The present invention relates generally to the field of detecting three-way call attempts in controlled telecommunications systems. In particular, the present invention relates to a system and method for preventing a user from successfully circumventing or masking a three-way call attempt by generating a continuous or constant noise. The system and method of the present invention may be utilized with any existing three-way call detection system or method. The present invention preferably monitors for periods of silence and whether the examined samples are below a pre-determined threshold. If the pre-determined threshold is exceeded, the present invention determines that an attempt to circumvent a three-way call has occurred if the amount of time between periods of silence exceeds the maximum allowable duration for continuous audio. The system also includes a continuous noise detection algorithm for detecting the presence of false positives in the audio steam.
Many institutions, such as prisons, nursing homes, mental institutions, etc., include controlled telecommunications systems that offer inmates or residents limited calling access. One reason for controlling use of the system is to prevent the institution from incurring unaccountable telephone costs. Other reasons for controlling access to the system include preventing harassing calls to outside parties, preventing fraudulent activities, etc. Therefore, systems in such environments often monitor and control the telephone activity of each inmate or resident. For example, systems may restrict calling to only certain telephone numbers. Systems may also have a means of maintaining call records for each inmate or resident, and a means for communicating with called parties to enable the called parties to prevent future telephone calls from inmates or residents. In short, the communications system used in a regulated institution must employ unique monitoring and control functions often unnecessary in other types of telecommunications systems.
In order for the methods of monitoring and control to be effective, it is important to prevent inmates or residents from exploiting any loopholes that can be used to bypass the control features of the system. For example, inmates or residents have been known to use three-way calling to have an outside party connect the inmate or resident to a blocked number. A three-way call is initiated when the remote called party depresses the hook switch on the telephone, generating a hook flash signal. The caller is temporarily put on hold while the called party establishes a connection with a third party. Then, all three parties can converse. Using three-way calling, the inmate or resident may utilize the institution's call system to, among other things, access blocked telephone numbers, for example, to perpetrate additional criminal activities, or harass certain parties.
It is therefore critical to carefully monitor all outgoing telephone calls for three-way call attempts. Without such monitoring, many of the system's control features of a telecommunications system can be rendered ineffective. Currently, there are systems and methods known in the art for detecting three-way call attempts. Many of these systems however, are inaccurate and subject to both false positives and false negatives. Also, many of these systems are effective only in certain types of telecommunications systems.
For example, one such system known in the art for detecting three-way call attempts monitors for pulses of energy indicative of a hook-flash by detecting the frequency of the energy pulse to determine if it is characteristic of a hook-flash (i.e., a three-way call attempt). Specifically, the system includes a low pass filter for passing energy signals having frequencies below 500 Hertz (“Hz”), preferably in the range of 100 to 300 Hz, and an energy detector for detecting specific electrical energy pulses passing through the filter and having a predetermined minimum magnitude. The system also includes a software window analyzer, which cooperates with the energy detector to detect specific events, such as sound, occurring on the telephone line during a predetermined time window after the detection of the aforementioned energy pulse. The software window analyzer includes a timer means that is activated by the detection of the energy pulse, and a sound means for detecting the occurrence of sound on the telephone line during at least one of multiple windows of time defined by the timer means. The non-occurrence of sound on the telephone line during a specified time window is used by the system to confirm that the detected energy pulse is in fact a three-way call attempt. A counter means is further implemented for counting specific energy pulses detected by the energy detector during the time window when the remote party is using a pulse-dial telephone. This system, by simply monitoring for a pulse composed of certain frequencies, is often inaccurate and cannot operate in digital systems.
A similar system is also designed to detect the presence of an energy pulse indicative of a hook-flash. Specifically, the system is designed to detect a pulse that is comprised of frequency components below 500 Hz and above a predetermined threshold. The existence of the hook-flash is confirmed by digital signal processing equipment which identifies a rapid drop-off in energy, which is indicative of a hook-flash signal. Optionally, the hook-flash may be further confirmed by including software for cooperating with the energy detector to ascertain whether sound has occurred in the telecommunication during a predetermined period following the first hook-flash signal.
Still another known system includes three-way call detection circuit that uses digital signal processing to identify a third party connection. The system operates by establishing a baseline background noise. The system identifies a drop in noise level below the established baseline background noise as an indication that a three-way conference call has been attempted by the called and/or calling party.
Yet another known system monitors all connected telephone lines for indicia representative of a three-way call attempt. For example, the system may monitor for a digital PCM signal or a period of silence, followed by a release pulse, followed by yet another period of silence. Upon detection of a possible three-way call attempt, the three-way call detection circuit examines the digital signals to determine the spectral characteristics (i.e., time duration, frequency, and energy level) of a suspected release pulse of the suspected three-way call attempt. The system utilizes pattern recognition techniques to compare the suspected release pulse with a reference release pulse indicative of a three-way call attempt. The system also monitors for periods of silence before and after the suspected release pulse. If the system finds that the suspected release pulse is substantially similar to the reference release pulse and that the correct periods of silence surrounding the suspected release pulse are present, the system responds to the detection, for example, by disconnecting the telephone call, playing a recording, or creating a record of the three-way call attempt.
Yet another known system for detecting three-way calls monitors audio signals for features that distinguish voice and line-generated audio signals from audio signals produced by events associated with three-way call attempts. The distinguishing features used are pulse patterns that are strongly correlated with either audio signals generated by central office switching activity (‘clicks’) (reference features) or voice-generated audio signals (reset features). Audio signals are continuously monitored for reference and reset features over selected intervals or sampling windows. Sampling windows are reset whenever reset features are detected in the associated audio signal segment. Audio signals that are free of reset features and include reference features are tagged as potential click events. A three-way call event is declared when audio signals associated with consecutive sample windows are tagged as potential three-way call events. In this system, a control program samples the audio signal at the selected rate and sorts the sampled signals during a sampling window to produce a profile of the sampled audio signal. The profile comprises counters for tracking the number, strength (loudness), and separation of signal pulses. These counters may be compared in various combinations with counter values extracted from voice-generated audio signals (reset thresholds) and three-way call generated audio signals (reference thresholds) to declare a three-way call attempt, continue sampling, or reset the sampling window.
Another known system counts signal characteristics to detect three-way call attempts. The system samples audio from a telephone conversation, sorts the sampled signals into a profile of levels for the sampled audio signals, and monitors the profile of sampled audio signals for reset and reference conditions. In this system, a reset condition is a pulse pattern inconsistent with patterns generated by three-way call events. Reference conditions, in contrast, are pulse patterns identified from sampled audio signals that are consistent with patterns generated by three-way call events. If a reference condition is detected, the telephone call is tagged as having a possible three-way call attempt. The system concludes that a three-way call attempt has occurred when two consecutive tags have been made to the same telephone call.
Still another known system detects three-way calls by recognizing that each telephone connection has a characteristic reflection, or echo, idiosyncratic to that connection. The echo characteristics of a particular telephone connection are altered, for example, when a three-way calling feature is activated by the remote party at the original destination thereby adding a third party at a secondary destination. The system includes means for “zeroing out” or canceling the characteristic echo once a connection has been established by using an adaptive finite impulse response (FIR) filter. The system also includes response means for implementing a predetermined response when an undesirable event is detected. Examples of the responses which can be pre-programmed include call termination, playing a prerecorded message, generating a tone which may be heard by one or more parties to the call, muting the microphone of the local telephone and recording the date and time of the remote party's attempt to initiate the three-way call.
Other systems are known which incorporate methods of monitoring calls in telecommunications management systems. For example, the methods include means for detecting tones commonly associated with call bridging and call forwarding attempts. One such method is directed to the detection of tones such as ring signals, busy signals, special information tones (“SIT tones”), dual tone multi-frequency tones (“DTMF”), call progress tones or other similar tones characteristic of the placement of a telephone call.
Yet another known system incorporates spread spectrum techniques to detect three-way calls. The system measures delay times associated with multiple echoes of a reference signal transmitted over a two-way call. This initial echo characteristic is measured and recorded in an initial echo profile. Whether a three-way calling event has occurred is determined by virtue of changes in the delay times and number of echoes in each subsequent echo profile when compared with the initial echo profile. In view of the foregoing, a need clearly exists for a method and system of three-way call detection capable of accurately detecting three-way call attempts in analog and digital telecommunications systems. This method and system may be used in conjunction with any current system, but is preferably implemented within a system that detects three-way call attempts by analyzing the communications path between the originator and recipient in a telecommunications network.
The present invention embodies a three-way call detection circuit for use with an existing telephone management system, and is designed to reduce the number of three-way call attempts not detected by current three-way call detection techniques. The system of the present invention may be implemented in a variety of facilities including, but not limited to, penal institutions, mental institutions, nursing homes, rehabilitation centers, correctional facilities, government agencies, private and public business and the like.
Typically, a telephone management system used by such facilities consists of a multitude of telephones connected to a switchboard device. The switchboard device routes calls, performs voice prompts and responds to menu selections. Telephone calls placed by users of the telephone management system are routed through the switchboard device and connected to the proper outgoing trunk based on the type of call placed (e.g., collect, debit, etc.). An integrated cross point switch enables any telephone to access any available outgoing trunk.
The three-way call detection circuit of the present invention is utilized each time a telephone call is placed by a user of the telephone management system. The circuit constantly monitors all active trunk lines and telephone conversations. During a telephone call, the three-way call detection circuit monitors the connection for pulses of energy associated with the act of the called party initiating a three-way call. The system of the present invention monitors the presence of audio signals generated by the central office switching activity (hereinafter, “clicks”) indicative of a three-way call initiation attempt.
For a called party to initiate a three-way call, the called party typically depresses the hook-switch momentarily to put the calling party on hold and to call a third-party. The called party's depression of the hook-switch generates a hook-flash signal, which results in the central office generating a click on the inmate's telephone line. It has been shown that in certain instances a user can circumvent current three-way call detection systems by “covering up,” “masking,” or otherwise hiding a three-way call attempt by creating a constant or continuous noise (e.g., a constant hum or a constant hiss) while the three-way call is being attempted. Prior art systems are not designed to nor are they capable of detecting such a continuous noise. Thus, users of prior art systems can bypass the system simply creating continuous noise during a three-way call attempt.
The present invention provides for a method to detect circumvention attempts during a three-way call attempt. Specifically, during each call, the system monitors for periods of silence (e.g., by examining a number of samples from an audio stream to determine whether the sample is below a certain, pre-determined threshold). When the system preferably detects that the samples are below the pre-determined threshold, it continues to monitor the audio samples until the pre-determined threshold is exceeded. After the pre-determined threshold is reached, the system continues to examine the audio stream for the next period of silence. The system then determines the amount of time between the periods of silence to determine whether it exceeds the maximum allowable duration for continuous audio. If the elapsed time is greater than a pre-determined maximum duration, the system determines that an attempt to circumvent the three-way call system has occurred and appropriate action is taken.
As a further check to avoid potential false positives, the present invention provides a system and method to monitor signal power of the samples during the period when the pre-determined threshold is exceeded. A false positive occurs when the signal power is not evenly distributed across the period during which the pre-determined threshold is above the maximum allowed.
Therefore, it is an object of the present invention to provide a three-way call detection method and system for detecting three-way call circumvention attempts.
It is another object of the present invention to monitor the signal power of the audio signal for detecting three-way call circumvention attempts.
Furthermore, it is an object of the invention to accurately detect three-way call attempts and respond with a designated action (e.g., disconnect, flag, record, monitor, etc.).
It is another object of the invention to provide a three-way call detection method and circuit which stores all detected three-way call attempts in a central database.
It is still a further object of the invention to provide a three-way call detection method and circuit capable of monitoring a telephone conversation from the called party's side of the connection.
Additionally, it is an object of the invention to provide a three-way call detection method and circuit which is compatible with pre-existing telephone management systems.
Finally, it is a further object of the invention to provide a three-way call detection method and circuit compatible with both analog and digital telecommunications systems.
Other objects, features, and characteristics of the present invention, as well as the methods of operation and functions of the related elements of the structure, and the combination of parts and economies of manufacture, will become more apparent upon consideration of the following detailed description with reference to the accompanying drawings, all of which form a part of this specification.
A further understanding of the present invention can be obtained by reference to a preferred embodiment set forth in the illustrations of the accompanying drawings. Although the illustrated embodiment is merely exemplary of systems for carrying out the present invention, both the organization and method of operation of the invention, in general, together with further objectives and advantages thereof, may be more easily understood by reference to the drawings and the following description. The drawings are not intended to limit the scope of this invention, which is set forth with particularity in the claims as appended or as subsequently amended, but merely to clarify and exemplify the invention.
For a more complete understanding of the present invention, reference is now made to the following drawings in which:
As required, a detailed illustrative embodiment of the present invention is disclosed herein. However, techniques, systems and operating structures in accordance with the present invention may be embodied in a wide variety of forms and modes, some of which may be quite different from those in the disclosed embodiment. Consequently, the specific structural and functional details disclosed herein are merely representative, yet in that regard, they are deemed to afford the best embodiment for purposes of disclosure and to provide a basis for the claims herein, which define the scope of the present invention. The following presents a detailed description of the preferred embodiment of the present invention.
Referring to
Three-way call detection circuit 101 monitors connection 107 through interface 113. Specifically, interface 113 receives audio signals from connection 107 through monitor connection 112. In turn, interface 113 provides the signals to three-way call detection circuit 101 through interface connection 114. Alternatively, three-way call detection circuit 101 can receive data directly from connection 107.
In an alternative configuration, three-way call detection circuit 101 can monitor connection 102 between inmate telephone 103 and telephone network 105. If, for example, inmate telephone 103 is in an institution such as a prison, nursing home, school, detention center, hospital, etc., this enables three-way call detection circuit 101 to be internal to the institution. Three-way call detection circuit 101 is compatible with institutional telecommunications systems such as those in prisons, nursing homes, mental institutions, etc. Therefore, host computer 115 may be any computer in a telecommunications system, including a host computer in one of the institutions listed above. In these types of institutions, it is important to monitor all telephone calls for the presence of three-way call attempts to prevent, among other things, inmates or residents from accessing blocked or restricted telephone numbers.
Three-way call detection circuit 101 (which will be discussed in more detail below with respect to
Turning next to
During operation, three-way call detection circuit 101 monitors the telephone line communication between an inmate or resident and a called party by receiving audio data from interface 113 through connection 114 (see
Signals from connection 114 are received at interface port 211 and transmitted to both A/D converter 201 and energy detection circuit 206. A/D converter 201 converts analog telephone line data to a digital signal compatible with microprocessor 203 and digital signal processor 205. As will be discussed with respect to
Referring next to
Peak detector 309 isolates energy pulses in the filtered signal that exceed a predetermined magnitude. In one embodiment, the predetermined magnitude is approximately 6 Decibels (dB), although other magnitudes may be chosen in accordance with the invention. When such a pulse is detected, the output of peak detector 309 is driven high and a signal is sent to threshold detector 311, which is comprised of operational amplifier 308, and resistors 310 and 312. Preferably, resistors 310 and 312 are both 10 kΩ resistors. If a pulse is provided to threshold detector 311, it passes the signal to pulse stretcher 313. If no pulse is detected, threshold detector 311 does not output the received signal.
Preferably, pulse stretcher 313 is used to maintain the output of threshold detector 311 at its high level for 20 milliseconds. However, pulse stretcher 313 may be configured to maintain the output of threshold detector 311 for any time period. The stretched signal is then output on energy detect line 317 and analyzed by microprocessor 203 to determine if an energy pulse consistent with a three-way call click has been found (i.e., if the pulse has a magnitude of approximately 6 dBs or greater).
In one embodiment, three-way call detection circuit 101 detects continuous noise (i.e., power levels in a continuous noise audio stream) generated by a user in an attempt to circumvent three-way call detection. As such, the three-way call detection circuit 101 examines the audio data from the calling party as well as the called party to identify an initial period of silence followed by continuous noise and a second period of silence (also called ‘three-way’ event). The three-way call detection circuit 101 includes a digital signal processor 205 comprising a “continuous noise detection” algorithm (or CND Algorithm) for identifying areas in the audio stream where the signal level is continuously above a certain threshold. However, the CND algorithm may identify some parts of the audio stream as three-way events (i.e., silence, noise followed by silence), where no three-way event had been attempted. These events are characterized as ‘false positive’ and the CND Algorithm includes a Power Distribution Filter (PDF) Algorithm as well as a Zero Crossing Filter (ZCRF) Algorithm to identify these false positives. These Algorithms will be described in detail below.
Referring next to
Electronic switchboard device 405 regulates calls and connects them to the proper outgoing trunk line 411. Trunk line 411 may consist of a multitude of connections to any number of local, long distance, or international telephone service providers. The number of trunk lines 411 depends on the outgoing capacity desired by the institution. In addition, trunk lines 411 may be analog, digital, or any other type of trunk lines not yet contemplated. Electronic switchboard device 405 further incorporates an integrated channel bank, allowing calls to be processed over either analog or digital trunks as required by the telephone call system 401. Specifically, when one trunk line 411 is occupied and handling an outgoing communication, electronic switchboard device 405 automatically accesses an alternate trunk line 411 to handle the outgoing communication. If all trunk lines 411 on the system are in use, the call may be routed to an alternate system (not depicted). For example, electronic switchboard device 405 may be interconnected to a multitude of switchboards to allow for expansion of the system to meet the capacity desired by the institution. A cross point switch integrated into electronic switchboard device 405 may also accomplish this routing.
Multiple processors may also be incorporated into the architecture. This allows call processing even after parallel component failure. The architecture also provides for a sharing of the load between processors, which eliminates system overload during extremely busy periods. The multiple processors enable the system to handle large volumes of calls at any time, and ensure system integration.
Additionally, electronic switchboard device 405 performs the voice prompts heard by the calling party and the recipient of the call allowing the parties to respond to the menu selections. Electronic switchboard device 405 tests outgoing trunk lines as calls are placed and digitizes telephone audio for recording and/or biometric voice identification purposes. If no dial tone is present, one of trunk lines 411 may be taken out of service for a pre-programmed amount of time for maintenance. These capabilities are pre-programmed into the device's firmware. However, it is foreseeable that software and software upgrades may provide these services in addition to other services useful in the present invention.
A central site server 413 interfaces within the telephone call system 401 via a first serial port 415. In the preferred embodiment of the present invention, an RS-232 serial port is employed for the interference connection. However, it is foreseeable that other types of serial ports 415 commonly known in the art may be utilized. Serial port 415 may also be comprised of a direct hardware connection or may consist of a series of ports and connecting means commonly known in the art for connecting electronic devices. Serial port 415 is designed to allow firmware driven systems, such as electronic switchboard device 405, to interface with software-based systems, such as a PC designed system operating as a site server. All inmate and call information is routed through central site server 413. At central site server 413, user call information is digitized for efficient data transfer and efficient record keeping. Central site server 413 stores at least each user's financial transaction data. It is preferred that central site server 413 also stores the digitized audio used for voice prompts as well as each user's call restrictions, PIN, biometric verification data, etc. However, depending on the memory requirements, numerous site servers may be employed. It is foreseeable that older archived data may also be stored on an integral or a remote computer system database (not shown) or kept on additional storage devices on the central site server 413.
Three-way call detection circuit 101 is utilized each time a telephone call is placed utilizing telephone call system 401. Three-way call detection circuit 101 is connected to telephone bank 403 and constantly monitors all active trunk lines 411 and telephone conversations. During a telephone call, three-way call detection circuit 101 monitors the connection and looks for three-way call circumvention or masking attempts.
Referring now to
The process begins with step 500 and is followed by step 501 whereby the audio streams of both parties on the call are monitored. Preferably, the process examines a pre-determined sample size of the audio streams, such as in one non-limiting example, a sample size of 512 bytes, to determine whether the audio of the sample is below a pre-determined threshold (e.g., −20 dBm) (step 503). The process continues to loop until a sample falls below the pre-determined threshold (as shown in the loops of steps 501 and 503), and this event signifies a silence event. If the predetermined threshold is not met (i.e., the amplitude of the sample is less than the pre-determined threshold), the process continues to monitor the streams until the audio stream exceeds the pre-determined threshold (as shown in the loop of steps 505 and 507).
Once the pre-determined threshold is reached, the process starts a timer (T1) and continues to monitor the stream until the audio stream again falls below the pre-determined threshold (as shown in the loop of steps 509 and 511). Upon detection of the audio stream falling back below the pre-determined threshold, the process stops the timer (T2) and compares the elapsed time (T2-T1) to a pre-determined acceptable duration (step 513). In one non-limiting example, the pre-determined acceptable duration is set to a maximum allowable length of continuous audio not exceeding 1500 ms (milliseconds). If the elapsed time (T2-T1) is greater than the acceptable duration, the system determines that an attempt to circumvent the three-way call system has occurred and appropriate action is taken (step 515). If the elapsed time (T2-T1) is less than the acceptable duration, the process starts over (step 501) for the remaining duration of the call.
In another alternate embodiment as shown in
The method starts in step 600 and is followed by step 601 (step 601) whereby the audio streams of both parties on the call are monitored. Preferably, the process examines a pre-determined sample size of the audio streams, such as in one non-limiting example, a sample size of 512 bytes to determine whether the audio of the sample is above a pre-determined threshold (e.g., −20 dBm). If the sample does not rise above the predetermined threshold, the circuit 101 continues to loop (as shown in the loop of steps 601 and 603). When a maximum threshold is reached in step 603, the method proceeds to step 605 where circuit 101 starts a timer which continues for a set period of time (T). Step 605 is followed by step 607 and in this step, if the sample remains above the pre-determined threshold for the entire period T, the system 101 determines that an attempt to mask a three-way call attempt has occurred and appropriate action is taken in step 609. If the circuit 101 determines that the sample fell below the pre-determined threshold, preferably for at least a period P during period T (where T>P), the system determines that no three-way call circumvention attempt has occurred and continues to monitor for events indicative of an attempt to mask a three-way call attempt (step 601).
Using a similar method, a three-way call can be detected based on the continuous noise of a ringer during the original call. For example, the method starts in step 600 and is followed by step 601 (step 601) whereby the audio streams of both parties on the call are monitored. Preferably, the process examines a pre-determined sample size of the audio streams, such as in one non-limiting example, a sample size of 512 bytes to determine whether the audio of the sample is above a pre-determined threshold. In this embodiment, the pre-determined threshold should be set so as to detect a hook-flash signal. In an embodiment, the threshold detector can be substituted with a waveform analyzer or other hook-flash detector to specifically detect a hook-flash signal. If the sample does not rise above the predetermined threshold, the circuit 101 continues to loop (as shown in the loop of steps 601 and 603). When a maximum threshold is reached in step 603, the method proceeds to step 605 where circuit 101 starts a timer which continues for a set period of time (T). In an embodiment, the period of time (T) is set to a duration sufficient to allow for a successful third-party connection, and for at least one ring to be completed, such as for example 1 s.
During this period, the circuit 101 monitors the signals to detect continuous noise indicative of a ringer. For example, ringers are generally represented by repeating 200 ms of continuous noise followed by 400 ms of silence. Therefore, the circuit 101 can be configured to monitor for the continuous noise. In an embodiment, the duration threshold can be set to be slightly less than the duration of a ring, such as for example 150 ms. In an embodiment, the duration threshold is set to be greater than 100 ms and less than 200 ms. This helps to ensure adequate detection of the ring, while reducing false positives from other noise sources. In an embodiment, the circuit 101 can be additionally/alternatively configured to detect the repeating noise/silence pattern. This detection can begin with the power level of the audio signal exceeding a predetermined threshold. Because more than a single ring is being detected, the duration of the detection time period should be extended in this embodiment to a length sufficient for the connection of the call as well as to capture at least more than 1 period of the ringer pattern, such as for example 1.6 s. Once initiated, the circuit 101 will examine the power level of the signal to see if it substantially conforms to the noise patterns indicative of a ring. Based on a correlation of the detected noise pattern with the expected noise pattern, a determination can be made as to whether a ring has been detected. In an embodiment, the circuit 101 can additionally/alternatively detect the ringer based on a particular frequency associated with the ringer. In an embodiment, this can be performed by computing the Fourier transform of the audio stream during the noise portion, and compare the result to an expected frequency value.
Step 605 is followed by step 607 and in this step, if the sample meets the criteria of the monitoring described above, the circuit 101 determines that an attempt to make a three-way call attempt has occurred and appropriate action is taken in step 609. If the circuit 101 determines that the sample did not meet the criteria of the monitoring described above, the system determines that no three-way call attempt has occurred and continues to monitor for events indicative of an attempt to make a three-way call attempt (step 601).
Referring now to
Once the pre-determined threshold is reached, the process continues to monitor the stream in step 711 until the next sample in the audio stream again falls below the pre-determined threshold (silence detected). Upon detection of the audio stream falling back below the pre-determined threshold in step 713, the process increments silence count value by one in step 715. Otherwise, silence count value is reset in step 717. Next, in step 719, if the sample in the audio stream is in the silence threshold, them the sample is saved as “End of Noise Sample” in step 721. Next, in step 723, if the difference between “Beginning of Noise sample” and End of Noise Sample is within the minimum and maximum length for continuous noise, the method progresses to step 725 where a message is played to either party that they are in violating of making a three-way call and the telephone system 100 proceeds to disconnect the call. However, if step 723 is NO, then the method progresses to step 705 where the next sample is compared to a predetermined silence threshold. The method ends in step 729.
Also, and as was stated earlier in reference to the Continuous Noise Detection Algorithm, the circuit 101 may also monitor signal power to avoid false positives. Particularly, three-way call detection circuit 101 examines the audio data from the calling party as well as the called party to identify an initial period of silence followed by continuous noise and a second period of silence (also called ‘three-way’ event). The three-way call detection circuit 101 includes a digital signal processor 205 comprising a “continuous noise detection” algorithm (or CND Algorithm) for identifying areas in the audio stream where the signal level is continuously above a certain threshold for the time period T1, which in one non-limiting example, is set at 1500 milliseconds. However, the CND algorithm may identify some parts of the audio stream as three-way events (i.e., silence, noise followed by silence), where no three-way event had been attempted, such as for example, loud continuous speech. These events are characterized as ‘false positives’. Furthermore, the CND Algorithm may include a plurality of Algorithms, such as a Power Distribution Filter (PDF) Algorithm and a Zero Crossing Filter (ZCRF) Algorithm to identify these false positives. The CND Algorithm triggers the three-way event candidate as an output set of line PCM coded samples of audio with audio duration greater than or equal to 1500 milli seconds. The PDF and ZCRF Algorithms operate on part of the PCM coded sample set. The optimal sample length used for filtering was selected as trailing 60% of the three-way event sample candidate; this resulting set of samples is further referred to as ‘Audio Segment’.
The PDF Algorithm detects events that do not have signal power evenly distributed across the reported CND event time interval T2-T1. The PDF Algorithm splits the Audio Segment into a configurable number of segments, preferably of equal length. The system 100 calculates the cumulative power (i.e., the sum of absolute coded signal amplitudes for each part). The cumulative power is normalized by the length of a part. The PDF Algorithm calculates the ‘distance’ (i.e., a difference of resulting average amplitude values) between values of each part and ensures that the maximum distance value is lower than a predefined configurable parameter. All Audio Segments that fail the PDF check are declared as false positive three way events.
The Feature Zero crossing Rate (ZCRF) Algorithm filters out all three-way events where the Zero Crossing Rate (ZCR) is below a certain threshold that characterizes human speech. The ZCRF Algorithm calculates the number of zero-crosses (X axis crosses) of a waveform signal averaged by the number of samples in the Audio Segment. If the resulting value is greater than a preconfigured reference value, such Audio Segment is declared as a false positive three-way event.
While the present invention has been described with reference to the preferred embodiment and several alternative embodiments, which embodiments have been set forth in considerable detail for the purposes of making a complete disclosure of the invention, such embodiments are merely exemplary and are not intended to be limiting or represent an exhaustive enumeration of all aspects of the invention. The scope of the invention, therefore, shall be defined solely by the following claims. Further, it will be apparent to those of skill in the art that numerous changes may be made in such details without departing from the spirit and the principles of the invention. It should be appreciated that the present invention is capable of being embodied in other forms without departing from its essential characteristics.
This is a continuation application of U.S. patent application Ser. No. 13/958,137, filed Aug. 2, 2013, which is a continuation-in-part application of U.S. patent application Ser. No. 12/378,244, filed Feb. 12, 2009, now U.S. Pat. No. 8,630,726, issued Jan. 14, 2014, which are incorporated herein by reference in their entirety.
Number | Date | Country | |
---|---|---|---|
Parent | 13958137 | Aug 2013 | US |
Child | 14919402 | US |
Number | Date | Country | |
---|---|---|---|
Parent | 12378244 | Feb 2009 | US |
Child | 13958137 | US |