers. The latter must transmit very broad frequency bands, with the highest modulation frequency and the extent of the side bands arising on both sides of the operating wave1 being determined from the consideration that the limiting dimensions of light and dark may change every \(1/5\) mm.
The lowest modulation frequency may be very small if the brightness remains unchanged over a considerable extent, for example, over an entire line. These very broad frequency bands must be amplified independently of frequency, in proportion to the incoming amplitudes. For this purpose amplifiers with resistances are used. In this connection it is worthy—
Fig. 15. Transmission of a landscape photograph.
attention to the circumstance that the finite extent of the light spot during recording or reproduction eliminates the necessity of transmitting the upper frequencies. This follows from the following considerations. Suppose that the light spot is a tiny square with side length \(l\) and moves over the surface of the image parallel to one side with velocity \(v\) of the circumference of the drum. The brightness of the original image remains constant over at least the width \(l\) perpendicular to the motion. If the light spot passes over a sharply black stroke, then at the point of the jump the brightness is equal to \(\frac{dH}{dt}=\pm\infty\). But the current \(I\) lags even in instruments that are practically without inertia, i.e. with a sufficiently short transmitting wave. For in order for the current \(I\) to reach its new final value, a time \(t=\frac{l}{v}\) is required, calculated taking into account the extent of the light spot in the direction of motion; if \(l\) is equal to \(1/5\) mm and \(v\), for example, is \(3\) m/sec, then \(\frac{l}{v}=\frac{1}{15000}\) sec. This time is required by the current \(I\) in order to change linearly between its limiting values. It represents the duration of one half of the modulating oscillation.
Fig. 16. Transmission of an imprint.
Fig. 17. Transmission of a check filled out by hand.
Therefore the highest sharply defined frequency of modulation is \(m=7500\) periods. Shades of change in brightness at the boundaries of the stroke being imaged cause the appearance of all the higher frequencies of the Fourier series with fundamental number of oscillations 7500. Electrical reproduction of the image gives, in essence, only one fundamental oscillation1, and in order to obtain in the receiver all the shades of change in brightness of the original, one has to resort to artificial measures such as selecting the grid voltage in the amplifier tubes, the corresponding steepness of the tubes, etc. From this point of view, the element
The Karolus cell is more convenient than electromechanical light relays, since it in fact photographically reproduces all the electrical higher oscillations of the exciting voltage according to their magnitudes. It should be noted that the finite length \(l\) of the spot of the reproduced light, for its part, exerts an influence expressed in the broadening of the signs and the blurring of their edges. This, however, can be eliminated by precisely setting the “threshold of sensitivity” of the entire receiving apparatus. An exact theory of the sharpness of the edges of the image and of distortions due to the damping of high frequencies in phototelegraphy must clarify a whole series of important questions and, among other things, develop the question of the arising and cessation of oscillations in filter circuits and in the antenna, the law of blackening of the film, the influence of the apparent resistance on the magnitude
\[ \frac{dl}{dt} \]
during jumps in brightness; a detailed exposition of the question would take us too far.
The range of waves for the wireless transmission of images is the same as for waves of telephony, and lies, approximately, below \(4\) km, descending to short waves of \(10\)—\(100\) m.
The latter have been used with great success at night for work over long distances, for example, Nauen—Buenos Aires (about \(12\,000\) km). The range of short waves is at present still developing; principal attention is being directed to increasing the output of the transmitter and to the constancy of the frequency in accordance with the requirements of reception, as well as to obtaining daytime communication (at wavelengths up to \(20\) m, for the most part with the use of special antennas and reflector installations). It now seems not reckless to predict a great future for short waves in the field of telegraphic transmission of images. Reception in Europe of American telephone communications on a short wave is often quite distinct and so strong that, with the aid of Karolus elements, the modulation can be recorded without any special amplification. By means of short waves, in connection with the Karolus element, it is possible to achieve a very considerable increase in transmission speed, limited only by the sensitivity of the film. It may be considered attainable to transmit an image with an area of \(10 \times 10\ \mathrm{cm}^2\) in \(5\) seconds, with an image element of \(1/5\) mm. Technically this will become attainable by the use of special apparatus operating not with rotating cylinders, but allowing the photographing and reproduction of the image on a flat surface with the aid of special optical arrangements. The latter presents no difficulties, as has been shown by experiments with the Karolus element. In laboratory experiments with Karolus elements, photographs have already been transmitted over a distance in the course of \(5\) seconds. The attainment of favorable transmission times will lead to the realization of handwritten telegrams transmitted in the form of an image—“transoceanic radio letters”—and will bring about a revolution
in the modern forms of the telegraph. To the astonishing speed of this method is added the possibility of transmitting by telegraph, instead of ordinary handwritten signs, abbreviated ones, such as the Morse alphabet or stenographic signs. It has often been intended to maintain communication by high-speed Morse with the aid of short waves, transmitting one and the same telegram several times in succession with the greatest rapidity. With a known number of repetitions there is a probability of correctly deciphering each sign, despite atmospheric distortions. The realization of this task seems most favorable with the aid of the methods and apparatus of phototelegraphy, since in it the reproduced sign is the product of a proper series of separate impressions, repeated at large intervals of time. Consequently, one may imagine that a surface written in the Morse alphabet or printed will be transmitted as a black-and-white drawing. The speed in this case will be expressed in several thousand words per minute.
Television.
Photoelectric cells free from inertia and Karolus cells make it possible to go still further with modulation frequencies and to predict the early successful solution of the problem of seeing by telegraph. A photoelectric cell operates at a frequency above \(10^5\) periods/sec.; a Karolus cell—\(10^5\) periods without noticeable lag. For a satisfactory telegraphic image, regardless of its dimensions, the magnitude of a single element of the latter must amount to at least \(1/10000\) of the total area. Under this condition a distinct image is obtained, especially with moving processes. Experiments have shown that, for carrying out seeing by telegraph, it is sufficient to transmit the image ten times in 1 second. In this case the highest modulation frequency is equal to \(1\cdot10^4\) periods/sec. Such frequencies, according to the foregoing, can be transmitted both by photoelectric cells and by Karolus photo-relays. In this case, for transmission, only short waves are, of course, suitable, since the transmitting frequencies must be of a higher order than the upper limits of the modulation frequency. At \(5\cdot10^4\) periods/sec. and a transmitting wavelength of \(30\,m\), side bands only \(\frac{1}{2}\nu H\) wide are obtained, and the normal relation between the width of tuning and the sharpness of contour, known to us from wireless telephony, is realized. The necessity of several carrier waves, considered by many an inevitable condition of television, is not absolute with a raster equal, as is accepted, to \(1/10000\).
In television the images are received and reproduced according to the same scheme as in phototelegraphy, but with the aid of special, exceptionally fast-acting optical instruments, allowing, with good utilization of light, precise decomposition and
reproduction of the image. The separate parts of the transmitted image are, one after another, supplied so as to act upon the photoelectric element and cause in the latter current oscillations which, after sufficient amplification, modulate a short-wave transmitter. In the receiver, to the amplifier of the modulated oscillations there is connected a Karolus cell, which affects the brightness of the light beam before it enters the apparatus reproducing the image, in accordance with the values of the current sent from the transmitter. The chief difficulty lies in the question of the luminous intensity of the light relay, and not in the inertia of the latter. This is easy to see from the fact that, in the course of reception and reproduction point by point, each element of the surface is illuminated only for \(10^{-5}\) seconds and is then not illuminated for \(10^{-1}\) seconds. The physiological laws of the fusion processes arising on the basis of the inertia of the eye state that the total brightness perceived for one point under these conditions will amount to \(1/10000\) of the brightness given by the moving light spot. The instantaneous brightness of the light spot must therefore be exceptionally great, in order that the received image may come out sufficiently rich in contrasts. The Karolus cell makes possible the excitation of light intensities of considerable magnitude through the expenditure of very insignificant electrical work, and upon this rests its importance for application in television. At the present time no other light relay is known that operates under such favorable electrical and optical conditions. Dr. Karolus has succeeded, within a year, in obtaining with the aid of his cell very good telegraphic images that transmit all details and, despite this, are of high light intensity. The practical development of his apparatus will require some more time. At present efforts are directed toward the transmission of transparencies, and first of all motion-picture film, by means of the optical wireless telegraph. This preliminary solution of the problem of television does not present very great difficulties. The transmission of three-dimensional objects, in view of the relatively much smaller intensity of light when projecting images onto ground glass (the camera obscura for the photoelectric element), sets for amplification technique familiar problems. There can be no doubt, however, that an answer will soon be found for this problem as well, and that real television will be achieved in the near future.