DE3105214A1 - "SUB-TAPE EXCHANGE PROCEDURE" - Google Patents

Description

31052U

Licentia Patent-Verwaltungs-GmbH NE2-BK / Ruf / white

Theodor-Stern-Kai 1 BK 79/90

D-6000 Frankfurt

Partial band swap procedure

The invention relates to a method for interchanging sub-bands of equal width of a signal frequency band according to FIG Generic term of claim 1.

Such methods are known, for example from DE-PS 24 26 451, in which a circuit arrangement for achieving Interchangeability-independent frequency grid for the harmonics of basic speech frequencies in facilities for sub-band swapping, in which a speech signal band is divided into sub-bands of equal width on the transmission side made and a transfer band is formed by a swap of the sub-bands, in which therein contained subbands can be inverted, the reversal made on the receiving side being reversed is made, is described, which is characterized in that the frequency is at the lower limit of the voice band and half the sub-bandwidth are in the ratio of two whole numbers to one another. This circuit arrangement has five modulators, which by means of suitably selected,

"'J'IUbZ I

- 6 - *. BK 79/90

upper carrier frequencies at the outputs of the downstream Unit filter divides the voice band into five subbands. The sub-bands obtained in this way are then with the help of further five downstream modulators converted back to the original frequency position of the voice band, whereby an interchange of the sub-bands is achieved in that the modulators with a selectable selection of five ten carrier frequencies are supplied.

The known bandpass methods require a relatively high one Circuit effort.

The object of the present invention was therefore to provide a method for the subband swapping of the above type, which requires less circuitry and allows discrete-time or digital signal processing.

The solution takes place with those specified in the claims Means.

; The inventive method has the advantages that There is no need to switch between the carrier frequencies fed to the modulators, which is a saving in terms of circuit complexity. In a further development of the invention, the passband of the bandpass filter is set to one of the sub-bands, whereby two modulators and a carrier frequency generation can be saved. In Further developments of the invention provide information on how the frequency values for the carrier oscillations are to be selected depending on the division into sub-bands and the sampling frequency, so that a time-discrete or digital signal processing can be carried out in an advantageous manner. Both rectangular sinusoidal carriers at relatively low sampling frequencies and a simple modulator circuit can also be used. In further embodiments of the invention

31052Η

- 7 - "BK 79/90

favorable values for obfuscating a CCITT voice channel which is divided into five sub-bands.

Eb, the description of the invention now follows with reference to the figures:

Fig. 1 shows the block structure of a speech according to the invention obfuscation system by means of five-band swapping.

Fig. -2 shows spectra after frequency conversion with four Sinus carrier oscillations with time-discrete signal processing.

3 shows an example of a modulator implementation.

■ '

4 shows spectra after the frequency conversion with a square-wave carrier wave with a time-continuous wave Signal processing.

^{The basic block diagram of a speech obfuscation system} based on five-band swapping is shown in FIG. The input spectrum of the speech signal Xin (f) to be disguised can contain undesirable higher-frequency spectral components, which may have to be suppressed by an input-side low-pass filter TP1 in order to be able to adhere to the sampling theorem, for example in the case of time-discrete signal processing with a given sampling frequency fS. The resulting band-limited spectrum X (f) is calculated in the next step using the same bandpass filters

OQ divided into five sub-bands. There is also a previous modulation with corresponding carrier frequencies ω1, ω2 etc. necessary. The actual scrambling takes place in the next step using a coupling matrix M, which is generated by a random generator is controlled instead, with the sub-bands X (f)

oc ν = 1, 2 ... 5 are randomly rearranged in short time intervals according to {i, j, k, .1, mlc {1, 2, 3, ^, 5 '. Then the sub-bands {X (f) for ν = i, j, k, 1, m} with the same frequencies as the input

- 8 - BK 79/90

Modulators are implemented again and thus come in a scrambled order into the original frequency range of the speech spectrum X (f) to lie. Possible harmonics of the output spectrum such as B. unwanted modulation products or periodically recurring spectral components with time-discrete signal processing are eliminated by the low-pass filter TP2. Y (f) is the spectrum of the veiled continuous output signal. The system for decrypting disguised Speech signals likewise have the block structure shown in FIG. 1, with for correct decryption the receiving-side random generator ZG the same random sequence as the transmitter and synchronized with it has to deliver.

The method according to the invention also yields the saving two modulators when the pass band of the unit band pass filter BP1 to BP5 falls to one of the five subbands and one of the carrier frequencies on the input and output side receives the value 0. 2a shows the spectrum of the voice signal (telephone quality), the rule position fully representing the continuous-time signal and the additional spectral components symmetrically to integer multiples of the sampling frequency fS (here is the Upside down below fS drawn) with time-discrete processing appear.

The CCITT voice channel is 0.3 to 3.4 by the limits kHz, which means that when divided into five sub-bands, the sub-bandwidth is B = 620 Hz. Die.arithme- ■ jO table center frequencies of the sub-bands are then included

fM '= (2v - 1). § + 300 Hz

for ν = 1 to 5. For simple carrier frequency generation and for time-discrete signal processing it is beneficial to use the center frequencies

ft * · ♦ • * ·

31052U

- 9 - BK 79/90

f M _v = ν. B.

to modify slightly. This extends to 05

B B scrambling speech spectrum from j to 11 ^ (0.31 to 3.41 kHz)

In order to convert the sub-bands of the spectrum belonging to the continuous-time signal according to FIG. 2a into the same frequency band by means of square-wave carrier oscillations, the fundamental frequencies f _v for ν = 1 to 5 are to be arranged equidistantly at a distance B. According to the formula

f _v = (n _v - j). B (4)

for r ^ = 0, 1, 2 etc. ^v = 1 to 5, 1 = ε IN and i = 0, 1, ..., 1-1 as freely selectable parameters. In the simplest case for 1 = 1 and i = 0 one obtains f _v = n _v . B. The minimum value of the square carrier frequencies must not fall below 4B, since otherwise disruptive spectral overlaps occur due to the hermeticity of the spectra of real signals. This condition arises in the same way due to the line spectrum of the rectangular carrier oscillation according to Fourier. For η = 0 corresponding to f _v = 0, the specified restriction does not apply, since no frequency conversion is carried out here. In FIG. 4, the spectral relationships for continuous-time signal processing are shown in detail.

The same spectral relationships can also be found in the frequency range shown in FIG. 4 with a time-discrete one Realize the system if the sampling frequency fs is chosen appropriately. Here fs is to be determined in such a way that all Harmonics of all rectangular carriers f. For ν = 1 to be symmetrical to half the sampling frequency. Through this it is ensured that the lines of the rectangular carrier spectrum, which are already present, are due to the periodicity time-discrete signals no further undesired lines are added. The condition for this is

J IUOZ.

- 10 - BK 79/90

With reference to half the sampling frequency, the spectrum then occurs lines of the bevelled rectangular beam vibrations in the following Frequencies on:

O ^r ) for rn odd ~ - 2μ fp and

fs + for in _v even 75— - (2μ + 1). fp,

each for ju. an element from the set of integers Z = {-. -1, 0, +1 ... |. The sampling frequency thus results

fs = 2n. m. B = k. B.

V V

Effort Favorable values for k, the associated sample ^{frequency-l '7} fs and η are summarized in Table 1, wherein in the first four rows solutions with non-implementation of the sub-bands 3, ^l \ or 5, thus saving modulators Darge · is shown.

~ ⁽⁾ In addition to the Aufwandsm'inderung results in a lower sampling frequency, respectively.

ErfindungsgemäJt ¹ × the sampling frequencies can further degrade be used if vibrations to implement sinusoidal carrier "■>, let be composed of only a few different magnitudes of samples to corresponding consequences. Again, the relationships hold (h) and (5 The following applies to the sampling frequency

■ * ° Vf, = K (IV {1. n _v - lj. Γ, ν = 1, ... '>

or

• fs = ~ - -, k ε | N.

ν. . 3Ί052Η

- 1 1 - BK 79/90

Because of the sampling theory, however, k / l-11 must be observed to avoid disruptive spectral convolutions. In the following table 4, possible carrier frequencies are possible for the values k / l = 15,? 0, 24, 30 and 36 in corresponding five columns relative to the sampling frequency fs and as a function of η as well as for green ", t. values of i and 1 shown:

Table 4 f '/ fs
ν f _v / fs ς, / fs f _v / fs f _v / fs ■ ¹ . η - i 10 Π 0 O k η O no 1/10 1/10 1/12 O I 1/6 1/8 1/10 02 1/5 1/6 1/12 15th 03 3/10 1/4 1/6 04 3/10 1/4 1/5 05 1/6 06 1/2 2/5 1/3 07 3/8 3/10 20th 06 1/2 5 / 12- 1/3 1/4 Ο ■ 1/2 ^2/1 J Ι Π 1/2 1/3 λ? 5/12 25th ν-, 15th 20th 24 30th 1/2 18th 1 O O O 36 k / l 2 1 1 1 O i 1 30th 1

From table 4 it can be seen that in the case k / l = 15 with 1 = 2 and i = 1 only four of the permissible carrier frequencies result. Likewise, the last result reproduced there for k / l = 36 with i = 0 is unusable, since there are only a maximum of four of the permissible carrier frequencies in any frequency intervals of width 10B

- 12 - BK 79/90

kuraint-ii, so dal. ¹ , highest :; Four · Te i! · bands with the same filters can separated. The spectra after the frequency conversion; r.jrul for k / l = ? Λ \ in Fig. 2. For the values k / l − 2Ά and 30, the data for the carrier frequencies, again relatively 2, ur sampling frequency for the individual subbands and the center frequency of the bandpass filter are listed in Table 2, which are favorable for implementation. With regard to time-discrete signal processing, the most cost-effective case is the choice of the band center frequency of the bandpass filter -zu

sc = 5 B,

since in this case a subband signal can be obtained without frequency conversion.

Blank

Claims (1)

Claims Method for interchanging sub-bands of equal width of a signal frequency band by means of modulators, which are followed by the same bandpass filters and further modulators, characterized in that the modulators are permanently assigned certain carrier frequencies, and that between the bandpass filters (BP) and the downstream modulators (MI .. .) a coupling matrix (M) is provided for scrambling the sub-bands. 2. The method according to claim 1, characterized in that the coupling matrix can be controlled by a random generator (ZG). 3. The method according to claim 1 or 2, characterized in that the pass band of the bandpass filter is equal to one of the subbands and that one of the carrier frequencies is selected to be zero. 4. The method according to any one of the preceding claims, that ft · ·· 31Ü52U - 2 - BK 79/90 characterized in that the carrier frequencies have the values f _v = (n _v - i). B. with v = 1, 2 ... sub-bands of bandwidth B and i, 1 and n _v ε (N = i ₇ 2 ... (elements ε from the set of natural numbers jN) as freely selectable parameters, '5. Method according to Claim ¹ J, characterized in that the carrier frequencies are the values f _v = n _v . B. exhibit. 6. The method according to claim 4 or 5, characterized in that the modulators operate time-discrete and the relationship for the sampling frequency of the carrier oscillations fs = 2m _v . f _v , m _v ε IN = 1, 2 ... is applicable. 7. The method according to claim 6, characterized in that the carriers are sinusoidal and the sampling frequency increases fs = KGV (1. η - i}. τ ¹ ν 1 with KGV equal to the least common multiple ^ o or to fs = k. γ with k ε JN is chosen. 8. The method of claim 6, wherein the signal frequency band is the CCITT voice channel, characterized in that ' 3Ί052Η - 3 - BK 79/90 η = 5 sub-bands with the bandwidth B = 620 Hz are formed and that the frequency conversion with rectangular carriers takes place according to one of the following options in Table 1: Tabel 560 1 fs / kHz n1 n2 n3 n4 n5 k 1260 347.2 4th 5 7th 8th 1008 781.2 5 6th ■ 7 - 9 1008 624.96 4th 7th 8th 9 - 1680 624.96 6th 7th 8th 9 - 1041.6 4th 5 6th 7th 8th 9. The method according to claim 6 or 7, characterized in that η = 5 sub-bands are formed with the bandwidth B = 620 Hz and that the frequency conversion is carried out with sinusoidal carriers according to one of the following options in the table: Table 2 f / fs
V 1/6 f _v / fs f / fs
V f / fs
V part 1/4, 1/3, 1/4 1/5 3/10 tape
No. 1/8 3/8 1/6, 3/10 1/3, 1/5 1 1/3 1/6 1/3 ■ 1/6 2 3/8 5/12 1/8 1/10 2/5 3 ⁰ ! B. 1/2, 1/12 2/5 1/10 4th 5. 7B 7B 8B 5 24 24 30th 30, fM k / l 3 Ί U b 2 Ί 4 - 4 - BK 7979 0 10. The method according to claim 7 or 9, characterized in that the modulators consist of a multiplier circuit which multiply the signal for f / fs or integer multiples thereof according to the following table 3 with the factors specified there in the cycle of the sampling frequency: Table 3 f _v / fs 1/2 1/3 1M 1/5 1/6 1/8 1/10 1/12 1 . -1 ti 0 ± 1 0 0 Ϊ1 Factors -1 0.5 0 1 'io. 5 -1 ^ 1 ^ 0.268 io.618 -iT ¹ to.618 io.732