Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04H—BROADCAST COMMUNICATION
  • H04H60/00—Arrangements for broadcast applications with a direct linking to broadcast information or broadcast space-time; Broadcast-related systems
  • H04H60/35—Arrangements for identifying or recognising characteristics with a direct linkage to broadcast information or to broadcast space-time, e.g. for identifying broadcast stations or for identifying users
  • H04H60/45—Arrangements for identifying or recognising characteristics with a direct linkage to broadcast information or to broadcast space-time, e.g. for identifying broadcast stations or for identifying users for identifying users
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04H—BROADCAST COMMUNICATION
  • H04H60/00—Arrangements for broadcast applications with a direct linking to broadcast information or broadcast space-time; Broadcast-related systems
  • H04H60/35—Arrangements for identifying or recognising characteristics with a direct linkage to broadcast information or to broadcast space-time, e.g. for identifying broadcast stations or for identifying users
  • H04H60/38—Arrangements for identifying or recognising characteristics with a direct linkage to broadcast information or to broadcast space-time, e.g. for identifying broadcast stations or for identifying users for identifying broadcast time or space
  • H04H60/40—Arrangements for identifying or recognising characteristics with a direct linkage to broadcast information or to broadcast space-time, e.g. for identifying broadcast stations or for identifying users for identifying broadcast time or space for identifying broadcast time
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04H—BROADCAST COMMUNICATION
  • H04H60/00—Arrangements for broadcast applications with a direct linking to broadcast information or broadcast space-time; Broadcast-related systems
  • H04H60/35—Arrangements for identifying or recognising characteristics with a direct linkage to broadcast information or to broadcast space-time, e.g. for identifying broadcast stations or for identifying users
  • H04H60/38—Arrangements for identifying or recognising characteristics with a direct linkage to broadcast information or to broadcast space-time, e.g. for identifying broadcast stations or for identifying users for identifying broadcast time or space
  • H04H60/41—Arrangements for identifying or recognising characteristics with a direct linkage to broadcast information or to broadcast space-time, e.g. for identifying broadcast stations or for identifying users for identifying broadcast time or space for identifying broadcast space, i.e. broadcast channels, broadcast stations or broadcast areas
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04H—BROADCAST COMMUNICATION
  • H04H60/00—Arrangements for broadcast applications with a direct linking to broadcast information or broadcast space-time; Broadcast-related systems
  • H04H60/56—Arrangements characterised by components specially adapted for monitoring, identification or recognition covered by groups H04H60/29-H04H60/54
  • H04H60/58—Arrangements characterised by components specially adapted for monitoring, identification or recognition covered by groups H04H60/29-H04H60/54 of audio

Definitions

• the invention relates to a system for logging the viewing and/or listening behaviour of a number of persons with respect to television and/or radio transmissions, while one or more television sets and/or one or more radio sets can be present within a specific environment.
• a system which has been known for a long time makes use of a measuring circuit which within the radio or TV set in question is connected to the tuner or the oscillator, or to some other point of the tuner application.
• the measuring circuit establishes which channel has been switched on or what frequency has been tuned to. Up to the mid 80s this worked reasonably well. Then, however, a technological development took off both in the TV world and in radio, which demoted such a measuring method to an entirely inadequate measuring system.
• the changes that took place in the early 80s includes inter alia: a) More than one TV set per location
• VCR video recorders
• VCRs The video recorders (VCRs) arrived. If a VCR is put in front of the tuner of the TV set and the tuner is tuned to the TV set, the TV set being tuned, for example, to a fixed channel (for example channel 00 or channel 36), it will not be possible, if a transmitter is selected via the VCR, to measure this from the TV set.
• a cable operator receives the TV and radio stations via one or more aerial arrangements owned by the operator. The cable operators convert these transmissions to their own channel format, depending on their perception and options within the cable network. The operator provides information to the user regarding the format, for example which transmitter has been put on which cable channel.
• a satellite converter/tuner is connected to a TV set in the same way as a VCR.
• the solution which has to be sought for the VCR is analogous to the system with the satellite receivers.
• the invention then provides a system for logging the viewing and/or listening behaviour of a number of persons with respect to television and/or radio transmissions, where one or more television sets and/or one or more radio sets can be present within a specific environment, characterized - in that in the vicinity of each television or radio set an observing unit is installed at least provided with a microphone by which the sound produced by the set in question is recorded and a data transmitter via which the sound or a signal related thereto can be trans- mitted, in that within or in the vicinity of the said environment a base unit is installed, at least provided with a tuner via which it is possible to tune to any station which can also be received via the said sets, - a data receiver via which the signals of the data transmitters can be received, a processor which carries out the following steps: a) tuning the data receiver to the data transmitter of a selected observing unit, b) sequentially tuning the tuner to the stations to be received and each time comparing the sound signal received or a signal related thereto with the signal obtained via the data transmitter
• transponder which can be activated by an interrogation signal to emit a coded response signal
• a transceiver in that in the vicinity of each television or radio set a transceiver is installed at least provided with an interrogator via which an interrogation signal is emitted periodically and a response receiver via which the coded response signals can be received and stored at least temporarily, means via which the stored response signals or signals related thereto can be transmitted to the observing unit which is associated with the set in question and where the response signals can be combined with the signal to be fed to the data transmitter.
• Linking the transponder technology to the above-described system ensures that viewing and/or listening data of a set will not be measured and logged unless a test person with a transponder is actually present in the room or locality in question where the set is located.
• Figure 1 shows the apparatus which has to be installed in a home
• Figure 2 shows a central station which communicates with the apparatus in a number of homes.
• the system can roughly be divided into four system components: 1 Central station
• FIG. 1 schematically shows a home 10 which is provided with a number of rooms 12, 14 and 16.
• a receiver set 18a Located in room 12 is a receiver set 18a, and located in the room 14 is a receiver set 18b.
• Each receiver set 18a, 18b can be formed by a radio set or television set or a combination of the two.
• the objective is to measure the viewing and listening behaviour with respect to both sets 18a and 18b. Consequently, an observing unit 30a is installed in the vicinity of set 18a, while an observing unit 30b is installed in the vicinity of set 18b.
• the observing unit 30a is provided with a microphone 20 which is connected to an amplifier 22.
• the signal output by the amplifier 22 is subjected to processing in the processing circuit 24, for example in order to filter out a number of nonrelevant frequency ranges, to remove background noise and the like.
• the processed signal is then, via a transmitter 26, transmitted to the base unit 40, yet to be described, which is located in the room 16.
• the observing unit 30b is likewise, in the same manner, provided with a microphone 20, an amplifier 22, a processing circuit 24, a transmitter 26 and a control circuit 28.
• the base unit 40 located in the room 16 is provided with a receiver 32 via which the signals from the transmitters 26 can be received.
• the base unit is further provided with a tuning unit 34 by means of which it is possible to tune to all the stations which can also be received via the receiver sets 18a and 18b.
• the sound signals of the received programmes are fed to a processing unit 36 which functions in the same way as the processing units 24.
• a sound comparator 38 the signals coming from the processing unit 36 and coming from the receiver 32 are then com- pared with one another.
• the frequency to which the tuning unit 34 has been tuned (or for example the corresponding channel number) is stored in the control unit 42, as is the time at which it was established that the signals from the processing unit 36 and from the receiver 32 were identical. It is also recorded from which observing unit the signals received via the receiver 32 originated, and which observers (provided with a transponder) were present in the room in question.
• Figure 1 shows the base station in a separate room 16, there is reason whatsoever why the base station should not be put into one of the other rooms or possi- bly outside the environment 10, as long as the transmitters 26 are able to communicate with the receiver 32.
• the idea therefore is for the sound comparator 38 to compare the sound produced by the radio or TV with the sound produced by the tuning unit 34.
• the sound comparator 38 There is a complication in that no hardware link must exist between the radio and/or TV receiver and the system. For the produced sound to be recognized, the sound will have to be received by a microphone or some other device suitable for this purpose and be converted into an electrical signal.
• This signal could be compared directly with the signal obtained from the tuning unit 34.
• the comparison procedure can be considerably simplified, however, if the two signals undergo processing in processor 24 and processor 36, respectively.
• the two processors 24 and 36 will process the electrical signal in such a way that a frequency range from 1000 to 6500 Hz remains. Using active filters, various samples (between 4 and 20 in number) of variable frequency width will then be taken from this band. The processed signals thus obtained can then be compared with one another in various ways.
• the reception in the base station 40 is free from noise or interference
• the observing unit makes use of a microphone which will also intercept "background noise”.
• the background noise can be eliminated by looking only at those sections of the total signal which emerge above the level of the background noise.
• those amplitude peaks of the processed signals which emerge above a set value of the noise are counted, their location on the time axis is compared, and the average energy is compared. Since the emitted signal is still intelligible or perceptible in relation to a certain degree of background noise, the level of the noise can predefined (the said set value). Consequently, the "background noise" component can be eliminated in the equation, and this source of interference is not involved in the equation.
• the signals are compared only if the signal of the observing unit 30 is sufficiently high with respect to the background noise. Then, again, only the peaks of the signal will therefore be used for comparison.
• the sound comparator 38 use can be made, for example, of: a- Fourier analysis b- Adaptive delta modulation (edge direction comparison) c- Trending (tracing the entire peak contour)
• the comparison can be carried out for a number of samples, identity being estab- lished if two conditions are met.
• the first condition is:
• SS - slave sample sample from observing unit 30
• inpv input tolerance value
• # samples not equal to (0+inpv) ⁇ x
• x limit value set for the number of samples.
• the tuning unit in the base station will be tuned to the next station for the purpose of the next comparison cycle. This will continue right until the sound comparison meets the predefined standard.
• the preset stations i.e. the stations the tuning unit must tune to successfully, are automatically set during installation.
• the tuning unit runs through all good-reception signals and will store the corresponding frequency information in a preset memory.
• Establishing the station to which a set has been tuned makes sense only if at least one person is present in the room or locality where the set is located, said person then being assumed to be listening to the radio or to view the TV transmission.
• the system described here makes use of trans- ponders.
• Each person taking part in the study carries a transponder on his or her person. This transponder is activated by a modulated activation signal transmitted continuously by the observing units. Upon activation, the transponder emits an identification code which depends on the modulation and can be received by the observing unit.
• the observing units are each provided with an IR transmitter, via which activation signals are emitted in a predefined rhythm.
• Each transponder 50 is provided with an IR receiver 46 via which these activation signals can be received if the transponder is located in the same locality as the transmitter.
• Each transponder is further provided with a code transmitter 48 by means of which an identification code unique to each transponder can be emitted. These identification codes are then received by the code receiver 52 with which each observing unit 30 is fitted.
• the code transmitter 48 of the transponder emits radio frequency signals which do pass through the walls and in principle can be received by all code receivers 52 of all the observing units.
• each IR transmitter of each observing unit transmits a specific code along with the activation signal. This code is detected in the transponder and accordingly a response signal, coded with a specific code, is emitted. In the observing unit, the code in the received response signal is compared with its own (IR) code. Only that observing unit for which the two codes are related to one another will establish that the transponder is located in the same room.
• transponder As soon as a transponder therefore enters a locality, it will respond to the activation signal of the observing unit and transmit its identification code to the observing unit, which stores said code, whereupon the transponder is regarded as "present". Once the transponder no longer responds after three activations, the code is erased from the memory and it is assumed that the transponder is "absent".
• the observing unit which is certainly present in said locality will not communicate with the base station, with the exception of a "null message" which indicates to the base station that the observing unit in question has been activated and is functioning correctly. If the presence of a transponder is detected, however, it is reported to the base station together with the (processed) sound signal to be compared, so that it is known in the base station, once the comparison process is complete, which station the set in question has been tuned to and which person (transponder) is present in the same locality as the set.
• FIG. 2 schematically shows a number of houses lOx, lOy and lOz, each provided, in the above-described manner, with a number of observing units and a base station.
• the base station 30x, 30y and 30z is shown schematically in each house.
• Each base station is able to communicate with a schematically depicted central station 60. This can be achieved by collecting the base station 30 in the location to the telephone network, options for this purpose including not only two-wire, four-wire and ISDN links but also wire-free links, for example GSM.
• the selection device in the base station 30 can be adapted as required.
• a data transmitter (transceiver) 54 is present which can communicate with the data receiver (transceiver) 56 in the central station 60. Via these devices, all the data stored in the base stations can be transmitted to the central station.
• a practical specific embodiment of the system may further include details other than those discussed above.
• the base station can have an aerial system for: FM, MW, LW, SW and all TV frequencies.
• an aerial system for: FM, MW, LW, SW and all TV frequencies.
• the installed load required is such that the base station can be supplied by a battery supply as well as normal mains voltages.
• the base station can also be provided with an input device to give the participants in the study the additional option of recording their information digitally instead of using an analogue method (via a form).
• This option is provided in order to establish whether the person taking part in the survey has actually been able to take note of the transmission in connection with the questionnaire to be filled in subsequently, in relation with his rating of the programmes and commercials (active metering).
• the objective is for the stored data in the base station, such as appliance address, name/address/place of residence, and date and time of the "presence of the transponder wearer" is supplemented with the information of what station has been listened to or viewed.
• the base station can be provided with two tuners by means of which the available stations can be scanned.
• the reception sensitivity can be adjusted, so that stations can be selected in terms of the section quality.
• Tuner 1 is designed to receive the TV signal.
• Tuner 2 is designed to receive the radio signal.
• the tuners include an arrangement which enables them to search the frequency bands for radio and TV stations automatically and also to store them in a "preset memory", thus establishing on what frequency these stations are received.
• tuners are additionally fitted with a device which provides these presets with a station name.
• the sound information of the radio or TV set in the home will be received by a directional microphone or broad-band microphone. After processing in the comparator device, this signal will be compared (real time research) with the sound signals of the preset tunings presented sequentially. These signals too are processed in the same way as the microphone signal. These two signals are compared, and if they match within a predefined degree of identity, this signal (name) associated with the preset listing will be reported as the station that is being listened to or viewed.
• the preset comparison will take place. In the application, this option will usually be implemented as follows. The searching for the preset and comparison in the base station will not take place until a transponder is within hearing distance. Not until the transponder comes within "hearing distance of the observing unit" will this be noted down in the database together with all other (transponder) data. When the transponder moves back outside "hearing distance", a second note will be made. This allows the "listening time of the participant" to be calculated.
• the base station can, for example, be fitted with a device which listens for 8 transponders at 8 different frequencies. At intervals, (e.g. every 10 sec.) the transponder emits a brief "burst".
• Signal processing in the base station can take place in such a way that an elec- tronic device subdivides the sound spectrum from 1 to 6.5 kHz into frequency blocks of variable frequency lengths.
• the blocks of the two signals preset station tuning in base station, and microphone signal from the observing unit
• the blocks of the two signals will then be compared in real time.
• the walls inner and outer walls of normal houses
• the walls do not carry magnetic insulation. This means that if the transponders emit bursts continuously or at intervals, these will be received by all observing units within the range of the transponders. The observing units will then indicate that the transponder is in a number of locations at the same time. Of course, this is not the intention and is therefore of no use.
• the PABs will therefore have to emit a presence signal not related to magnetic waves or radio waves. This option of active reporting from the transponders is not the best solution either, given a limited a battery capacity.
• the transponders will therefore be defined as a passive component within the system.
• the observing units will give off, within the locality, a sufficiently strong signal which will be just beyond the frequency of visible light (on the IR side of the spec- trum).
• the observing unit will supplement the signal with its code, which is uniquely linked with its unique address.
• the transponders now know which code they must add to their periodic burst so as to emit exclusively and solely a coded signal which is recognized exclusively and solely by the observing unit in question.
• the other observing units which may also "hear" this coded signal will now not respond.
• the functionality of the transponders can be summarized as follows: a- detecting the unique IR code linked to the observing unit b- generating a unique transmission carrier wave frequency (batch ID) c- storing the observing unit recognition information (observing unit location modulation mode) d- modulating the carrier wave of the observing-unit information (location indication) e- formatting the modulation into a pulse/width signal (manually input information and transponder condition) f- the observing unit, after notification of somebody's presence, regularly pages the transponders present in accordance with an interval schedule. For example, the observing unit can page the transponders three times. If no further answer is received, this establishes that the transponder in question has left the room. g- physical recognition of the person taking part in the survey.
• the transponder can be designed as a button model and in various colours or shapes or with a number to enable simple visual identification and assignment to the same person taking part in the survey during the survey period.
• the [lacuna] the transponder identification code is equal to the allocated frequency of the transponder. This is laid down in a table.
• the unique observing-unit address code included by the observing unit in its tranponder trigger message is modulated by the transponder onto its transmission frequency.
• the transponder(s) will therefore report this code in modulated form to the observing unit. Only the observing unit having the correct address will therefore report this transponder as "present”. It is therefore entirely possible for other observing units also to receive this signal, but as the modulated address information does not match the observing unit, the latter will then not report a "presence”.
• the transponder is preferably composed of low-energy components and is in a sleep mode until the coded trigger pulse comprising the unique observing-unit address in question has been received.
• a lithium voltage source or another environmentally friendly voltage source will be chosen.
• the battery function is assumed or supported by an IR or photosensitive cell which converts this energy into an electric voltage.
• Specific other messages such as a "battery low” indication, can be transmitted at the same time in the ID message, so that the researcher, when reading out the data at the decoding centre, can also be informed of the battery status, or that the battery can be replaced by service personnel during a visit, etc.
• "real time” emotion information can be obtained by using the "switches” present on the button to give an indication of impressions or how the information listened to or observed is rated. This information is made known to the observing unit by the modulated signal being provided with a pulse/width coding.
• the functionality of the observing units can be summarized as follows: a- receiving sound (per microphone or direct link) b- conditioning the sound to give a predefined format c- activating the transponders plus providing location information of the observing unit d- establishing which transponders are transmitting (carrier wave selection) e- establishing what additional data are being reported by the transponders
• the functionality of the base stations can be summarized as follows: a- synchronizing the observing units b- sampling the conditioned sound presented and comparing the samples with the samples of the station's own tuners (information regarding TV or radio origin), c- optionally selecting the tuners (stable setting for radio reception) d- digitizing (all) the data e- receiving, listing and presetting all stations for the benefit of the master tuners (3 TV and 4 radio) (self-learning principle) f- building up a database file g- input of survey data, e.g. filling in the form via keyboard h- transmitting the data to a central station via GSM, for example i- reading in the favourite presets for each person taking part in the survey, so that identity can be established considerably more rapidly.
• the functionality of the central station can be summarized as follows:
• the central station consists of a PC, system and application software.
• the functionality is that of processing built-in data files to give useful application information for the benefit of the market research companies and their clients.
• the functionality of the CPU further comprises the retrieval, collection and recording of the data from the outstanding HBs in an event-driven, object-oriented open database, as well as transmitting messages to the outstanding HBs and attending to and checking the plurality of desired settings for the HBs.
• any custom metering application can then be connected.
• Echo-suppressing circuits are especially applied in the telephones comprising a loudspeaker and a microphone (handsfree calling, carkits, etc.).
• loudspeaker and microphone there is always the risk that the generated sound is received by the microphone which will be the cause of echo, reverberation and regen- eration phenomena.
• the incoming and outgoing signals are acoustically separated in the telephone receiver so that no or only a very small signal from the loudspeaker receives the microphone.
• the audio separation is hardly realizable.
• a part of the loudspeaker signal can be added with negative sign to the microphone signal.
• the source of the unknown signal is in principle the microphone, it can be a direct coupled signal from a TV or amplifier, but the problem around a microphone as a source is larger and builds therefore a worst case situation for the principle of signal recognition.
• the source of the reference signal is a radio receiver (tuner).
• Echo-suppressing circuits are applied in "handsfree" kits and telephones and can be subdivided into two main systems.
• the most simple version uses the signal received from the loudspeaker to attenuate the volume of the reproduced sound. Therewith the voice of the speaker is only hardly reproduced by the own loudspeaker, but in essence a simplex communication system is created. The two interconnected speakers are able to talk each in its turn, however, speaking simultaneously is not possible.
• This version of an echo-suppressing circuit does not apply signal recognition and is within the scope of this document further not relevant.
• the second method uses a controllable transversal filter of which the characteris- tics are adapted such that the acoustical characteristics of the signal path from loudspeaker to microphone are imitated as good as possible.
• the imitated signal is thereafter inverted and added to the microphone signal. If the imitated signal sufficiently resembles the real incoming signal then the summation will add to the effective suppression of the echo signal.
• the correcting signal ref (n) should resemble as good as possible the signal which is received through the microphone.
• This acoustic signal, received by the microphone is filtered by the acoustic characteristics of the loudspeaker, the housing, the space in which the apparatus is present and the characteristics of the microphone.
• the filter which should generate ref (n) has to be able to imitate all these acoustic properties.
• the amplitude behaviour, the frequency behaviour as well as the running time/phase behaviour should correspond as much as possible with that of the acoustical system of loudspeaker, housing, space, and microphone.
• n+1 summations with n+1 weighing factors or filter coefficients c(0..n) are necessary, see figure 4.
• the maximum delay in that case is equal to n times the delay of one delay element.
• the problem is that the coefficients of the filter have to be determined. In the literature algorithms are published for that purpose, each algorithm having its own specific advantages and disadvantages.
• any output signal which is present after the correcting circuit can be presented as a consequence of the imperfection of the filter.
• the circuit After switching on the echo-suppressing circuit, the circuit has to determine the desired characteristic of the filter.
• the speed with which the filter can be adapted to its optimum is the reaction speed of the echo suppresser.
• the error signal comprises the real error signal, being the differ- ence between the ref (n) after the filter and the real ref (n) present in mic(n), and the noise x(n).
• a microcontroller is nothing more or less than a computer, embodied as one sin- gle integrated circuit.
• Microcontrollers are destined for control functions and are in general less suited for functions where intensive calculations are involved. Calculations such as a Fourier- analysis can be done by a fast microcontroller but these calculations do not leave any space for other functions as the determination of correlation between spectra of different signals.
• the speed of microcontrollers is between 1 and 20 MIPS (Mega Instructions Per Second). Because of its characteristics a microcontroller is hardly suited for echo-suppressing calculations but is on the other hand suitable for taking decisions based on the results of the echo-suppressing circuit.
• DSP Digital Signal Processor
• microcontroller The configuration and the global functioning of a Digital Signal Processor (DSP) and a microcontroller are basically the same.
• the DSP's have such an internal structure that operations on data and numerical calculations with high speed can be performed.
• Microcontrollers are more suited for controlling external apparatuses and less suited for calculations.
• DSP's are able to perform a standard multiply-raise operation in one in- struction cycle This instruction is very handy in algorithms such as digital filters.
• DSP's have the possibility to perform different memory accesses in one single instruction.
• DSP's have one or more serial or parallel I/O -interfaces and specialized I/O- management mechanisms to enable a very fast input and output of data to be operated.
• the floating point DSP's are available both in fixed point- or floating point-embodiments, the floating point DSP's are in general more expensive than the fixed point-DSP's but have the advantage that they are easy to use in applications with a lot ol ⁇ floating point-calculations.
• the speed of a normal DSP is between 10 to 50 MIPS (Mega Instructions Per Second).
• DSP digital signal processor
• a DSP is pre-eminently suitable for an echo-suppressing system although in most cases circuits have to be added for converting analogue signals into digital signals and vice versa.
• a sound card has a stereo channel it is possible to play a reference signal on the right channel and play the "unknown” or sound to be identified on the left channel.
• Both signals can be produced simultaneously and applied to the measuring circuit.
• COOL EDIT is selected because this program has the necessary capacities for performing operations on sound. Furthermore, this program is available in an evaluation version and the evaluation period is long enough to be able to compose the signals.
• the applied signal operations are:
• the original signals are supplied to the reference input of the echo-suppressor whereas the second signal, eventually after operation, is applied to the microphone input of the echo-suppressor. Thereafter it is determined what suppression of the reference signal in the microphone signal can be obtained and, if possible, within what time this suppression can be obtained.
• the sinusoidal signals are periodical by nature and delays in the signals can be interpreted as phase shifts of less than one period. For that reason the FIR-filter in the echo-suppressing circuit does not have to compensate for delays larger than those of one single period time for these signals.
• the coefficients of the filter for the section with delays larger than one period interval will therefore be adapted to a 0 value. That means that for periodical signals only a restricted number of coefficients have to be calculated and one may expect therefore that the response time of the echo-suppressing circuit will be relatively short.
• Non-periodical reference signals With speech and music the same suppression factors can be obtained but the response times are less favourable, up to more than one second, especially if there are large delays in the microphone signal in relation to the reference signals. This can be explained from the fact that in those cases more coefficients of the filter will not be equal to 0 so that more internal calculation operations are necessary to determine the most optimum coefficients. Signal delays
• the signal delays are for the most part larger than the control range of the FIR-filter and furthermore the amplitudes of the delayed signals were relatively large.
• the signal part can be suppressed in a relatively small degree in relation to the whole signal and a measurement of the suppression is not performable through the usual measuring system.
• the "Long Hallway” and “Parking garage”-effects do not pose a problem, because this type of signals are not hindered by delays longer than one period.
• Tone control The consequences of a tone control are observed by treating a noise signal with filters which give preference to lower or higher frequencies or suppress these frequencies.
• a noise signal is applied because this is the only signal whereby one may assume with certainty that the frequencies to be influenced are represented in the signal.
• a tone control influences the amplitude and the phase of the signal, however, both in a degree within the boundaries of the control range of the FIR filter. On the ground thereof one may expect that the echo suppression is not impeded by a tone control.
• a filter which is adapted to the acoustic and further properties of the signal path may reach excellent results even in case the amplitude of a disturbing signal is in the same order as the amplitude of the hidden and sought after reference signal.
• an echo suppressing signal is capable of recognizing signals it is not directly usable for such a recognition. The reason therefore is that the echo suppressing circuit suppresses the requested signal respectively to signal to be recognized and does not indicate in any way that the signal is found.
• the efficiency of the echo suppressing circuit can be derived from the signals which are directed to the input of the echo suppressing signals as well as the signals which are supplied by the echo suppressing circuit.
• the echo suppressing circuit has not provided any usable contribution, still a method for signal recognition is necessary, the problem is greatly simplified by the echo suppressing circuit.
• the signals mic and mout will not show the mutual differences which may appear between mic and ref.
• the spatial acoustic and the setting of the tone control etc. do not have any influence on the relation of the mic- and mout-signals.
• the filtered signal can be subtracted from the mout-signal, if the signals mic and mout have identical content the resulting difference signal will have the 0-value. In case the ref-signal is suppressed then this signal will appear as difference signal.
• a cross correlator is a circuit which compares two signals by multiplying them. This multiplication can be done by a DSP, a Microcontroller or an analogue (four quadrant) multiplier. If there is some similarity in the multiplied signals then multiply- ing implies for that signal raising to the square, which always results into a positive value. Equality of the signals will therefore result into the generation of a DC voltage which was not present before.
• FIG. 6 illustrates a block diagram for this method. However, if standard echo suppressing circuits are used the output signal of the
• FIR-filter is not available. This signal can be derived from a negative summation of the output signal with the microphone signal. Figure 7 illustrates this principle.
• the echo suppressing circuit processes the signal mic(n) not only by negatively adding the signal ref (n) but in branch ⁇ A ⁇ the phase and amplitude characteristic are
• the added filter may also have a fixed setting.
• the signal ref (n) regained in this manner can thereafter be used to determine the correlation between the reference signal and the microphone signal through the above mentioned method.
• Dependent on the internal structure of the echo signal it may be necessary to preprocess the mic(n) signal just as was done in the regaining circuit of ref (n), using a filter before multiplying the signals ref (n) and mic(n).
• Appearing errors can be compensated partly by adding an offset voltage which is derived from a rectified input signal.
• Phase characteristic Phase characteristic
• phase errors are not predictable because the frequency content and amplitude of signals, appearing in practice, are unknown and variable. It therefore depends how far the erroneously compared frequency components are of interest within the resulting correlation signal.
• a suitable correction of the phase characteristic of the echo suppressing circuit is therefore of great importance for the usability of the correction signal, taking into account the multiplicity of possible signals which in practice may appear. If a signal would comprise frequencies for which the phase errors are large then an effective recognition would not be possible.
• the input signal comprises 50% of the requested signal and for the main part consists of other signals then this ratio of 1 :2 will be found back between the am- plitude of the correlation signal and the amplitude of the input signal.
• the measure of degree results from divisioning the correlation signal by this comparison signal. In case of complete correlation the division will result into a value 1, in all other situations the value is between 0 and 1 and forms a measure of correspondence.
• the signal refl(n) as well as the signal mic(n) are subjected therefore to additional processing.
• the processing carried out on the refl(n) signal can be considered, however, also as a processing from the FIR-filter and can therefore remain out of consideration.
• a(n) is a function representing the phase and amplitude processing of the signal path mic ⁇ mout. It is now assumed that the mic(n) signal comprises an unknown signal x(n) and the reference signal ref(n) has undergone a processing caused by the acoustics of the room k(n). The value of k(n) may be 0, in which case the reference signal is not represented in the mic(n) signal.
• ref2(n) b(n) * ⁇ x(n) + k(n) * ref(n) ⁇ - mout(n)
• the summing circuit does not perform any signal operation within the bandwidth of ref2(n) of mic(n), or represents a neglectable factor in re- lation to the operations already performed by the echo suppressor.
• the second part of the equation comprises the factor ref(n) 2 which results into a positive value independent on the signal shape of ref(n).
• the section b(n) 2 * k(n) 2 * ref(n) 2 therefore results into a DC voltage
• the section 1 b(n) 2 * k(n) * ref(n) * x(n) results into an AC voltage which can be suppressed by the low-pass filter.

Landscapes

• Engineering & Computer Science (AREA)
• Signal Processing (AREA)
• Mobile Radio Communication Systems (AREA)
• Circuits Of Receivers In General (AREA)

Applications Claiming Priority (2)

Publications (1)

Family

ID=19767765

Family Applications (1)

Country Status (2)

Citations (5)

• 1998
  • 1998-09-08 NL NL1010024A patent/NL1010024C2/nl not_active IP Right Cessation
• 1999
  • 1999-09-08 WO PCT/NL1999/000557 patent/WO2000014914A1/en not_active Application Discontinuation

Patent Citations (5)

Also Published As

Similar Documents

Legal Events