Vibroacoustic masking systems.

Vibroacoustic camouflage systems..

Anatoly Anatolyevich Khorev, Doctor of Technical Sciences, Professor

VIBROACOUSTIC MASKING SYSTEMS

The article discusses the principles of construction, main characteristics and methodological recommendations for the installation of vibroacoustic masking systems.

Protection of acoustic (speech) information is one of the most important tasks in the general complex of measures to ensure the information security of an object or institution and is carried out using passive and active methods.

Passive methods of information protection are usually implemented during the construction or reconstruction of buildings at the stage of developing design solutions, which allows for the types of building structures, methods of laying communications, and optimal locations for the placement of dedicated (protected) premises to be taken into account in advance.

In case of technical impossibility of using passive means of protecting premises, or if they do not ensure compliance with the required standards for sound insulation, active protection measures are used, consisting of creating masking acoustic and vibration interference for acoustic speech reconnaissance means.

Acoustic masking is effectively used to protect speech information from leakage through a direct acoustic channel by suppressing the microphones of reconnaissance means installed in such structural elements of protected premises as: door vestibule, ventilation duct, behind a suspended ceiling, etc.

Vibroacoustic masking is used to protect speech information from leakage via vibroacoustic and acousto-optic (optical-electronic) channels and consists of creating vibration noise in elements of building structures and in utility lines. Vibroacoustic masking is effectively used to suppress such means of interception of information as electronic and radio stethoscopes, as well as laser acoustic reconnaissance systems.

The process of speech perception in noise is accompanied by losses of the constituent elements of the speech message. Verbal speech intelligibility is used as an indicator for assessing the effectiveness of vibroacoustic masking systems. It is characterized by the number of correctly understood words and reflects the qualitative area of ​​intelligibility, which is expressed in categories of details of the report being compiled on the conversation intercepted using technical intelligence means. The criteria for the effectiveness of speech information protection largely depend on the goals pursued when organizing protection, for example: to hide the semantic content of the ongoing conversation, to hide the topic of the ongoing conversation, etc.

Practical experience shows that it is impossible to compile a detailed report on the content of an intercepted conversation if the verbal intelligibility is less than 60-70%, and a brief report-annotation is impossible if the verbal intelligibility is less than 40-50%. If the verbal intelligibility is less than 20-30%, it is significantly difficult to establish even the subject of the conversation, and if the verbal intelligibility is less than 10%, this is practically impossible even with the use of modern noise reduction methods.

A typical vibroacoustic masking system includes: a noise generator, a set of vibration emitters, a set of acoustic emitters (sound speakers), as well as equipment necessary for adjusting and setting up the system.

The main characteristics of noise generators that affect the effectiveness of speech information protection include: the type and frequency range of the generated interference, their amplitude-frequency characteristic and noise quality factor, the number of linear outputs, the maximum number and types of vibration emitters connected to them, as well as the ability to adjust the power and spectrum envelope of the interference in each channel.

The role of terminal devices in vibroacoustic masking systems, which convert electrical noise oscillations into acoustic oscillations of the speech frequency range, is usually performed by small-sized broadband loudspeakers, and those converting electrical noise oscillations into vibration oscillations are vibrators, usually of the electromagnetic or piezoelectric type.

In practice, analog, digital and combined noise generators have found wide application.

A large group of analog noise generators are devices whose operating principle is based on amplifying the oscillations of primary noise sources, the latter being vacuum, gas-discharge, semiconductor and other electronic devices and elements.

A temporary random process, similar in its properties to noise oscillations, can also be obtained using digital noise generators that form chaotic (pseudo-random) sequences of binary symbols and transform them into sequences of chaotic pulses.

In acoustic and vibroacoustic masking systems, the following types of noise interference are usually used:

1 – “white” noise (noise with a constant spectral density in the speech frequency range);

2 – “pink” noise (noise with a tendency for spectral density to drop by 3 dB per octave toward high frequencies);

3 – noise with a tendency for spectral density to drop by 6 dB per octave toward high frequencies;

4 – noise “speech-like interference” (noise with an amplitude spectrum envelope similar to a speech signal).

In accordance with the requirements of the State Technical Commission under the President of the Russian Federation, the interference generator must generate noise oscillations in the frequency range from 175 to 5600 Hz. Fig. 1 shows the calculated dependences of the verbal intelligibility W on the integral signal-to-noise ratio q, measured in the frequency range from 175 to 5600 Hz with various types of noise interference [2].

As can be seen from the graph, the most effective are “speech-like” interference and interference of the “pink” noise type (noise with a tendency for the spectral density to fall by 3 dB per octave towards high frequencies). However, it is necessary to remember that this type of spectrum envelope should not be created at the output of the noise generator, but at the output of the vibration emitter, which has its own amplitude-frequency characteristics.

Fig. 1. Dependence of verbal intelligibility W on the integral signal-to-noise ratio q in the frequency band from 175 to 5600 Hz:
1 – “white” noise;
2 – “pink” noise;
3 – noise with a spectral density drop of 6 dB per octave towards high frequencies;
4 – noise “speech-like” interference

As an example, Table 1 and Fig. 2 show the average levels of the vibration noise signal in octave frequency bands generated by the VN-GL, VN and VNT-2 vibration emitters when a noise voltage of the “white” noise type [2] is applied to their input.

As can be seen from Table 1average acceleration levels measured in octave frequency bands for different vibration emitters may differ by 20–30 dB or more. Therefore, the ability to adjust the noise signal spectrum is an important characteristic of a noise generator.

Table 1. Average levels of vibration noise signal in octave frequency bands generated by VN-GL, VN and VNT-2 vibration emitters

Fig. 2. Average levels of vibration
noise signal in octave frequency bands, generated by vibration emitters VN-GL, VN and VNT-2

An important characteristic of the noise generator is the noise quality factor Ksh, showing the degree of similarity of the generated interference signal with “white” noise. The higher the value of the noise quality factor, the fewer regular components in its spectrum and, therefore, the more difficult it is to isolate the hidden speech signal when using various noise cleaning methods.

In accordance with the requirements of regulatory and methodological documents on counteracting acoustic speech intelligence, the noise quality factor of vibroacoustic masking systems must be at least 0.8.

The calculation of the standardized noise quality factor is performed in accordance with the regulatory document “Means of active protection of electronic equipment facilities from information leakage through side electromagnetic radiation and interference. Basic technical requirements (Ministry of Radio Industry, 1987) based on the values ​​of the asymmetry coefficients and excess of the probability density of the generated noise distribution obtained using the device for studying the correlation characteristics of the X6-4.

Fig. 3 shows the typical density of the noise voltage distribution at the output of the vibroacoustic masking system generator.

Fig. 3. Density distribution of noise voltage at the output of the noise generator

At present, a large number of different active vibroacoustic camouflage systems have been created, which are successfully used to suppress speech information interception means. These include: the Baron, Zaslon-2M, Kedr, Kabinet, Ravnina-2K, Skit-AR, Skit-MVA, Sonata-AV, Shelest-4K, Shorokh-1 (2) , VV 301, LGSh-401, ANG-2000, SI-3001, SI-3002, SI-3030, VNG-006, VNG-012GL systems, etc. The main characteristics of some of them are given in Table. 2 and 3 [1, 4, 5].

Let's consider the principles of constructing noise generators using the example of the noise generator of the VNG-012GL vibroacoustic camouflage system (Fig. 4) [1].

Fig. 4. Structural diagram of the VNG-012GL system noise generator

In the block of analog noise generators, electrical signals with random parameters (noise) are formed, which are sent to the block of digital signal processor, where their analog-to-digital conversion, level adjustment and formation of the amplitude-frequency characteristic of noise on four channels are carried out. It is possible to adjust the spectrum of the noise signal on each of the four outputs separately in octave or 1/3 octave frequency bands. This allows for the optimal formation of the amplitude-frequency characteristic of the interference.

From the digital signal processor block, the formed noise signals are sent to four digital-to-analog converters (DAC1 — DAC4), and then to high-voltage amplifiers, to the outputs of which piezoelectric vibration emitters are connected. A low-frequency amplifier ULF 1 is connected in parallel to the high-voltage amplifier of the first channel, to the output of which acoustic systems or electromagnetic vibration emitters can be connected.

The low-frequency amplifier ULF 2 is connected to the block of analog noise generators via an electronic level controller. The amplitude-frequency characteristic of the noise in this channel is formed using analog filters and is not adjustable.

For outputs 1 through 4, a special program allows you to set the width of the frequency bands for spectrum adjustment: 1/3 or 1/1 octave (common for all channels), as well as adjust the integral signal level and the levels of its spectral components for specified bands in each channel individually. For output 5, the program allows you to set only the integral level.

Thus, the noise generator has five independent noise signal outputs, allowing you to adjust not only the signal amplitude in each of them, but also its amplitude-frequency characteristic.

The interference parameters are adjusted after the system is installed at the commissioning stage using a personal computer (Fig. 5), connected to the generator via the RS-232 interface through a computer interface unit. The interference characteristic set depending on the protection requirements is remembered and stored in the non-volatile memory of the generator, which allows the generator to operate autonomously (without a computer).

Along with noise interference, vibroacoustic masking systems use speech-like interference, for example, a “speech choir” (simultaneous conversation of several people).

Fig. 5. User interface of the program in the mode of adjusting the amplitude-frequency characteristic of interference:
1 — button for switching the generator operating mode adjustments from one-third octave to single-octave frequency bands and back;
2 — “Setting the maximum level”;
3 — channel number;
4 — indication of the status of all generator settings;
5 — output signal level controller;
6 — output signal level controller in the band

Fig. 6 shows the block diagram of the noise generator of the vibroacoustic camouflage system “Baron-2”, which in addition to the usual analog noise generator for generating interference includes three radio receivers (RM1 – RM 3), independently tuned to different FM (VHF-2) radio stations [4]. In the noise generator of the vibroacoustic camouflage system “Baron-U”, instead of radio receivers, special “phoneme cloners” are used, synthesizing speech-like interference by cloning the main phonemic components of the speech of specific individuals.

Fig. 6. Block diagram of the noise generator of the vibroacoustic camouflage system “Baron”

The formation of interference signals takes place in two stages. In the first stage, using a computer and special software, pseudo-speech is synthesized from the voice recording of one or more people by cloning the main phonemic components of their speech, which is a certain sequence of signals. In the second stage, the interference synthesizer, whose memory contains pseudo-speech, randomly takes random fragments from this sequence of signals, which are then fed to the input of the interference channel [4].

The generator also provides for the regulation of the spectrum of interference signals in four independent channels in the frequency ranges: 60 — 350 Hz; 350 — 700 Hz; 700 — 1400 Hz; 400 — 2800 Hz; 2800 — 16000 Hz.

The effectiveness of the vibroacoustic masking system is largely determined by the correct choice of installation locations and methods for attaching vibration emitters.

The required number of vibration emitters is determined based on their locations, the design and materials of the enclosing surfaces, window openings and utility lines, as well as the effective radius of vibration emitter suppression on the corresponding surfaces.

Usually, the effective suppression radius of a vibration emitter is understood to be the maximum distance along the surface from the place of its installation to the place of installation of the sensor (contact microphone) of the reconnaissance means (for example, a stethoscope), at which, with the maximum level of the noise signal supplied to the sensor and the location of the source of the hidden speech signal (test generator) at the minimum possible distance from the place of installation of the sensor of the reconnaissance means, the required efficiency of suppression of the reconnaissance means is ensured (i.e. at which the value of the signal-to-noise ratio in at least one of the octave bands becomes equal to the maximum permissible value given in the regulatory and methodological documents on counteracting acoustic speech reconnaissance (NMD ARR) of the State Technical Commission of Russia).

The effective suppression radius depends not only on the characteristics of the vibration emitter itself, but also largely on the characteristics of the surfaces being noised, and is therefore determined experimentally. When moving 1 m away from the vibration emitter installation site, the noise level it creates decreases by approximately 3–6 dB.

Usually, the following recommendations are used when installing vibration emitters.

For noise reduction of wallsVibration emitters are installed on the center line between the floor and the ceiling. The distance between them should be no more than 2 x ri, where ri is the effective suppression radius of the vibration emitter on the corresponding type of surface (concrete, brick, etc.). It is advisable to install vibration emitters as close as possible to the possible installation locations of reconnaissance sensors. If only one vibration emitter is required during installation, it is usually installed in the center of the wall.

When noise is present in the ceiling or floor, the required number of vibration emitters is Nis selected from the condition:

, where S is the area of ​​the noise-absorbing ceiling (floor, ceiling), m2.

The distance between the vibration emitters should also not exceed 2 x ri,.

The installation of vibration emitters on the surface of building structures is carried out, as a rule, using a special spike, fixed in the ceiling with epoxy putty (Fig. 7[1]). The vibration emitter is screwed onto the tenon after the putty has polymerized, usually 2–3 hours after its application.

When making noise in windows, vibration emitters are attached either in the center of the window frame, or one on each element of the window glazing, or on the window frame. To attach vibration emitters to the glass surface, use a special adhesive for bonding metal to glass (Fig. 8 [1]).

Fig. 7. Example of mounting a vibration emitter on a wall

Fig. 8. Examples of mounting vibration emitters on glass

When noise is present in utility linesVibration emitters are installed on each incoming/outgoing pipe. Vibration emitters are mounted on utility lines (pipes) using a clamp (Fig. 9 [1]).

Fig. 9. Examples of installing vibration emitters on pipe communications

The required number of acoustic emitters is determined based on the calculation of one emitter for each ventilation duct or door vestibule or 25–30 m2 of suspended ceiling (in this case, the acoustic speakers are installed behind the suspended ceiling near possible locations for installing reconnaissance microphones).

The vibroacoustic masking system installed in a dedicated (protected) room should not itself create additional technical channels for information leakage, for example, due to electroacoustic transformations of acoustic signals or parasitic generation. Therefore, after installing the system, it is necessary to test it for compliance with the “Standards for the Protection of Speech Information Processed by Technical Means from Leakage Due to Spurious Electromagnetic Radiation and Interference (PEMIN)” (State Technical Commission of Russia, 1998).

The adjustment and assessment of the effectiveness of the vibroacoustic masking system is carried out during its certification using certified measuring equipment for general-purpose control [3].

The choice of locations (control points) for the placement of control equipment elements depends on the type of speech intelligence equipment for which the protection of speech information is carried out.

If the location of the speech signal source is known (desk, place of conversation, etc.), the installation point of the test acoustic signal source is located at the location of the speech signal source. If it is impossible to determine the specific location of the speech signal source, the test acoustic signal source is located at a distance of 1 m from the nearest enclosing structure in the reconnaissance hazardous direction and at the same distance from other enclosing structures and objects.

The control points for installing the acoustic sensor (measuring microphone) are the locations of possible placement of speech reconnaissance equipment (parking lots, bus stops, rest benches, windows of nearby buildings, etc.). If it is impossible to install the measuring microphone in the actual locations of possible placement of speech reconnaissance equipment, the control points are placed on the border of the controlled (protected) zone.

When monitoring the security of speech information from vibroacoustic speech reconnaissance equipment, the control points for installing a measuring contact microphone (vibroacoustic sensor) are the surfaces of various enclosing structures, utility lines and other objects that are located in reconnaissance-hazardous directions, external to the source of the speech signal, as well as possible places on utility lines (building structures, etc.) accessible to unauthorized persons.

To control compliance with the standards for protecting speech information from optical-electronic speech reconnaissance equipment, a contact vibroacoustic sensor is also used. It is attached with a special paste or glue to the outer surfaces of window glass or other reflective plates that vibrate under the influence of speech acoustic signals, and the normal to their surface coincides with the reconnaissance-hazardous direction.

After taking measurements, the verbal intelligibility of speech W is calculated for each control point. If the calculations show that W> Wп (where Wп(normative (threshold) value) the corresponding channel of the system is adjusted (or the signal level in the corresponding channel or in the corresponding octave band is increased). When W Ј Wп the set signal levels in the corresponding channels and spectral bands are recorded and the system can be accepted into operation.

Literature

1. Equipment for vibroacoustic protection of premises “VNG-012GL: User's Guide. Moscow: IKMC-1, 2003. 25 p.
2. Equipment for vibroacoustic protection of premises VNG-012GL: Technical conditions. Moscow: IKMC-1, 2003. 50 p.
3. Zheleznyak, VK, Makarov YuK, Khorev AA Some methodological approaches to assessing the effectiveness of speech information protection//Special equipment. 2000, No. 4, pp. 39 – 45.
4. Baron series vibroacoustic protection systems: Information materials. M.: NPC Nelk, 2003. 18 p.
5. Modern security technologies: Catalog. M.: TsBI “Maskom”, 2003. 52 p.

Table 2. Technical characteristics of noise generators of vibroacoustic masking systems

Table 3. Technical characteristics of vibration emitters