United States Patent 3773977

An information signal train which may, for example, be an audio signal, is divided into equal time intervals and temprarily stored. Each stored element is read out in a random pattern so as to randomly "scramble" the arrangement of the elements as compared with their original arrangement. The "scrambling" of the elements is continuously monitored so as to prevent more than one element from being shifted to the same (new) time position and further to prevent any gaps in the transmitted signal train. The process is effectively reversed at the receiving end which is provided with a similar random generating and deciphering means operating in synchronism with the enciphering means at the transmitter facility. Shifting of the element positions may be confined to a group containing a predetermined number of elements or alternatively may be progressively shifted to new positions without concern for limiting shifting in time of elements to a group of a predetermined number of elements. Additional techniques may be employed to reverse polarity of selected elements or to randomly superimpose other signals upon selected ones of the elements in a random fashion.

Application Number:
Publication Date:
Filing Date:
Primary Class:
Other Classes:
380/37, 380/275
International Classes:
H04K1/06; H04L9/34; (IPC1-7): H04K1/06; H04L9/00
Field of Search:
178/22 340
View Patent Images:
US Patent References:
2547515Secrecy system1951-04-03Zenner
2453659Secret telegraph signaling1948-11-09De Bellescize
1953918Cryptographic system and apparatus1934-04-10Bellamy

Primary Examiner:
Borchelt, Benjamin A.
Assistant Examiner:
Birmiel H. A.
What is claimed is

1. A method of enciphered information transmission, in which the plain-language signals to be transmitted are split up into elements of equal length whose original time sequence is modified by interchange prior to transmission and restored after transmission by a reverse process of interchange, the elements interchanged (and the reverse process carried out at the receiving end) through a process of storage in an information store, characterized by the steps at transmitting and receiving ends of: (a) generating coinciding aperiodic cipher signals (w); (b) generating control signals (i FIG. 2) which are derived from the cipher signals which determine the storage position of the individual elements in the information store; (c) storing binary pulses, each in a multi-position store in which each position is assigned to an information element; (d) modifying the binary pulses in an irregular sequence by hunting pulses (h) in the sense that occupied positions are evacuated (or empty positions are occupied) in accordance with the state of said cipher signals at any given time, by moving from each position which has already undergone the aforesaid change in occupancy, i.e., has either been evacuated or occupied, to those positions which have not yet been evacuated or occupied, without modifying the occupancy of any position which has already changed (e) producing control signals (i) depending on these changes to determine the storage position of the corresponding element in the information store, so that automatic monitoring of the occupancy of the information store is achieved whereby, on the one hand, control signals which are required to avoid omissions of individual elements are ensured, and on the other hand control signals which lead to the repetition of individual elements, being suppressed.

2. A method as claimed in claim 1, characterized in that the positions in the information store (NS) are initially cleared; in that after each determination of a storage position an element is supplied to this position; and in that the extraction of the elements from the information store takes place in accordance with a progressive sequence.

3. A method as claimed in claim 1, characterized in that the elements are supplied to the storage positions in the information stre (NS) in a progressive sequence; and in that after each determination of a storage position, an element is extracted from it.

4. A method as claimed in claim 1, characterized in that binary pulses are stored in first and second independent stores in order to determine storage positions; determining the storage positions to which the elements are supplied by altering the binary data of pulses in said first store and altering the state of binary pulses in the second store to determine the storage positions from which the elements are subsequently extracted.

5. A method as claimed in claim 1, characterized in that the elements consist of sections of the unmodified plain-language signal.

6. A method as claimed in claim 5, in which the plain-language signal is converted into digital form by sampling and analog/digital conversion of the sampling pulses, characterized in that the elements each consist of a specific number of coded sampling pulses of the plain-language signal.

7. A method as claimed in claim 1, in which the plain-language signal is provided in analogue form, characterized in that the elements consist of a specific number of amplitude-modulated pulses which are obtained by periodic scanning of the plain-language signal.

8. A method as claimed in claim 1, characterized in that each storage position in the multi-position store (BS, FIG. 4), corresponds with a specific position in the information store (NS); in that the number of multi-position store positions corresponds with the number of elements in a group; in that at the commencement of element interchange in a group, all positions in the multi-position store operating as occupation register have the same occupancy, i.e., all are occupied or all are free; in that with any modification of the occupancy of a register position by a corresponding control signal (i), a corresponding position (corresponding, that is to this register position), in the information store is determined; and in that the number of the associated register position, which number is assigned to said hunting pulses, is formed in accordance with the rules of binary addition, from several pulses of the cipher signal.

9. A method as claimed in claim 1, characterized in that in the multi-position store (BR, FIG. 5) operating as occupation register, binary pulses are stored and advanced therethrough step by step; in that the content of the information store (NS) is likewise displaced vis-a-vis fixed tappings; in that the first stage of the occupation register always exhibits a first (state of) occupancy, corresponding to the "occupied" or "unoccupied" conditions; in that by means of hunting pulses (d), which depend at least partially upon the cipher signal (w) individual positions of the occupation register are converted from the said first to the said second state of occupancy; in that the hunting pulse operation in each case commences with the last stage of the occupation register and moves back step by step until a stage is converted from the first to the second occupancy; in that this change in state of a stage takes place in such stage still exhibits a first state of occupancy and if, furthermore, a blocking pulse (a) dependent upon the size of the signal, does not occur; and in that with each conversion of a register position from the first to the second occupancy, a corresponding control signal (d) determines a position (corresponding to this register position) in the information store (NS).

10. A method as claimed in claim 1 characterized in that the information store (NS) consists of several individual independent stores operating in synchronism (M1 -M6, FIG. 7); and in that the control signals for driving the associated individual stores are displaced by varying amounts, through additional registers (HR).

11. A method as claimed in claim 1, characterized in that the information store (NS) consists of several individual stores; in that the occupation register (BR, FIG. 7) consists of several individual registers; in that in the hunting operation, the individual occupation registers are successively tested as to their binary states; and in that each new hunting pulse commences with the particular individual register next in succession.

12. A method as claimed in claim 1, characterized in that each individual information store and each individual occupation register has several positions; and in that in the hunting operation a position in each occupation register is tested and in the ensuing hunting pulse a position in each individual occupation register is tested (FIG. 9).

13. A method as claimed in claim 1, characterized in that with changing directions of information transmission, recording and extraction in and from the information store change correspondingly, although the driving of the storage positions remains the same.

14. A method as claimed in claim 1, characterized in that the information elements are time-compressed prior to transmission, by increasing the rate of store extraction, and are expanded again at the receiving end by slowing the rate of store extraction.

15. A method as claimed in claim 1, characterized by additional enciphering of the information elements by reversal of selected elements in accordance with special cipher signals.

16. A method as claimed in claim 1, characterized in that the cipher signals (w) are generated from intermediate signals (s) by means of a cipher computer (SC, FIG. 12); in that in order to produce these intermediate signals, programme signals (u) are generated using programme generators (PS) which consist of shift registers with facility for feedback of output pulses to the input via logic switching circuits; in that several intermediate signals are extracted from selectable shift register stages through selector switches (FIG. 13); and said cipher computer (SC, FIG. 14) consisting of at least two shift registers in association with switches S1, S4), inverters (S2, S5) and contact breakers (S3, S6) which interrupt the pulsing signals, these switches being controlled by the intermediate signals (s).

The present invention relates to method and apparatus for transmitting data in enciphered fashion and more particularly relates to apparatus and method for enciphering information transmitted, in which the plain language signals to be transmitted are divided into elements of preferably equal time length whose original time sequence prior to transmission is modified by interchange and restored after transmission by a reverse process of interchange, the elements being in part at least interchanged (and the reverse process carried out at the receiving end) by dissimilar time shifts through the utilization of a storage process within an information store.


Methods of the type referred to hereinabove and devices for carrying them out are well known in the prior art as evidenced, for example, in Swiss Pat. Nos. 22,742 and 232,768. However, a very important provision in such systems employing an interchange process, and which is fundamental to the method of the present invention, is that there must be no repetitions and no omissions of individual information elements (i.e., signal sections); the interchanged sequence must contain all the elements of the sequence forming the original plain language signal (i.e., a speech signal), without any overlap between any two elements and without any gaps. With a periodic period of interchange, as used, for example, in the methods disclosed in the aforesaid patent specifications, the compliance with this condition is something which can be achieved of course without difficulty, by appropriate adjustments. This kind of periodically recurring interchange process, however, is unsatisfactory from the cryptological point of view since the periodicity as well as the cipher may be easily detected and reconstructed by unauthorized third parties who may receive the message. Consequently, it has also been proposed in the prior art (for example, see Swiss Pat. No. 220,056) that the interchanges be controlled by correspondingly prepared punched tapes. However, handling of tapes of this time introduces many attendant problems, along which are the necessity of a tape of great length, the need for synchronism between transmitting and receiving ends at the time of start-up and halting of the punched tapes, as well as the fact that the tapes are relatively fragile and are incapab e of being used over and over again.


An object of the present invention is to overcome these drawbacks. In accordance with the invention, this is achieved in that the transmitting and receiving ends are provided with coinciding aperiodic cipher signals which are produced by cipher signal generators and supplied to ancillary devices whereby special control signals are derived from the cipher signals which are utilized to determine the storage positions of the individual elements in the information store; in that additional devices each containing an additional storage means serve as an occupation register in which binary pulses are stored, each of which is assigned to an information element; in that by means of hunting pulses (h) which depend in part at least upon the cipher signals, the (state of) occupancy occupany of individual positions in the occupation register is modified in an irregular sequence in the sense that occupied positions are evacuated (or empty positions are occupied;) in that, when a hunting pulse tests a register position which has already undergone the aforesaid change in occupany, i.e., has either been evacuated (or occupied,) the hunting operation is continued for register positions which have not yet been changed in this manner, without further modification of the occupancy of any register position which is already changed; and in that, with any change in occupany, a control signal (i) is produced which determines the storage position of the corresponding element in the information store, so that automatic monitoring of the occupancy of the information as stored is achieved, whereby on the one hand, control signals which are required to avoid omissions of individual elements is insured, and on the other hand, control signals which lead to repetition of individual elements are suppressed.

In one embodiment of the present invention, first and second stores are provided which are capable of storing a predetermined number of elements of equal time length. One of said stores is filled while the other is emptied and the process is then reversed whereby the store being emptied is emptied in such a manner as to interchange the positioning of the stored elements in an aperiodic fashion with monitoring means continuously monitoring the interchange operations so as to prevent the occurrence of overlapping or gaps in the interchanged signal train. The process is reversed at the receiving end.

In another embodiment of the present invention, a single store is provided in which continuous interchange of the signal elements is performed.

It is therefore a primary object of the present invention to provide a novel apparatus and method for enciphering a signal transmission by aperiodic interchange of the time position of the signal elements and monitoring the interchanged operation so as to prevent the occurrence of overlapping of elements or the occurrence of gaps within the transmitted signal train.

Another object of the present invention is to provide a novel apparatus and method as described herein wherein the deciphering of the enciphered signal train occurs by a technique reverse from that set forth in the previous object.


These as well as other objects of the present invention will become apparent when reading the ensuing description and drawings in which:

FIG. 1 is the block diagram illustrating the fundamental principles of the invention;

FIG. 2a illustrates information signal wave forms showing the manner in which the wave form is broken down into individual elements and further illustrating the interchange and reversal of the individual signal elements;

FIG. 2b symbolically represents the interchange of signal elements at the transmitting end and the reverse operation at the receiving end;

FIG. 3a symbolically represents the principle of interchange of signal elements within equilength groups ("leapfrogging window");

FIG. 3b schematically depicts another interchange principle which avoids closed groups ("sliding window");

FIG. 4 illustrates an example of a device for implementing the method of the invention, which device employs the group interchange technique;

FIG. 5 is an examply of another variant embodiment of the invention with continuous interchange of signal elements;

FIG. 6 schematically represents the mode of operation of the device shown in FIG. 5;

FIG. 7 is a further example of the invention employing an information store consisting of several individual magnetomotor stores;

FIGS. 8a and 8b schematically represent the mode of operation of the device of FIG. 7;

FIG. 9 illustrates an example of the invention employing individual stores for the information store and separate occupation register for the individual information stores;

FIG. 10 is a schematic illustration of the principle of time compression of the interchanged signal elements, prior to transmission;

FIG. 11 shows the block diagram of a device for additional concealment signals which are added to the interchanged information signals;

FIG. 12 illustrates the block diagram of a cipher signal generator employed in performing the method of the present invention;

FIG. 13 is a detailed illustration of a portion of the cipher signal generator of FIG. 12; and

FIG. 14 is a block diagram of the cipher computer employed in the cipher signal generator.


Referring initially to FIG. 1, there is shown therein a block diagram of an enciphering system in which at the transmitting end (Index I) and receiving end (Index 2) the cipher signal generators SG1 and SG2 produce the aperiodic cipher signals w. These signals are supplied to the additional devices ZE1 and ZE2 , respectively, and are processed thereby in such a manner that at the outputs of these devices control signals i are produced, which signals satisfy the special conditions in terms of repetition and omission in the ensuing process of element interchange. The breakdown of the plain language signal x1 into equilength (timewise) elements and the interchange of these elements in accordance with the control signal i, take place in the cipher modulator SM1 where the enciphered signal z1 is produced. In order to restore the plain language signal x2 at the receiving end from the receive signal z2 and by a reversal in the process of interchange of the elements, the cipher demodulator SM2 is provided.

The plain language signal x may, in accordance with FIG. 2a, consist of a periodic train of oscillations, of the type, for example, occurring in a speech signal. This signal is divided into the sections or elements (i.e., time intervals) s1, s2, . . . , of uniform length throughout and which generally have no fixed relationship to the periodicity of the signal. These elements are stored for dissimilar times, for time-interchange purposes, so that a new signal z is produced. In order to make it more difficult for unauthorized persons to carry out the reverse process of interchange, it is also advisable to reverse the polarity of selected ones of the individual elements. Thus, for example, the element s5 ' has been produced by reversing the polarity of element s5. If the plain language signal is already available in digital form or is otherwise placed in this form by an analog-to-digital conversion, then the "elements" naturally consist of a specific number of bits, the number being the same in each element.

The interchange of signal elements at the transmitting end and the reverse procedure of interchange at the receiving end, taking also into account the individual cases of reversal of polarity which lead to the production of the s' type elements are symbolically illustrated in FIG. 2b wherein the elements s1 -s7 are interchanged in time so as to form the signal z1. This signal is received as z2 by the receiving end and is processed in such a manner as to rearrange the individual elements to return the elements to their original arrangement, as shown symbolically by the signal arrangement x2.

As shown in FIG. 3a, the interchange can take place within individual groups of elements of length F or can, for that matter, take place progressively, avoiding closed groups, as shown in FIG. 3b. The displacement of each individual element from its original position, is in each case restricted to the cross-hatched area which, in the case of FIG. 3a, leapfrogs ("leapfrogging window" technique) from group to group, or as in the case of FIG. 3b, shifts from element to element ("sliding window" technique).

A device SM for producing time-interchange between the elements of a plain language signal x1, as well as an additional device ZE for producing the control signals i for this interchange function, are shown in FIG. 4. These control signals must be derived from the aperiodic cipher signal w of the cipher signal generator SG, in such a way that in the interchange operation, no repetitions and omissions occur. Each element of the plain language signal x1 will be assumed for example, to consist of a train of k analogue pulses (scanned values), which have been obtained by periodic scanning of the original signals. A group of, for example, six elements for six k scanned values of the plain laungauge signal is then supplied via the double-throw switch W2 to the analogue shift store NS2 which is comprised of six k cells. The transfer through the register during the introduction of these scanned values is effected by pulsing signals eo which are applied through switch W3. At the same time, the elements of the previous group already stored in NS1 (which is substantially identical in design and operation to NS2), are extracted in the modified sequence through the output switch W4. The double-throw switches W0 and W1 are so designed that a control pulse i2, for example, with W1 in the solid line position illustrated in FIG. 4, causes k individual pulses of the pulsing signal eoo to be supplied to those storage cells Z2 ' of the register NS1 which are assigned to pulse i2. These k individual pulses thus form a pulse train (pulsing signal) j2 ' which brings about the extraction of the k scanned values of the second element, from the store.

In a similar fashion, the other elements are extracted in a sequence which is determined by the control signal i. During the extraction operation, gaps and repetitions must be avoided.

In order to fulfill the above requirements the device ZE is provided with an occupation store BS. The six cells of this store are initially filled by the pulse e6, with the commencement of each sixth element group, and is subsequently individually emptied by the hunting pulses h. The hunting pulses are produced from the cipher signals y1, y2, y3 of the cipher signal generator SG, with the aid of code converter CW which produces an appropriate (single) output pulse dependent upon the state of the three input pulses which it receives from binary adder AD. Since the three binary outputs from signal generator SG is capable of producing a combination of eight possible output pulses h1 . . . h8, and only six are actually required to empty the BS register, feed back of the pulses h7, h8 to binary adder AD is provided for. Adder AD is capable of performing a binary addition of three units to bring about a change in the hunting pulses h1 -- h6 which hunting pulses will be automatically changed each time a pulse h7 or h8 is generated. The emptying of a cell in register BS results in the generation of a corresponding control pulse i. However, if this cell has already been emptied, then an output pulse k is produced which once again results in a binary addition operation in AD to cause another change in the hunting pulse h, until a cell in BS which is still occupied is encountered and emptied. This procedure is repeated with each pulse e1 of the pulsing signal, each such pulse corresponding to a stored element. Thus, in this fashion, the result is achieved in that each of the control pulses in occurs only once and that none is omitted. This insures the desired unbroken and unrepeated extraction of all elements from register NS1.

After this emptying operation, a change pulse e'6 reverses the state of switches W1 through W4. At that time, the shift store NS1 is filled with elements of the plain language signal while the elements of the store NS2 which has previously been filled, are extracted in a changed sequence. The interchange between the information elements thus takes place in group fashion in accordance with the "leapfrogging window" principle of FIG. 3a. The pulse output w4 and device VM of FIG. 4 are employed as an additional enciphering device, as will be described in greater detail hereinbelow.

A device for the continuous interchange of signal elements in accordance with the "sliding window" principle of FIG. 3b is shown in FIG. 5. By way of information store NS once again an analog shift store can be used or for that matter some other known delay system with several supply or extraction points may be employed. For example, it is possible to employ magnetic tape recording with moving audio signal carriers. The supply of the elements r1, r2 . . . of the plain language signal x1 is controlled by the switches U1, U2, . . . in accordance with the actuation by control signals d1, d2, . . . , whose sequence must again satisfy special conditions in order to avoid repetitions and omissions. To insure that these conditions are complied with, the occupation register BR is provided in association with hunting switches S1, S2, . . . . The information register contains the monitored cells indicated in cross-hatched fashion while the additional cells are designed as unmonitored shift cells in order to reduce the complexity of the system. The cell content is shifted in rhythm with the pulsing signal e1 which corresponds with the pulsed rate of the individual signal elements. The still empty left-hand end of the register can be occupied cell by cell by the pulses d5, d4, . . . and the occupancy is checked by the monitoring signals b4. If the last cell P17 in the register is still empty, the absent occupation pulse b1 has the result that from the periodic pulse train c1 an individual pulse d1 is extracted and supplied to this cell so that the latter two are filled. If the cell is already occupied, the individual pulse is supplied as a pulse c2 to the hunting switch S2. This switch is once again controlled by any occupation pulse b2 which arrives and also by the blocking pulse a2 which occurs with a probability of A2. It is only the simultaneous absence of a2 and b2 that the storage cell P13 is filled by a pulse d2. Otherwise, a transmission pulse c3 is supplied to the hunting switch S3 so that monitoring is repeated. Any transmission pulse c5 coming from the switch S4, finally, if necessary, fills the initially always empty first cell P1 of the register. Thus, the function of the hunting switch obeys the following logic relationships, where the bar in each case indicates the negation condition:

dn = an bn cn (1) cn+ 1 = (1 - an bn) (2) ub.n

Since the possibly still empty last cell is filled in all cases and because at any rate the first cell can be occupied when all the other monitored cells have already been occupied, it is ensured that in each case one of the cross-hatched, monitored cells is filled and that thus a corresponding position pulse dn is supplied to the associated contact breaker Un. Thus, in the store two, one of the cross-hatched cells is occupied by a single element and a first cell (right) will ultimately always be occupied. Of course, there is still no guaranty that in the interchange process there is a like probability of occurrence of all the shifts, something which would be desirable in order to make the reverse process of interchange as difficult as possible for anyone trying to break the cipher. In order to achieve the best results, statistical considerations must be employed. The probabilities of the signals a, b, c and d are indicated by A, B, C, D respectively and the probability of the negation condition a = 1-a and by A = 1-A. Thus, we have

Dn = An Bn Cn (3) Cn+ 1 = (1 - An Bn) (4) ub.n

It is desirable that in each case the next element in the plain language signal should be supplied to each of the available cross-hatched positions with the same probability, i.e., the condition:

Dn = 1/N (N = sign number of elements per group) (5)

should s apply. This makes the probability of the occupation which is already taken place:

Bn = 1 - n/N (6)

and the probability of occupation of the position which is still free:

Bn = n/N (7)

and from this we obtain:

Cn = Cn-1 - Dn-1 = N - n + 1/N (8) An = N/n(N - n + 1) (9)

For the optimum probability of the blocking pulses a, we thus obtain the following, for N=5:

A1 = 0

A2 = 3/8 = 0.375

A3 = 4/9 = 0.444

A4 = 3/8 = 0.375

A5 = 0

In order to obtain blocking signals a2, a3, a4 whose probabilities correspond as closely as possible to these values, this from the cipher signals y1, y2, . . . y10 which occur with a uniform probability distribution of 50 percent, the cipher signal converter marked SW can be used and is made up of the logic gates Lo (logic "OR") and Lu (logic "AND") as is shown in FIG. 5. Precise adherence to these probability values is of course generally not necessary and thus simpler circuits can be employed to produce the desired blocking signals.

The operation of the interchange system of FIG. 5 will be explained in somewhat more detail, making reference to FIG. 6. As shown in FIG. 6, the reference KS refers to the plain language signal, Gr the group (of information elements in the plain language signal), VS the interchanged signal, SpGr the storage group (in the information store), INr, the internal number (of the storage position within a group), FNr the serial (of the information store positions). The elements of the plain language signal (KS) are sequentially numbered (right-hand margin) and are also ordered in groups of four elements each. Element 4 passes via a switch U4 (FIG. 5) to the store and is there recorded at position 17 on the moving data carrier NS. Element 5 of the plain language signal passes, after four pulses of the pulsing signal, via switch U1 to the data carrier which has in the meantime advanced four steps; i.e., it is not recorded at position 5 on the data carrier, which position was originally disposed at this location, but at position (5+4) = 9. The position fixed in relation to the apparatus is, in each case, marked by a square (with the element number) and the coordinate relating to the data carrier by a circle (with the element number). Similarly, element 9 is recorded at position 17 (of the apparatus) and at position 25 (of the data carrier) and so on. These elements, which in each case appear at the first position in a group, exhibit neither repetitions nor omissions after recording, for the reasons explained earlier. In a similar manner, the elements which are in each case located at the second position in a group, are recorded. For example, element No. 2 of the plain language signal is once again supplied by a switch U1 to the data carrier which is now, however, moved by one step so that the element arrives at position 6 on the data carrier instead of position 5. Element 6 of the plain language signal, which is supplied via a switch U2 would, if the carrier were at a standstill, arrive at position 9 thereof. However, it receives an additional shift of (1+4)=5 steps because the carrier has moved by five steps at the time of recording, and so on. From an examination of the positions of recording (marked by circles) thus determined, of all the elements on the data carrier, it can be seen that at no point have two elements been recorded on top of one another. In other words, in the illustration, there is no point at which two circles are superimposed upon one another. Finally, at the bottom edge, the data carrier is illustrated, this time showing the elements with their original numbering. Beneath it, the occupation register, which controls recording, has been schematically illustrated together with the extracted signals d1, d2, . . . d5.

Instead of one continuously moving carrier, as shown in FIG. 5, it is also possible to work with several circulating carriers which, for example, may take the form of magnetic tape wheels or magnetic drums M1, . . . M6 which are driven through a common shaft as shown in FIG. 7. In order to select the eligible storage positions, once again an occupational register BR with associated hunting switches S1, . . . S4 can be used. The control pulses consequently produced would primarily be suitable for producing an effective recording upon the continuously moving data carrier. The rotating data carriers are assigned a data carrier of this kind in such a way that on M1 the four elements of a first group of the moving carrier are stored, on M2 the four elements of the second group, and so forth. The supply to the storage wheel memory devices must be progressively switched so that said relationship is maintained. To this end, the additional shift register HR1 is provided, through which the position pulses d are transmitted. The first four position pulses can be relayed directly in the form of corresponding control pulses i. The next group of four position pulses must, however, be moved one step to the left in HR1 so that a recording, which, for example, by way of U4 would take place in group 5 of the tape which has meanwhile been moved, now takes place through store M5 which corresponds to this group. Similarly, the next four position pulses, which thus correspond to the third group of the plain language signal, must be shifted two places to the left HR1 and so on. Signal extraction from the source M1, . . . M6, on the other hand, is controlled by extraction pulses k1, . . . k6 from the register HR2 through which circulates a single control pulse so that if the control pulses k1, k2, . . . Extraction shifts from store to store and takes place at the same locations on the rotating data carriers, as would correspond to scanning of a continuously transfer data carrier in fixed relationship to the equipment.

The effect of this type of recording can best be appreciated from FIG. 8a. The references to this figure correspond to those of FIG. 6 except that in the case of VS (interchanged signals) the designation SpGr (store group) has been replaced by Sp (store), since we are now dealing with individual stores. The recording control program, i.e., the sequence of the control pulses d, is, in this case, the same as in the arrangements of FIGS. 5 and 6. Because of the aforesaid additional shifts, however, the recording positions (marked by squares) which are fixed in relation to the equipment, i.e., the numbers of the individual rotating stores appear at the shifted locations indicated by the arrowhead lines. For example, element 5 of the plain language signal is routed not by way of switch U1 but instead by way of switch U2, and is stored in M2 (See FIG. 7). Similarly, in FIG. 8a, the recording position (in a fixed relationship to the equipment) of the plain language element 6 has an additional left-hand displacement (marked by the square). The storage positions of the rotating data carriers occur in this illustration in the oblique zones as drawn in, for example, for store M6. The first occasion of occupation is, in each case, marked by the circled number of the original element, while the corresponding continuing occupancy is simply marked by circles. Thus, it will readily be seen how the last position in store 6 is finally occupied by the plain language element. The shift register HR for insuring the additional shifts is drawn in at the bottom edge.

The extraction of the stored elements from the register is illustrated in FIG. 8b where the occupancies of the stored positions are once again indicated by the encircled numbers and then by dots. Since recording commenced with plain language element No. 1, in the first extraction cycle individual storage positions still remain unoccupied. Extraction takes place successively from store M1, M2, and so forth. It is in each case indicated by cross-hatching of the first group position corresponding to the read-out head. The movement of the storage positions beneath the read-out head is indicated by the horizontal arrows. The progressive extraction times correspond to the numbering at the right-hand margin, and it can be seen from this that the extracted elements appear in the same interchanged sequence as in FIG. 6.

Another interchange device for several data carriers M1, . . . M6, is illustrated in FIG. 9. Here, separate occupation registers BR1, . . . BR6 are provided for the individual stores, whose outputs are returned to their inputs via the respective switches W11, . . . W16, thus indicating the maintenance of the occupation condition over several cycles. An auxiliary register HR is occupied by 3 circulating pulses which in each case bring about the closure of 3 associated switches, e.g. W22, W23, W24. With polarity change in a storage cell at HR, an appropriate output signal appears, in this case it is k6, which on the one hand initiates sampling of the hunting switches S, with a corresponding starting point (in this case at S1), and on the other hand, via the corresponding switch U, brings about the emp-tying of an individual store (in the present case M6). The three neighboring occupancies in HR are ensured by using the polarity change pulse k6 to drive individual stages, polarity reversal taking place in VO. With the indicated position of pulses in the register HR and the switches W22, W23, W24 set accordingly, the switches S2, S3, S4 are supplied with random pulses P2, P3, P4 of probability P. When triggered by the starting pulse k6, therefore, a train of hunting pulses is propagated via the switches S1, S2, S3, . . . in a manner similar to that indicated in FIG. 5, until a still empty cell of one of the registers BR is occupied. The position pulse dn which produces this occupancy at the same time brings about the storage of a plain-language signal element in a corresponding section of the individual store Mn. Instead of the relay (indicated in FIG. 7), of the succeeding position pulses by a second register HR1, what happens this time is that the occupation monitoring function is advanced by virtue of the fact that through the shifting of the pulses through the register HR, both the control of the switches Sn and the function of the occupation register BR, experience a cyclic displacement by one step. This also applies to the switches Un and the stores Mn, with the result that with each group change, i.e., with each pulse e4 of the pulsing signal, the same change is produced in the store input as would occur with a continuous store of the kind described in FIG. 5. This also applies to the emptying of the individual stores by the control pulses kn. The occupation register (e.g., BR6 in FIG. 9) assigned to the emptied store, is emptied at the same time by interrupting the feedback through the agency of the feedback switch (e.g., W16).

The devices in the indicated examples are suitable without further explanation for performing the reverse process of interchange of the signal elements at the receiving end, the ciphered signals z2 being supplied via the leads indicated and the reverse-interchanged signals x2 being extracted at the points shown. The functions of supply and extraction of the elements to and from the stores are simply exchanged. With magnetic storage, erasing will conveniently be carried out directly by the new recording. However, an additional erasing function can be provided which comes into operation directly after signal extraction.

The storage methods referred to here are intended purely as examples:

Other known methods can be employed however, such as, for example, storage in the form of electrical charges of capacitive data carriers, electronic storage of the kind known for example from radar work, ultrasonic delay, piezoelectric storage, magnetic wire, film or core storage and so on.

Depending upon the storage method, the time of compression of the individual storage elements or of the scanned values contained in an element is possible. The individual elements s*1, s*2, . . . of the cipher signal of FIG. 10 are produced. This achieves the result that the linear distortions of the transmission channel do not produce any unwanted cross-talk between the positionally displaced elements. At the receiving end, by element expansion and reverse interchange, plain-language signals are recovered and the time-dispersion of the transmission channel has no undesired effects in this context. Compression is achieved by the use of pulsing signals eoo of somewhat higher frequency than the signals eo (See FIG. 4), i.e., by extracting the individual elements from the stores at a slightly faster rate. Similarly, element expansion at the receiving end is achieved by somewhat slower extraction from the stores.

It can be seen from FIG. 2a that reverse interchange is simplified for unauthorized monitors, by the particular amplitude values at the ends of the elements, i.e., such a monitor can in each case pick out elements whose end amplitudes match one another. This possibility is made more difficult by reversing the polarity of individual elements (s'5 in FIG. 2a), so that additional edge amplitudes are produced which increase the already large number of possible solutions. Polarity reversal is effected quite simply using the device controlled by additional cipher pulses, e.g., the (sign-modulated) switch VM of FIG. 4, which is operated by the cipher signal y4.

One highly effective measure to render the detection of associated elements more difficult consists in the addition of specific concealment signals at the transmitting end, which signals, with undistorted transmission, can be subtracted again at the receiving end. These concealment signals may conveniently be obtained from special cipher signals using a digital-analogue converter, possibly coupled with shaping by special filters. This kind of device, with the digital-analogue converter D/A and the filter BP, is shown in FIG. 11. Besides this concealment signal condition DM and the position modulator LM, once again a sign modulator VM is indicated.

The cipher generator SG used to produce the cipher pulses, can, as FIG. 12 shows, consist of a program signal generator PG, a cipher selector SE and a cipher computer SC. The programme signals u are generated in the programme signal generator in accordance with a specific logic law, e.g., using a shift register, whose output pulses, taken from two points, are fed back via a modulo-2 logic system and the switch So shown in FIG. 13 to the input. Synchronization with transmitted programme pulses g is made possible by initial injection of this pulse train through the switch So until the fedback pulses coincide completely with the new incoming pulses. This condition is detected by the correlator KO which then automatically switches So to the feedback position for further operation on its own.

In the cipher selector SE, in addition to the permanently wired and possibly exchangeable line matrix MA, an additional switching of the extracting signals s1, s2, . . . s6 using the decade switches SW, is provided for. These intermediate signals are thus dependent in an unambiguous way upon the switch positions. They serve to control the actual cipher computer SC. This, as FIG. 14 shows, can consist of the two registers R1, R2 in association with the switches S1, . . . S6 which are controlled by the intermediate signals s1, . . , s6. The registers have periodic feedback via switches s1 and S4. The after effect of pulses injected earlier disappears in the course of time thanks to partial interruption of the feedback function. A change in signal in the feedback channel is brought about by product formation between the fedback signals and the intermediate signals s2, s5 in S2, S5. Finally, the pulsing signals eo for the registers are periodically interrupted by the switches S3, S6 so that the circulating pulse trains have no specific period.

It can therefore be seen from the foregoing description that the present invention provides the method for interchanging signal elements of a transmitted message in a non-uniform manner wherein the transposed signal elements are introduced into new time slots wherein the monitoring technique employed assures the fact that there is no overlapping of signal elements nor are any gaps provided.

Although in the foregoing preferred embodiments of this novel invention have been described, many modifications will now become apparent to those skilled in the art and it is therefore preferred that this invention be limited not by the foregoing description but only by the appending claims.