MXPA96001658A - Microphone selection process for use in a multiple microphone switching system, powered by the - Google Patents

Abstract

The present invention relates to audio systems and, more particularly, to systems for selectively connecting voice frequency circuits to an audio line in response to video signals.

Description

MICROPHONE SELECTION PROCESS FOR USE IN A VOICE-POWERED MULTI-MICROPHONE SWITCHING SYSTEM TECHNICAL FIELD This invention relates to audio systems and, more particularly, to systems for selectively connecting talk circuits to an audio line in response to speech signals.

BACKGROUND OF THE INVENTION Many companies now consider teleconferencing as an inexpensive means of communication between staff in dispersed locations and thus reducing the need for business travel. In an audio teleconferencing arrangement, a number of speakers in one place are put in communication with a number of speakers in one or more remote locations via a telephone connection. The quality of transmission between the separate groups of speakers generally depends on the position of each speaker with respect to a microphone and the loudspeaker device in each location. With a single microphone and a loudspeaker device in a room in the place where the conference is held, the transmission is subject to degradation because some of the speakers are generally at a greater distance from the optimal microphone and loudspeaker device. It is well known that the use of a plurality of microphones properly distributed in each place where the conference takes place, improves the quality of the conference system. The outputs of the microphones are summed and the summed outputs are applied to a communication link established between the places. In such an arrangement, each speaker can be within an acceptable distance from one of the microphones, so the pickup of the vocal frequency is of relatively good quality. With all the microphones turned on at the same time, however, several undesirable effects occur. The total noise pickup is much higher than for a single microphone. The artificial reverberation effects caused by the pickup of delayed signals from the remotest microphones severely reduce the quality of the conference transmission. In addition, electroacoustic instability of the plurality of always-on microphones can easily result. It is therefore desirable and also known in the art to improve a switching arrangement that allows only the microphone closest to the speaker speaking to be activated so that the pickup of reverberation and noise is minimized. Such an arrangement is commonly known as a "voting circuit." In the "voting circuit" arrangement, the one who speaks loudest can capture control and block the other speakers in their place. This automatic switching between microphones, responds to the input of the highest level of vocal frequency that appears alternately in the different microphones, however, can also result in interruptions in transmission that adversely affect the intelligibility and can result in interference undesirable caused by the transitory noise of the room. For example, a loud noise in one of the places where the conference is made can completely interrupt the controller microphone. Furthermore, since only one microphone operates at the same time, the transfer of control from one microphone to another, such as that caused by the movement of the speaker who is talking from one position to another in a place in the room, can bring result in a variable quality vocal frequency transmission, interruptions in the transmission, and reverb effects that may vary with the position of the speaker who is speaking. Several arrangements for teleconferencing have been proposed and used previously to select a single microphone from a plurality of conference microphones and to transmit the signal only from the selected microphone. Such arrangements are described in, for example, US Patent No. 3,730,995 issued to M. V. Matthe s on May 1, 1973, US Patent No. 3,755,625 issued to D. J. Maston on August 28, 1973, US Patent No. 4,449,238 issued to B. H. Lee, et al. on May 15, 1984, and US Patent No. 4,658,425 issued to SD Julstrom on April 14, 1987. Another example of a teleconference arrangement is described in co-pending US patent application serial number 08/239771 filed on May 9, 1994 in favor of DJ Bowen and commonly awarded to the same beneficiary along with this application. In this copending application, a voice-activated switching arrangement is provided for the selection of one or more microphones according to the levels of the output signal of each of the microphones. Also, the voice-activated switching arrangement described in the copending application employs directional microphones to reduce the degradation of the speech frequency signals due to the pickup and reverberation of noise. These directional microphones are located in a common circular housing and have sensitivity response patterns that extend outward from the center of the housing. The voice-activated switching arrangement also employs an algorithm or voting process to select the actuation of the appropriate number of those microphones to effectively detect each person speaking in a room. The voice-operated switching arrangements described above have been satisfactory in minimizing the degradation of the speech frequency signals due to reverberation and noise pickup. It has also been satisfactory to make the microphone selection technique appear to occur in a very normal way, without, for example, trimming the syllables when the microphones go from the off to the on state. However, it is desirable to simplify the execution of the microphone selection technique so that this technique can be performed in a limited amount of processing time. Such simplification could release a processor more regularly from other necessary calculations or allow the use of a less powerful and cheaper processor in the switching arrangement.

BRIEF DESCRIPTION OF THE INVENTION In accordance with the present invention, a microphone selection process becomes relatively constant in terms of the processing requirements through the use of the combination of values that provide a measure of the quality of the received speech frequency signal in each of a plurality of microphones. Such combinations of values are derived in such a way that they provide an indication of the microphone that receives the voice frequency signal better. Each of the microphones has a pattern of supercardioid responses, and collectively the microphones are placed to provide coverage of the entire area of a typical conference room.

According to one aspect of the invention, the microphone selection process selects the microphone that best receives the voice frequency signal by comparing a value of received signal energy in each of the microphones with that received in each of the other microphones. More specifically, the pairs of microphones are examined, to determine the direction of origin of the vocal frequency, looking for a pair of microphones where the vocal frequency is strong in the forward-facing microphone, ie the microphone directed towards the microphone. vocal frequency source, and weak in an associated backward-facing microphone, ie the microphone directed away from the source of the vocal frequency. Since a null, which is located on the back of each microphone, is narrower than a beam or main sensitivity pattern, which is located on the front of each microphone, this null is more sensitive to and therefore a better indicator of the direction of origin of the vocal frequency than the main beam. The combination of the energy values of the signal for a forward-facing microphone and its associated backward-facing microphone respectively provides advantageously a specific combination value which is compared with each of the other microphone pairs of the switching arrangement. The pair of microphones having the best combination value is then easily determined by identifying and selecting the microphone that best receives the voice frequency signal.

BRIEF DESCRIPTION OF THE DRAWINGS The invention and its mode of operation will be more clearly understood from the following detailed description when read with the attached drawings in which: FIGURE 1 is a block level diagram of the arranged conference microphone circuit, arranged in accordance with the present invention; FIGURE 2 is a top plan view of a conference arranged housing for housing the microphone circuit shown in FIGURE 1; FIGURE 3 is a front view of the arranged conference housing shown in FIGURE 2; FIGURE 4 is a teleconferencing system in which the present invention can be oyed; FIGURE 5 is a flowchart of a process suitable for incorporation into the digital signal processor shown in FIGURE 1, according to the invention; FIGURE 6 is a flowchart of a process that shows in greater detail a portion of the process shown in FIGURE 5; Y FIGURE 7 is a flow diagram of a process that shows in greater detail a portion of the process shown in FIGURE 5.

Through the drawings, when the same elements are shown in more than one figure they are designated by the same numerical reference.

DETAILED DESCRIPTION OF THE INVENTION Referring now to FIGURE 1, a block level diagram of the microphone circuit arranged for conference (CAM) 100 is shown. Included in the CAM 100 circuit is a digital signal processor (DSP) 110, five separate input circuits which consist of the amplifiers 121 to 125 and the respective associated linear CODECs 131 through 135. Each of these input circuits is associated with each of the first-order gradient microphones contained in a CAM housing 200 shown in FIGURE 2 and described hereinafter. The CAM 100 circuit also includes a selection logic circuit 140 for selecting each of the five input circuits to respectively provide its microphone signal to the DSP 110 via five serial-in-parallel-out (SIPO) or serial transducers. a in parallel 141 to 145. The output of the DSP 110 is provided to an output circuit comprising a linear CODEC 150 and an output amplifier 151. The DSP 110 and the linear CODECs 131 through 135 and 150, and selecting logic 140 all receive synchronization information from a synchronization circuit 153. Five light-emitting diodes (LEDs) 152-1, -2, -3, -4, -5 are included in the CAM 100 circuit to provide a visual indication of the initial calibration of the CAM 100 circuit as well as to provide a general visual indication to the individuals present in the conference room of which general area of the room is being covered by the microphone or microphones selected by the CAM 100 circuit. In operation, each analog input signal of each microphone input to the CAM 100 circuit is amplified respectively by one of the linear amplifiers 121 to 125. Amplifiers suitable for use as amplifiers 121 to 125 are commercially available. Such an amplifier is the MC34074 unit available from, for example, Motorola. From each amplifier 121 to 125, the associated analog signal is coupled respectively in the 16-bit linear CODECs 131 to 135 where each analog signal is digitized. CODECs suitable for use as CODECs 131 to 135 are commercially available. Such CODEC is the AT &amp unit; T7525 available from, for example, AT &T Corp. Economical Law CODECs are also available that adequately provide the desired functions required from CODECs 131 through 135 and 150. From CODECs 131 to 135, each signal 16-bit digitized is serially loaded in two cascaded 8-bit serial to parallel registers. Five pairs of these cascaded registers respectively comprise the serial to parallel (SIPO) converters 141 to 145. Serial to parallel converters suitable for use as converters 141 to 145 are known in the art and are available from, for example, Motorola as part number MC74299. The microphone input signals are weighted and summed by the DSP 110 to form the desired unit microphone output signal. The DSP 110 may, illustratively, comprise physical elements or hardware of the digital signal processor such as the DSP 16 or DSP32C of AT &T Corp. together with the read-only memory (ROM) for storing the programs and programming systems or software, which perform the processing operations described hereinafter, and the random access memory (RAM) for storing the results of the DSP. Through the use of the selection logic circuit 140, the DSP 110 selects sequentially each of the ten registers from serial to parallel in cascade in converters 141 to 145 and reads in this data, 8 bits at a time through the lower 8 bits of its parallel port. The DSP 110 provides a control signal to the selection logic circuit 140 on the line 101 at the appropriate time to allow the selection logic circuit to activate one of the appropriate registers and therefore provide the correct 8-bit data signal to the DSP 110. Decoder circuits suitable for use as a selection logic circuit 140 are known in the art and are available from, for example, National Semiconductor as part number 74154. After the data entry signals of the five microphones are received in the DSP 110 and processed, as described in detail hereinafter, a 16-bit digital output signal is transmitted in series from the DSP 110 to the linear CODEC 150 in the microphone output circuit. The output signal of the CODEC is then amplified and conditioned by the amplifier 151 to provide a standard analog microphone output signal.

The microphone output signal is not limited to one or two microphone output signals, but is the weighted sum of all microphone output signals. A variable weight factor is assigned to each microphone and used to gradually turn on or off the signal from each selected or activated microphone that is connected to an audio line. The weight factor is typically higher for the selected microphones and zero for the non-selected microphones. Since these weighted factors are gradually adjusted, the selection of a microphone and the changes in the level of background noise are therefore less noticeable to the users. During the transition intervals in conversations the weight factor can be relatively greater for several microphones simultaneously. A linear CODEC suitable for use as CODEC 150 is available from, for example, AT &T Corp. as part number AT &T7525. A suitable amplifier for use as an amplifier 151 is available from, for example, Motorola as part number MC34074. The synchronization circuit 153 includes a 26 MHz crystal oscillator for the DSP 110 as well as a 2048 MHz signal used by the CODECs for data synchronization and transmission.

Shown in FIGURE 2 is a top plan view of a CAM 200 housing that includes an upwardly facing loudspeaker 210, microphones 220-1, -2, -3, -4, -5, and LEDs 152- 1, -2, -3, -4, -5 embedded in this accommodation. In the described embodiment, the housing of the CAM 200 is configured with a plurality of directional first order gradient microphones of the type described in US Patent 5,121,426 which was issued on June 9, 1992. These microphones are mounted in a shaped housing. of pentagon illustrated by the American Patent 327,479. The plurality of first order gradient microphones, illustratively shown as five, are placed in the pentagon or in the circular shaped housing generally oriented outward from the center of the housing and from the supercardioid response patterns. The array of microphones provides coverage of the entire area of a room which is very useful in a conference phone application. Since only one person speaks at a time during normal operation, background noise and reverberations are minimized by activating only the microphone that receives the best vocal frequency for that person.

According to the described embodiment, the circuits shown within FIGURE 1 are located within the housing of the CAM 200 and are arranged to compare the output signals of each of the microphones 220-1, -2, -3, -4 , -5 to determine which of one or more of those microphones is providing the strongest vocal frequency signals. In response, selected microphone signals or microphones are transmitted to a conference participant at a remote location without the reverberation that normally results when more than one microphone is activated. The loudspeaker 210 is located at the null of the polar response pattern of each of the microphones embedded in the housing 200. The null of the polar response pattern resides between the main lobe and the adjacent lateral lobe. This particular null is located at 125 ° -which take into account the particular position of the microphones around the perimeter of the housing 200. This operation is achieved by placing a microphone element, as described in the US Patent No. 5,121,426, in the accommodation, thus forming a supercardioid polar response pattern. Although only the polar response pattern associated by a single microphone 220-4 is shown in FIGURE 2, the response patterns of each of the microphones in the housing are identical. It should be noted that the housing and the microphones contained therein cooperate to determine the shape of the response pattern. FIGURE 3 shows a front view of the CAM 200 housing to illustrate the relative placement of three of the 220-2 microphones, 220-3 and 220-4, and to demonstrate that such units can be packaged in an attractive way in a low profile product. Shown in FIGURE 4 is a modality of the teleconferencing system that includes the CAM 200 housing placed in the center of a conference table 405. The CAM 100 circuit, incorporated in the housing of the CAM 200 is connected to a control unit 410 in the system by means of a cable 401 which can either pass through the table 405 via a hole drilled in it or can rest on the top of the table. This cable contains the appropriate connections for transmitting both the microphone output signal from the CAM housing 200 to the control unit and the input signal to the loudspeaker 210 from the control unit 410. The cable also includes the connections for transmitting power to the control unit. a conventional electrical power source (not shown) in the CAM 100 circuit, which provides the electrical power for the circuit shown in FIGURE 1.

The control unit 410 is interconnected to an annular tip telephone line (not shown) via line 402 to provide the conventional telephone service to the teleconferencing system. The control unit receives the microphone output from the amplifier 151, as shown in FIGURE 1, and also directly provides an output signal for the loudspeaker 210, shown both in FIGURES 2 and 3. A control unit suitable for use as control unit 410 is described in US Patent No. 5,007,046 entitled "Computer Controlled Adaptive Service Microphone". This control unit provides an improved switched loss loss adaptive service microphone, which dynamically adjusts its switching thresholds and other operating parameters based on an analysis of the acoustic environment and the conditions of the telephone line. The control unit described in the referred patent receives an output from a microphone and provides an input to a loudspeaker to provide a service microphone array. The microphone output signal provided by the amplifier 151 is easily replaceable by the microphone shown in the described service microphone array. An alternative control arrangement suitable for use as a control unit 410 is described in U.S. Patent No. 5,016,271 entitled "Echo Canceller Suppressor Service Microphone". Almost all and all double operations are regularly achieved with this alternative control arrangement since the reception path remains open at all times and the transmission path has its gain reduced only to the level necessary to suppress the excess of reverberant return echo. Although the control unit 410 is shown as part of the CAM 100 circuit, it should be understood that such a control unit can also be integrated into the electronics within the housing of the CAM 200. Moreover, it should also be understood that the CAM 100 circuit, when used The well-known wireless telephone circuit, such as in the AT &T Corp. 5500 HT cordless telephone set, can also be assembled to obviate the need for any wiring between each other and a base unit or control unit that connects to the Ring tip telephone line. Such a suitable wireless telephone circuit is also described in U.S. Patent No. 4,736,404. For this wireless telephone circuit as well as for the CAM 100 circuit, a battery can be used to provide a suitable operating electrical power source. Referring next to FIGURE 5, there is shown a flow diagram illustrating the operation of the DSP 110 in executing the microphone selection operation. The functions provided by the DSP 110 are advantageously determined by a process or program stored in the associated read-only memory (not shown). The process is introduced in step 501 where the initialization parameters are set. As part of those parameters, the weight factor, described hereinafter, of any of the five microphones, illustratively 200-1, is set to 1, thereby effectively turning that microphone ON. When this microphone is ON, an initial cut of the syllables is not advantageously perceived by the speakers because some voice frequency signal will always be transmitted, even if this is attenuated due to the relative position of the microphone ON to the person who is talking. Other certain initialization parameters are executed according to U.S. Patent No. 5,007,046. Once this initialization is performed and verified in decision 502, the circuit is ready to input signal data and the process advances to step 503. During each sampling period or every 125 μs, each of the microphone inputs is sampled in step 503 to determine the maximum absolute values in the voice frequency energy input. Also in each sampling period, the input value for each microphone is adjusted according to its assigned weight factor and then the weighted outputs of all the microphones are summed into a common audio line. The maximum absolute values of the microphones are obtained from 16 samples during a cyclic period of 2 milliseconds (ms) to obtain the largest absolute maximum value that occurs within this period of time for each microphone. If during this cyclic period of 2 ms, a maximum value subsequently measured is greater than one previously measured and the maximum value stored, then the maximum value previously stored is replaced with the maximum value subsequently measured. If the maximum value measured above is greater than the maximum value subsequently measured, however, then the maximum value measured above is retained in the memory. The maximum absolute value for each of the five microphone inputs is thus determined in step 503 during each cyclic period. The 16 samples obtained during each cyclic period allow to trace the envelope of the signal for each microphone at 300 Hz, the lowest frequency of interest. If the 16 samples in the vocal frequency energy have not been measured for each microphone in step 503, as determined in decision 504, the process advances to step 505 where the weighted output is calculated for each microphone. This calculation is made according to the data processing speed or every 125 μs. If the CAM 100 has been recently activated, the initialization parameters, according to that provided in step 501, determine the weighted output and thus the input signal from the moment in which the initially selected microphone is connected to the line analog output at this point in the process. Once the initialization is completed, however, the microphones in the CAM 100 are configured either in the ON or OFF state or in the transition between those two states according to the acoustics present in the room. After 16 maximum input values have been measured in the vocal frequency energy for each microphone as determined by decision 504, the selected maximum input values are used to calculate a logarithmic value, for example, a calculation of logio or decibel, of the signal for each of the five microphone inputs in step 506. These logarithmic values, which simplify the calculations of the strength of the relative signals, are then used in step 507 to determine The energy of the relatively long and short term envelope for each of the five maximum inputs of the microphone, the determination of the energy of the long and short term envelope is described in greater detail hereinafter with reference to FIG. 6. FIG. The energy of the envelope determined in step 507 is used by an algorithm or voting process in step 508 to select the microphone signal inputs not that they will be passed through the exit. In executing the selection process in a described modality, the voting algorithm makes comparisons based on the maximum microphone signal, selecting either 1) the current microphone, 2) an opposite microphone; or 3) both the current and the opposite microphone, if their vocal frequency signal levels are relatively strong; or 4) under less restrictive criteria, the microphone with the strongest signal. Taken in the given order, each of the above comparisons is made in a less restrictive way than the preceding one. If the levels of the vocal frequency signal of the current and the opposite microphone are not strong enough, the voting algorithm can choose any microphone based on less restrictive thresholds. When the levels of the voice frequency signal are close to the level of background noise, the voting algorithm makes comparisons only between the microphones currently selected and the opposites, remaining with the selected microphone if the comparisons are not conclusive. Once the microphone inputs for activation or deactivation are selected in step 508, the variable weight factor for each microphone is updated in step 509 during each cyclic period of 2 ms and those weight factors are used below for determine the level of the signal for each microphone that remains connected to the output. Thus, according to its selection or non-selection, the output of a microphone remains, either ON, OFF, or a transmission occurs to one or the other of those two states in the calculation performed by step 505. As shown in FIG. noted, the output of the CAM 100 circuit is a weighted signal derived from all the microphones, not only those selected by the voting algorithm to be activated or configured ON by this algorithm. In this way, when a microphone is selected to be activated by the voting algorithm, its input is gradually added to or makes a greater percentage of the output signal. Similarly, when a microphone is not already selected or set to OFF after having been selected by the voting algorithm, its output is gradually removed from the output signal. The initial cut of the syllables is also not advantageously perceived because at least one microphone is left at all times, and the vocal frequency generated anywhere in the room will be immediately detected and transmitted, even if attenuated. The activation and deactivation weight factor for each microphone is shown by: Wi = Wi + 0.05 if the microphone., Is set to ON t = i - 0.01 if the microphone is configured OFF O = S Ii i i-l where: Wi is the weight factor for the microphone that has an interval between 0 and 1.0; Ii is one of the five microphone inputs, and O is the output value for the sum of each mic weighted signal.

In this way, a microphone that is turned on is activated five times faster than a microphone that is turned off. A major advantage of this activation and deactivation arrangement is that any background noise that is not removed by the noise removal process described below is less noticeable if it is added and removed together with the microphone signal. This arrangement also allows you to turn ON multiple microphones at the same time due to differences in the delays in the weight factors to activate and deactivate the microphones. In this way, any undesirable side effects of the voting algorithm quickly switch between the microphones, such as those caused in difficult switching, (by immediately returning a microphone completely on, or completely switched off) are eliminated. In this way, in fact, many people can talk respectively and activate different microphones at the same time. The degree to which each person continues speaking, your microphone will remain ON or activated. Referring now to FIGURE 6, there is shown a flow diagram illustrating the steps involved in obtaining the measurements of the relative strength of the signals for each of the microphones by the CAM 100 circuit. Those steps 601 through 604 they are all part of step 507 executed in FIGURE 5. Since the voting algorithm determines when one or more people are talking and then activates the microphone or microphones that best receive those vocal frequency signals, a critical component of this calculation is to determine correctly when the input signal of a microphone is of a vocal frequency and not of noise. The steps executed by the flow diagram of FIGURE 6 advantageously provide this information to be used by the voting algorithm.

The strength of the received signal is calculated as in step 601 by averaging the maximum absolute value selected for each microphone input, each maximum absolute value is selected from those occurring during a cyclic period of 2 ms. There is an average short and long term energy generated that represents the strength of the speech frequency signal and the strength of the noise signal respectively. The different averaging factors are selected depending on the slopes of the input values are positive or negative. When the slope is positive, the input values increase in force and when the slope is negative, the input values decrease or decrease in force. Both averages are calculated as 0.2In + (1-0.2) recsn-i if In > rec Sn-l recsn = 0.005In + (l-0.005) recSn-? yes In < recSn- í 0.00024In + (1-0.00024) recin-i if I "> rec? n-? recín = 1 0.025In + (l-0.025) rßc? "_? yes In < recln-l where: recs and reci are the averages of the respective short and long term signal; In is the value of the maximum signal for each input during the current cyclic period; and In-? is the value of the signal for each input during the previous cyclic period.

Both recN and recN quantities are used to calculate the strength of the speech frequency signal. The rec amount is a measure of the background noise. The rec3n quantity is a measure of the intermittent signals such as the voice or any other high-pitched noise, along with any background noise. As indicated in step 602, the strength of the speech frequency signal or the value of the energy of the tracked signal, rectn for each microphone is calculated by subtracting the long-term rec rec the short-term average in this way : recn = recsn - rec? n or VOCAL FREQUENCY = (VOCAL FREQUENCY + NOISE) - NOISE Since these are logarithmic values, the straight amount is not the difference in magnitude between the average values of the short and long-term signal, but rather the ratio of the magnitudes of those two values. The values of the tracked signal of each microphone are then classified as in step 603 to determine the maximum and minimum tracked signal energy values, RECMAX and RECMIN respectively, among all the microphones. The spread is then calculated, which is the difference between RECMAX and RECMIN, in step 604. Since the background noise level is effectively removed from each microphone input, the DIFFUSION must be close or zero when they are not. present intermittent signs. When the DIFFUSION is greater than zero for some threshold, then the voting algorithm interprets this as an indication that the voice frequency signal is present and then searches the value of the strength of the signal traced respectively for each microphone to determine the source of the vocal frequency signal. DIFFUSION is a measure used to indicate that an intermittent signal such as a voice frequency signal is present. In response to the presented input parameters, the selection process selects the microphone that best captures the sound or voice frequency signal. In the selection of this microphone, the values of the signal strength tracked for the microphones are compared with others. More specifically, we examine pairs of microphones, to determine the direction and origin of the vocal frequency, looking for a pair of microphones where the vocal frequency is strong in the forward-facing microphone, that is, the microphone directed towards the source of the vocal frequency, and weak in the microphone oriented backwards, ie the microphone directed away from the source of the vocal frequency. It is assumed that the vocal frequency is in the null of the microphone oriented backwards. The null of each microphone is narrower, and therefore more sensitive to the direction, than the main beam. The combination of the two microphones provides a better measure of the directionality of the voice frequency signal.

Referring now to FIGURE 7, there is shown in accordance with one embodiment of the invention, a flow chart showing the additional steps incorporated in step 508 of FIGURE 5 using the DIFFUSION, RECMIN and RECMAX values in the selection of the microphone or appropriate microphones to be activated. As indicated at the beginning, the voting algorithm determines if a voice frequency signal is present and selects the microphones, or beams, that optimally or better receive the voice frequency signals. It uses the values of the tracked signal for each microphone or beam, the beam pattern is indicative of a particular microphone, and the RECMAX, RECMIN and DIFFUSION values make the decisions. As also indicated at the beginning here, the microphones 220-1, -2, -3, -4, -5 are mounted in a pentagon-shaped housing as clearly illustrated in FIGURE 2. Thus, it is considered that each of the plurality of microphones has two opposed microphones. For example, the microphone 220-1 has two generally opposite microphones, the microphone 220-3 and the microphone 220-4. When the CAM 100 circuit is in its active or ON state, the relative input energy levels of each microphone input are determined and either one or two microphones are selected and remain on.

According to the described modality, the use of the calculation of the DIFFUSION is to determine if there is an intermittent signal such as the vocal frequency present in the room. Since the RECMIN and RECMAX values are relative to the background level noise level, both will be zero if a voice frequency is not present. Even in very noisy environments, the RECMAX value is also an indicator that a vocal frequency is present, however, such a vocal frequency is less likely to be a single source in such an environment. In executing the voting algorithm, the microphone selection process as shown generally in step 507 of FIGURE 5, decides whether to reconfigure any microphones that are configured, ON to OFF, or reconfigure any microphones from OFF to ON. As indicated at the beginning here, this voting procedure never turns all the microphones OFF. In response to the DIFFUSION, RECMAX and RECMIN values, the microphone selection process selects the microphone that best captures the voice frequency signal. In the selection of this microphone, the energy values of the tracked signal for all the microphones are compared with each other. More specifically, pairs of microphones are examined to determine the direction of origin of the vocal frequency, looking for a pair of microphones where the vocal frequency is strong in the forward-facing microphone, that is, the microphone directed towards the source of the microphone. vocal frequency, and weak in the rear-facing microphone, that is, the microphone directed away from the source of the vocal frequency. Since the null of each microphone is narrower than the main beam, this null is more sensitive and therefore a better indicator of the direction of origin of the vocal frequency than the main beam. In this way, the energy combination of the signal of the two microphones provides a simplified but completely adequate measure to determine the direction of origin of the vocal frequency. In co-pending US Patent Application Serial No. 08/239771, numerous comparisons are made in the execution of a microphone selection process that identifies the microphone or microphones that are directed towards the source of the speech frequency signal. Due to the numerous comparisons made for each possible condition of the CAM 100 circuit, there is a significant amount of redundant calculations. For example, as described in patent application Serial No. 08/239771, for the typical case when a microphone or beam is currently selected to be ON, and the DIFFUSION is large, the process is continuously recycled through the processing steps described here to determine if 1) the same microphone should continue ON, 2) an opposite microphone should be selected in its place, or 3) the same microphone and the opposite microphone should remain both ON. If none of the three tests prove to be satisfactory, for the worst case condition, the process then verifies each entry and chooses the first entry that exceeds a minimum threshold amount that indicates the presence of a level voice frequency signal. low. In other circumstances, choose to remain with the currently selected microphone. Although the worst case scenario can not happen very often, it is the one that requires the greatest processing time. Often in programmatic systems, a limited amount of processing time can be guaranteed on a periodic basis, although larger quantities are hardly available on request. It is more typical to require an amount to be made a sum of the processing amount in a localized amount of processing time. And acceptable results are obtained in the microphone selection process that limit the analysis that provides the best performance in the worst case at the expense of a certain deterioration of the possible operation in the best of cases. To make the selection process of the microphone executed by the voting algorithm relatively constant in terms of processing requirements, the combination values, indicative of a measure of the "goodness" or figure of merit of each microphone, are calculated through of a five-step process described here later. Through the use of those combination values, a comparison can be made with a single number advantageously, instead of making a series of calculations and comparisons. A zero reference combination value is selected to be an indicator of a measure of the best acoustic condition, and any other value other than zero is a measure of how far the measured combination value of the best acoustic condition is removed. If, by way of example, the value of the energy of the tracked signal rectifies for each of the microphones 220-1 to 220-5, shown illustratively in FIGURE 2, it appears as: 1) Microphone 220-1 220-2 220-3 220-4 220-5 rectn 1 2 3 4 5 where, in this example, RECMAX = 5, RECMIN = 1, and DIFFUSION = 4 RECMAX is the energy value of the maximum tracked signal that occurs in a microphone that exceeds the value of the energy of the tracked signal that occurs in any of the other microphones. The RECMIN is the value of the energy of the minimum tracked signal that occurs in a microphone that is smaller than the value of the energy of the tracked signal that occurs in any of the other microphones. And the DIFFUSION is the difference between RECMAX and RECMIN. The first step, as illustrated in step 701 of FIGURE 7, in determining the value of combinations to determine the difference between the value of rectn for each microphone and RECMAX: 2) Microphone 220-1 220-2 220-3 220-4 220-5 RECMAX-reCtn 4 3 2 1 0 The next step, shown as step 702 in FIGURE 7, is to identify and associate the opposite microphone (Mic-op) having the smallest or smallest rectified signal energy value. In the described embodiment employing a pentagon-shaped housing, the opposite microphone can be either the second or third microphone, after a microphone (ie, that the 220-3 has the opposite microphones 220-5 and 220- 1) . Due to the energy values of the tracked signal assigned illustratively, the microphones are paired like this: 3) Microphone 220-1 220-2 220-3 220-4 220-5 Mic-op 220-3 220-4 220-1 220-1 220-2 Once the opposite microphone for each microphone is identified, the difference between each value of the opposite microphone rectn and RECMIN is calculated and this value is then displayed under its associated microphone as follows: 4) Microphone 220-1 220-2 220-3 220-4 220-5 rectn -RECMIN 2 3 0 0 1 Finally, the combination value for each pair of microphones is calculated from the sum of the values of the microphones obtained in steps 2 and 4 for the pair of microphones as follows: ) Microphone 220-1 220-2 220-3 220-4 220-5 Value of 6 6 2 1 1 combination The results, like those provided in this illustrative example, indicate that any microphone 220-4 or microphone 220-5 could be good choices and either or both are selected by the process, since they have values of combination of only 1 from the ideal value of zero. Such results, such as those obtained in this example, are unexpected since the initial values of rectn * - the energy value of the signal traced respectively for each microphone, are chosen simply as a sequence of integers to make clear and easy the understanding. Finer delineations can be obtained in rectn and be easily employed to a simpler microprocessor and such variations were anticipated. When such delineations are used, the ideal case occurs or one that fits best between the source of the vocal frequency and a microphone, which is very common in practice. Various other modifications of this invention were also contemplated and may obviously be frequented by those skilled in the art without departing from the spirit and scope of the invention as hereinafter defined by the appended claims. It is noted that in relation to this date, the best known method for carrying out the aforementioned invention is that which is clear from the present description of the invention. Having described the invention as above, property is claimed as contained in the following:

Claims (32)

1. A method for selecting a microphone of a plurality of microphones for transmitting voice frequency signals from the microphone to an output line, the method is characterized in that it comprises the steps of: measuring the level of the vocal frequency signals that appear in each of the plurality of microphones that respond to vocal frequency sounds, the level of the vocal frequency signals in each of the plurality of microphones is determined by a source address for the source of the vocal frequency sounds; arranging the plurality of microphones in pairs of microphones, a first microphone in each pair of microphones has a master beam sensitivity pattern that extends in a first direction and a second microphone in each pair of microphones has a null pattern that extends generally in the first direction; combining the level of signals of vocal frequency signals that appear in the first and second microphones in each of the pairs of microphones to obtain the combination values of the pair of microphones; Y compare each combination value of the pair of microphones to identify a pair of microphones in which the first microphone receives the best vocal frequency sounds.

2. The method according to claim 1, characterized in that it also includes the step of mounting the plurality of microphones on the perimeter of a circularly shaped housing to orientate it outwards from the center of the housing and form the patterns of supercardiode response.

3. The method according to claim 2, characterized in that the circular shaped housing has a pentagon shape.

4. The method according to claim 2, characterized in that the plurality of microphones are gradient microphones of the first order.

5. The method according to claim 4, characterized in that the plurality of microphones are arranged in the housing by the mounting step to provide coverage of the area for the vocal frequency sounds emanating in a room.

6. The method according to claim 1, characterized in that the comparison step further includes the step of comparing each combination value of the pair of microphones with a reference combination value, the reference combination value provides a measure of a better condition acoustic for the first microphone to receive the vocal frequency sounds.

7. The method according to claim 6, characterized in that the comparison step further includes the step of selecting a combination value of the pair of microphones that is close to the reference combination value.

8. The method according to claim 1, characterized in that the voice frequency signals are the measured energy values of the tracked signal.

9. The method according to claim 8, characterized in that the measurement step further includes the step of determining a difference between the value of the energy of the signal tracked in each of the plurality of microphones and an energy value of the tracked signal maximum in any one of the plurality of microphones.

10. The method according to claim 9, characterized in that the step of arranging further includes the step of identifying the second microphone of each pair of microphones, the second microphone being one of at least two microphones having master beam sensitivity patterns that are they extend in a second direction which is generally opposite to the first direction and the second microphone is one of at least two microphones that have a smaller tracked signal energy value.

11. The method according to claim 10, characterized in that the determination step, which responds to the identification step, determines a difference between the value of the energy of the signal traced in the second microphone and an energy value of the minimum tracked signal in any of the plurality of microphones.

12. The method according to claim 11, characterized in that the comparison step further includes the step of comparing each combination value of the pair of microphones with a reference combination value, the reference combination value provides a measure of a better condition acoustic so that the first microphone receives the vocal frequency sounds.

13. The method according to claim 12, characterized in that the comparison step further includes the step of selecting a combination value of the pair of microphones that is close to the reference combination value.

14. The method according to claim 13, characterized in that it also includes the step of connecting to the output line the first microphone in the pair of microphones with the combination value of the pair of microphones that is closest to the reference combination value.

15. A voice-operated switching apparatus for selecting a microphone of a plurality of microphones for transmitting voice frequency signals from the microphone to an output line, the apparatus is characterized in that it comprises: means for measuring the level of the voice frequency signals that appear in each of the plurality of microphones that respond to the vocal frequency sounds, the level of the voice frequency signals in each of the plurality of microphones is determined by a source address of the source of the vocal frequency sounds; means for arranging the plurality of microphones in pairs of microphones, the first microphone in each pair of microphones has a master beam sensitivity pattern that extends in a first direction and a second microphone in each pair of microphones has a null pattern that is usually extends in the first direction; means for combining the level of speech frequency signals that appear in the first and second microphones in each of the pair of microphones to obtain the combination values of the pair of microphones; Y means for comparing each combination value of the pair of microphones to identify a pair of microphones in which the first microphone best receives the vocal frequency sounds.

16. The voice-operated switching apparatus according to claim 15, characterized in that it further includes means for mounting the plurality of microphones on the perimeter of a circularly shaped housing to orient them outwardly from the center of the housing and form the supercardioid response patterns .

17. The voice-operated switching apparatus according to claim 16, characterized in that the circular-shaped housing has a pentagon shape.

18. The voice-operated switching apparatus according to claim 16, characterized in that the plurality of microphones are gradient microphones of the first order.

19. The voice-operated switching apparatus according to claim 18, characterized in that the plurality of microphones are arranged in the circular housing through mounting means to provide coverage of the area of the vocal frequency sounds emanating from a room .

20. The voice-operated switching apparatus according to claim 15, characterized in that the means of comparison further includes means for comparing each combination value of the pair of microphones with a reference combination value, the reference combination value provides a measure of a better acoustic condition for the first microphone to receive the vocal frequency sounds.

21. The voice-operated switching apparatus according to claim 20, characterized in that the comparison means further include means for selecting a combination value of the pair of microphones that is closest to the reference combination value.

22. The voice-operated switching apparatus according to claim 15, characterized in that the voice frequency signals are the measured energy values of the tracked signal.

23. The voice-operated switching apparatus according to claim 22, characterized in that the measuring means further includes means for determining a difference between the value of the energy of the signal traced in each of the plurality of microphones and a value of maximum tracked signal energy in any of the plurality of microphones.

24. The voice-operated switching apparatus according to claim 23, characterized in that the arrangement means further include means for identifying the second microphone for each pair of microphones, the second microphone being one of at least two microphones having patterns of main beam sensitivity extending in a second direction which is generally opposite to the first direction and the second microphone is one of at least two microphones having a smaller tracked signal energy value.

25. The voice-operated switching apparatus according to claim 24, characterized in that the determining means, which respond to the identification means, determine a difference between the value of the energy of the signal traced in the second microphone and a value of the minimum tracked signal energy in any of the plurality of microphones.

26. The voice-operated switching apparatus according to claim 25, characterized in that the comparison means further include means for comparing each combination value of the pair of microphones with a reference combination value, the reference combination value provides a measure of a better acoustic condition so that the first microphone receives the vocal frequency sounds.

27. The voice-operated switching apparatus according to claim 26, characterized in that the comparison means further include means for selecting a combination value of the pair of microphones that is closest to the reference combination value.

28. The voice-operated switching apparatus according to claim 27, characterized in that it further includes means for connecting to the output line the first microphone in the pair of selected microphones.

29. A voice-activated switching system, characterized in that it comprises: a plurality of circuits for receiving speech frequency sounds and for sharing the vowel frequency sounds in speech frequency signals; means for measuring a level of the speech frequency signals that appear in each of the plurality of circuits, the level of the speech frequency signals in each of the plurality of circuits is determined by a source address of the source of the vocal frequency sounds; means for arranging the plurality of circuits in pairs of circuits, a first circuit in each pair of circuits includes means for optimally detecting signals originating in a first direction and a second circuit in each pair of circuits includes means for detecting Optimally the signals originating in a second direction, the second direction is generally opposite to the first direction. means for combining the level of the voice frequency signals appearing in the first and second circuits in each of the pair of circuits to obtain the combination values of the circuit pairs; Y means for comparing each combination value of the pair of circuits to identify a pair of circuits in which the first circuit receives better the vocal frequency sounds.

30. The voice-operated switching system according to claim 29, characterized in that the comparison means further include means for comparing each combination value of the circuit pair with a reference combination value, the reference combination value provides a measure of a better acoustic condition for the first circuit to receive the vocal frequency sounds.

31. The voice-activated switching system in accordance with the claim 30, characterized in that the comparison means further include means for selecting a combination value of the circuit pair that is closest to the reference combination value.

32. The voice-operated switching system according to claim 31, characterized in that it also includes means, responsive to the selection means, for connecting the first circuit in the selected circuit pair to an output line.