0TK0
`(11) Patent Number:
`5,579,124
`115
`United States Patent
`Aijala et al.
`[45] Date of Patent:
`Nov. 26, 1996
`
`
`[54] METHOD AND APPARATUS FOR
`ENCODING/DECODING BROADCAST OR
`RECORDED SEGMENTS AND MONITORING
`AUDIENCE EXPOSURE THERETO
`
`4,425,642
`
`1/1984 Moses et al.
`;
`;
`(List continued on next page.)
`FOREIGN PATENT DOCUMENTS
`
`scssossssessesstsssetsne 370/76
`
`[75]
`
`Inventors: Victor A. Aijala, Arnold; Gerald B.
`T1986 Canadaeeresrsnreenrenee HOAN 7/16
`1208761
`Cohen, Gaithersburg; James M.
`Canada...
`. HO4H 9/00
`2036205 12/1991
`
`Jensen, Columbia; Wendell D. Lynch,
`8/1985 France.
`Ossoo
`wyoj2558002
`
`Silver Spring: Juan C. Urbi, Laurel, llWO21/11062,7/1991.WIPOsnnnnnnsnsnn
`of Ma.
`OTHER PUBLICATIONS
`
`[73] Assignee: The Arbitron Company, Columbia
`Ma.
`,
`
`;
`
`;
`[21] Appl. No.: 396,342
`[22]
`Filed:
`Feb. 28, 1995
`
`Related U.S. Application Data
`[63] Continuation of Ser. No. 976,558, Nov. 16, 1992, aban.
`doned.
`
`The Institute of Electrical and Electronics Engineers, Inc.,
`“Spread Spectrum Techniques”, IEEE Press, 1976 (p. 15).
`SMPTE Journal, Society of Motion Picture and Television
`Engineers,Inc., vol..101, No. 8, Aug. 1992 (pp. 538-549).
`Namba, Seiichi, et al., “A Program Identification Code
`Transmission System Using Low-Frequency Audio Sig-
`nals”; NHK Laboratories Note; Ser. No. 314, Mar. 85.
`Primary Examiner—Thai Q.Tran
` 48S#Stant Examiner—Robert Chevalier
`Attorney, Agent, or Firm—Curtis, Morris & Safford, P.C.;
`Eugene L. Flanagan, Il
`
`[56]
`
`.
`References Cited
`U.S. PATENT DOCUMENTS
`
`[51]
`Ite Cho ceeccccssnneseseeseees HOAN 5/76; HO4N 5/00
`
`[52] U.S. Ch. cisccssssesssscssrsnssessecseerseees 386/96; 381/2; 386/98—[57] ABSTRACT
`
`
`[58] Field of Search ou...
`cess 381/14, 2, 3, 4;
`Methods and apparatus for encoding and decoding informa-
`455/2, 49.1, 53, 67; 358/335, 341, 343;
`tion in broadcast or recorded segment signals are described.
`348/1, 2, 3, 5
`In certain embodiments, an audience monitoring system
`encodes identification information in the audio signal por-
`tion of a broadcast or recorded segment using spread spec-
`trum encoding. A personal monitoring device receives an
`acoustically reproduced version of the broadcast or recorded
`
`2,470,240=5/1949 Crosby .....sesscsseccsecersseneateceesensee 250/27 signal via a microphone, decodes the identification infor-
`2,573,279
`10/1951 Scherbatskoy.......
`. 3467—mation from the audio signal portion despite significant
`
`2,630,525
`3/1953 Tomberlin et al.
`...
`~ 250/6
`ambient noise, and stores this information, automatically
`
`2,660,662
`11/1953 Scherbatskoy.......
`” one
`providing a diary for the audience member which is later
`2,766,374 10/1956 Hoffmann........
`” 1979/2
`uploaded to a centralized facility. A separate monitoring
`3,004,104 10/1961 Hembrooke..
`i 346/37
`device decodes additional information from the broadcast
`3,397,402
`8/1968 Schneider.....
`
`3,492,577
`1/1970 Reiter et al.
`..
`sae 325/31
`signal, which is matched with the audience diary informa-
`
`3,760,275
`9/1973 Ohsawaet al.
`ww. 325/31
`tion at the central facility. This monitor may simultaneously
`
`3,803,349
`4/1974 Watanabe.....
`178/5.8 R
`send data to the centralized facility using a dial-up telephone
`3,845,391
`10/1974 Crosby..........
`sees 325/64
`ine, and receive data from the centralized facility through a
`
`4,025,851
`5/1977 Haselwood et al
`--- 325/31
`signal encoded using a spread spectrum technique and
`
`4,225,967
`9/1980 Miwaet al.
`......
`455/68
`modulated with a broadcast signal from a third-party.
`
`4,230,990
`10/1980 Lert, Jr. et al.
`...
`- 455/67
`eeeeeeeeeeeeeee 370/11
`4,238,849
`12/1980 Gassmann 0...
`2/1981 Schmidt et al. oes 381/14
`4,252,995
`
`110 Claims, 10 Drawing Sheets
`
`INFORMATION
`
`| BROADCAST
`“NYTT 4
`
`110
`
`ADDITIONAL
`
`100
`
`AUDIO
`PROGRAM
`
`ENCODED
`AUDIO
`PROGRAM
`
`D
`
`Sony Exhibit 1019
`Sony Exhibit 1019
`Sony v. MZ Audio
`Sony v. MZ Audio
`
`

`

`5,579,124
`
`Page 2
`
`U.S. PATENT DOCUMENTS
`
`4,450,531
`
`5/1984 Kenyonet al. o.ccccsssecsssssseson 364/604
`
`4,547,804 10/1985 Greenberg .
`. 358/142
`4,613,904
`9/1986 Lurie......
`358/142
`
`4,618,995 10/1986 Kemp .
`ene 455/2
`
`4,626,904 12/1986 Lurie......
`.. 358/84
`1/1987 Greenberg oo...ceeeceseeesees 358/142
`4,639,779
`6/1987 Lert, Jr. et al. sessssernrrreeneienes 358/84
`4,677,466
`. 358/84
`4,697,209
`9/1987 Kiewit etal. .
`
`4703476 10/1987 Howard
`370/76
`..
`4,718,106
`1/1988 Weinblatt
`we 455/2
`4,805,020
`2/1989 Greenberg .....
`we 358/147
`4,843,562
`6/1989 Kenyon et al.
`ccccccssscsssesseeenen 364/487
`
`
`
`4,876,617 10/1989 Best et al. occccccccccsssssessesssssessee 360/60
`
`fosters aise von wrrveenernavnnnorecennenggngaanaseans pen
`
`CIMCT on.
`ceescceatsessencsnersensoceeneatens,
`
`7/1990 Kramer wn.cecsscesssesesrseresesee 358/142
`9/1990 Welsh et al.
`cccscssccsccssssessssssesssssen 455/2
`
`....ccscccssssseusscecs-see 358/142
`10/1990 Greenberg
`11/1990 Gross el al.
`cccccsccscsssssssecscsessccsseee 380/3
`6/1991 Call sesssscssessntssenssntsnsesnnescsssee 455/2
`5/1992 Best et al. ccecsscccscsssssssssssssssseeseeee 380/3
`6/1992 Mankovitz
`
`5/1993 Sherman uo...
`.
`6/1994 Preusset al.
`
`.
`>
`4,945,412
`4,955,070
`4,967,273
`4,972,471
`5,023,929
`5,113,437
`5,119,503
`5,213,337
`5,319,735
`
`

`

`U.S. Patent
`
`Nov. 26, 1996
`
`Sheet 1 of 10
`
`5,579,124
`
`WYd0ud
`
`o1any00!
`
`WVEI0Ud
`
`|_-SLI
`
`ISvoavoud|404
`cana
`
`Q3009N3
`
`olanv
`
`NOLLWWHOSNI
`
`“TWNOILIGQV
`
`

`

`U.S. Patent
`
`Q3Q0NA
`
`olany
`
`WVa90ud
`
`Noy. 26, 1996
`
`Sheet 2 of 10
`
`5,579,124
`
`Ve‘Sls LSVvoavoud
`
`
`

`

`U.S. Patent
`
`Nov. 26, 1996
`
`Sheet 3 of 10
`
`5,579,124
`
`JéSla
`
`

`

`eeeaSeeeeeeee10Olrirtroodtitoorottoodtttiohodtlioogdgitao66
`
`0|jiOoo|COTorto1Pioiooit4ottfJa~aT¥Tag“914
`
`eeeeeeeeee
`
` | | | | |
`
`Noy. 26, 1996
`
`AY‘-
`
`—Ab-
`
`
`
`Pirrrd?ttbrtjoo00gdd000O00OFFTEtTttbia
`
`U.S. Patent
`
`
` r260778078008|O00E ze6z|Ol”207THEGHC||UtMOEOSEAnt=|eeeTeeSIEBOEOO,228|0
`
`
`“snot&§@oorOr a
`880777009662962
`
`—=4b
`
`v¢“dla
`
`Sheet 4 of 10
`
`—— AL
`
`SINIOd929
`
`Ae
`
`qb
`
`
`
`

`

`U.S. Patent
`
`|||||
`
`~~a[pbbtrttrttpooooocoooootrrrriridetI|
`
`-——-——mpP||||0||LZL_L111
`
`d¢Sls
`
`Nov. 26, 1996
`
` So=Samyown_oevaDN
`~-+aypOLOOFTOT
`10io|LLLLTTr9¢“SI
`OTITIOOTiOoO1FoTOOT — —\ —
`
`H¢“Sls
`
`—~k-
`
`— —\+ —
`
`eeeee
`
`

`

`U.S. Patent
`
`Nov. 26, 1996
`
`Sheet 6 of 10
`
`5,579,124
`
`ADDITIONAL
`
`x(t)
`185
`
`INFORMATION
`
`FIG. 4B
`
`

`

`Nov. 26, 1996
`
`Sheet 7 of 10
`
`35,579,124
`
`WD0ud
`
`U.S. Patent
`
`
`PTTTTapartsssTT7Lsv3qvoua|!wOd|SLi
`
`olny||G3G09N3|ollwali|WWYDOYNdrn5ned
` 091os!8616606!||Tinga|V-HSINI}WaC0H|:961"6)Leeor
`
`
`WNOLLIGGYChSd>)ALINOVA
`NOILLVIUO-NI()x
`olnymI
`
`()9
`
`Q4ZIWYLN39D
`
`
`
`ONISSIIONdVIVE
`
`
`

`

`Le|
`
`US. Patent
`
`Nov. 26, 1996
`
`TWNOILIGQY
`
`Wvy9oud
`
`oleny008 NOILYWHOSNI
`
`¢‘Sls
`
`|9“Ola
`
`-087e03Na|.004
`
`fatalge=aYOSN3SOF
`LTT=
`aanOe[ae<ele
`
`Z
`
`
`
`Os?
`
`Sheet 8 of 10
`
`5,579,124
`
`
`
`
`
`dWVLSSWIL(3)6
`
`TTTTTTeTT
`
`
`

`

`U.S. Patent
`
`Nov. 26, 1996
`
`Sheet 9 of 10
`
`5,579,124
`
`g(n)
`
`ADDITIONAL
`
`INFORMATION
`
`

`

`U.S. Patent
`
`Noy. 26, 1996
`
`Sheet 10 of 10
`
`5,979,124
`
`0g
`
`
`
`092
`
`300930
`
`
`
`300930gc/
`
`W307|e;pm|6ON
`
`sq0030etaN/LSv30Vv0uE
`
`
`
`
` 07/sed|400930gowsdYSNNL
`
`of90d|
`
`W3dq0W>
`
`
`
`QNISSHI0UdVIVO
`
`d3ZWaLNaD
`
`ALNIVS
`
`|||
`
`AN3W93S
`
`
`
`

`

`5,579,124
`
`1
`METHOD AND APPARATUS FOR
`ENCODING/DECODING BROADCAST OR
`RECORDED SEGMENTS AND MONITORING
`AUDIENCE EXPOSURE THERETO
`
`This application is a continuation of application Ser. No.
`07/976,558, filed Nov. 16, 1992 now abandoned.
`
`BACKGROUND OF THE INVENTION
`
`The present invention relates to encoding and decoding
`broadcast or recorded segments such as broadcasts trans-
`mitted over the air, via cable, satellite or otherwise, and
`video, music or other works distributed on previously
`recorded media, as well as monitoring audience exposure to
`any of the foregoing.
`Broadcast segments include live programs, taped pro-
`grams, commercials and the like. These segments may be
`aired according to a wide variety of schedules, for example,
`national coverage, particular geographic coverageortofill
`otherwise unreserved programming slots. Furthermore, the
`scheduled broadcast time may be uniform nationwide or
`vary according to a broadcaster’s local considerations.
`There is a need to independently detect when segments,
`such as commercials, were actually broadcast over a given
`channel or by a given station.
`There is also a need to monitor the audience for broadcast
`segments because rates charged for broadcast
`typically
`depend on audience size. Further, some market research
`techniques involve testing the effect of broadcast segment
`frequency and/or nature on consumer purchase decisions.
`There are several conventional methods of detecting the
`identity of broadcast segments. However, each of these
`methods is limited in at
`least one respect, such as its
`complexity, its intrusiveness or inconvenience to audience
`members, or its vulnerability to errors caused by a noisy
`environment.
`
`In one such method, each of a numberofselected audi-
`ence members maintains a diary of which programs he or
`she viewed or heard. This methodrelies on the voluntary and
`timely cooperation of the selected audience members.
`Advertisers, advertising agencies and broadcasters have in
`the past expressed concerns that media experiences may not
`have been fully reported by respondents in their diaries. In
`particular, it has been inferred from survey data that the
`media experiences of young children, teens and young men
`are especially underreported.It is thought by some that such
`groups are either unable to complete the written diaries or
`find this task to be particularly tedious and thus neglect to
`enter complete information.
`To avoid the perceived drawbacks of manual recording,
`passive recording methods have been sought. Such passive
`recording methods would be characterized by the presence
`of a device which attempts to sense,
`in real
`time,
`the
`broadcast segments to which an audience member
`is
`exposed and record this information, which would later be
`retrieved at or uploaded to a centralized data processing
`facility. Since the information would be collected in com-
`puter readable form, data processing could be carried out
`readily with the use of a passive recording apparatus.
`Information collected by passive recording would be free of
`human error, and in this respect would enjoy improved
`reliability.
`Devices known as “personal passive people meters”,
`which are small and portable, have been proposed. Such
`devices are intended to be carried by persons whose broad-
`
`15
`
`20
`
`25
`
`30
`
`35
`
`45
`
`50
`
`35
`
`60
`
`65
`
`2
`cast segment exposure would be monitored. These meters
`would permit viewer/listener determinationat the individual
`level, which is highly desirable.
`A major problem in passive recording is to correctly sense
`the segment to which a viewer is being exposed. The
`proposed approaches involve attempting to identify both
`unmodified broadcast segments, and segments modified
`before broadcast to make them more readily identifiable.
`One approach to identification of unmodified segments
`involves pattern recognition. Each segment
`is analyzed
`before or after broadcast and its analyzed characteristics
`determine its “broadcast signature”. A table of broadcast
`signatures is created by, or made available to, each moni-
`toring station. In operation, a monitoring station attempts to
`analyzethe characteristics of a segment being broadcast and
`matchit to oncof the broadcast signatures,that is, recognize
`its pattern. This approach usesrelatively complicated tech-
`nology and is cumbersome to implement due to the need to
`enable each monitoring station to recognize new segments
`as they are introduced.
`Several identification approaches involve modifying the
`broadcast segments to provide a code which thedetecting
`equipmentis designed to recognize. An advantage of these
`approaches is that the monitoring stations need not be
`updated as new broadcast segments are created.
`U.S. Pat. No. 3,004,104 (Hembrooke) proposed to sup-
`press a narrow bandof frequencies (10 Hz wide) in a portion
`of the voiceband (1000 Hz) at timed intervals according to
`a predetermined code. However, if the suppression is short
`enough to be imperceptible as information to an audience
`member, then the suppression may be susceptible to inter-
`ference from ambient noise sources.
`
`It has also been proposed to modulate the audio frequency
`subcarrier with an identifying code of narrow bandwidth
`(100 Hz) and short duration (3 seconds) at the start and end
`of each segment. This technique is unsatisfactory because
`the metering equipmentfor a vieweror listener who tunes in
`a momenttoolate and tunes out a momenttoo early fails to
`sense the identifying code, and because it is vulnerable to
`noise.
`
`It has been proposed in the alternative to mix subaudible-
`frequency identifying codes with conventional audio in the
`program segments. This technique assumes the monitoring
`station would receive the broadcast, prior to audible repro-
`duction by the reception equipment, since some reception
`equipment is of poor quality and might not reproduce this
`information with sufficient fidelity for a personal metering
`device to recognize it. Thus, this technique is unsuitable for
`a personal meter of the type which monitors acoustic signals.
`A technique proposed for use with a musical recording
`comprises eliminating a sequence of six frequency bands
`from an audio signal, with the sequence varying during the
`course of the signal, and in place ofthe eliminated frequen-
`Cies, inserting a sequenceofcodesignals. This technique can
`be circumvented, since it
`is fairly easy to remove the
`included signals. Further, this technique is vulnerable to
`noise, especially acoustic noise.
`OBJECTS AND SUMMARY OF THE
`INVENTION
`
`Objects of the present invention includc the following:
`to provide information concerning broadcast or recorded
`segments to which audience members have been exposed;
`to provide information concerning the broadcast or
`recorded segments to which audience members have been
`exposed despite the presence of significant ambient noise;
`
`

`

`5,579,124
`
`3
`to provide methods and apparatus for encoding audio
`signals in which the codes are imperceptible as information
`to audience members;
`to detect which segments were actually broadcast in a
`given time period;
`to provide media exposure records for audience members
`to a centralized facility;
`to receive information from a centralized facility via an
`encoded transmission hidden within a pre-existing transmis-
`sion channel.
`
`information is
`In one aspect of the present invention,
`encoded in audio signals to be broadcast or recorded. A code
`signal having a predetermined bandwidth is modulated with
`an identification signal having a narrower bandwidth than
`the predetermined bandwidth to produce an encodedsignal.
`The encoded identification signal is mixcd with the broad-
`cast or recorded audio signal to produce an outputsignal.
`In another aspect of the present invention, an encoded
`broadcast or recorded segment signal including an audio
`signal portion having an encoded identification signal are
`reccived. The encoded identification signal is produced by
`modulating a code signal having a predetermined bandwidth
`with an identification signal having a narrower bandwidth
`than the predetermined bandwidth. The audiosignal portion
`is correlated with a copy of the code signal to recover the
`identification signal.
`the receiving and correlating is
`In some applications,
`carried out by a personal unit worn orcarried on the person
`of an audience member,
`that produces a record of the
`broadcast or recorded segments to which the audience
`member has been exposed. This record, with identification
`of the audience member,is uploadedto a centralizedfacility.
`A separate monitoring unit performs receiving and cor-
`relating in like manner as the personal units and may also
`extract additional information contained in the broadcast or
`recorded segment to produce a full record of what was
`broadcast. This monitoring unit communicates with the
`centralized facility to upload information thereto.
`The centralized facility matches the individual audience
`records with the additional information pertaining to the
`items in these records to provide a full record of who was
`exposed to what, and when.
`The above, and other objects, features and advantages of
`the invention, will be apparent in the following detailed
`description of certain illustrative embodiments thereof
`which is to be read in connection with the accompanying
`drawings forming a part hereof, and wherein corresponding
`parts and components are identified by the same reference
`numerals in the several views of the drawings.
`BRIEF DESCRIPTION OF THE DRAWINGS
`
`FIG. 1 is a block diagram of an encoder in accordance
`with an embodimentof the present invention;
`FIGS. 2A, 2B and 2C are block diagrams of personal
`monitors for use with the encoder of FIG.1;
`FIGS. 3A-3Kare frequency use charts used in explaining
`the embodiments of FIGS. 1, 2A, 2B and 2C;
`FIG. 4Ais a block diagram of an encoder in accordance
`with another embodimentof the present invention;
`FIG. 4B is a block diagram of an apparatus for program-
`ming a ROM ofthe FIG. 4A encoder with time domain code
`signals;
`FIG. 4C is a block diagram of an encoding system in
`accordance with an embodimentof the present invention;
`
`4
`FIG. 5 is a block diagram of an encoder in accordance
`with a further embodimentof the present invention;
`FIG.6 is a block diagram of a personal monitor for use
`with the encoder of FIG. 5;
`FIG. 7 is a block diagram of an encoder in accordance
`with still another embodiment of the present invention,
`FIG. 8 is a block diagram of a personal monitor for use
`with the encoder of FIG. 7; and
`in
`FIG. 9 is a block diagram of a monitoring unit
`accordance with still another embodiment of the present
`invention.
`
`15
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`50
`
`55
`
`60
`
`65
`
`DETAILED DESCRIPTION OF CERTAIN
`ADVANTAGEOUS EMBODIMENTS
`
`In certain advantageous embodiments, the present inven-
`tion adds identifying information to the audio portion of a
`broadcast segment before the segment is broadcast using a
`spread spectrum technique selected from among several
`alternatives, and includes a passive monitoring device which
`operates without human action to sense the identifying
`information in the broadcast segment and record it. The
`terms “meter” and “metering device” are sometimes used
`herein to refer to devices such as passive broadcast moni-
`toring devices. At periodic intervals, the recorded informa-
`tion in each meter is uploaded to a centralized data process-
`ing facility for permanentstorage.
`In such embodiments,
`the spread spectrum techniques
`employed typically encode identifying information having a
`relatively low data rate and formed into an identification
`signal having a narrow bandwidth, referred to herein as
`X(w), x(t) or x(n). As usedherein,the term “signal”includes
`both an electrical signal and a representation of information
`which is stored, processed and/or transmitted, as well as any
`other form in which information is embodied. The term
`“bandwidth” as used herein includes a difference between
`frequency band limits as well as a frequency interval or
`range of frequencies. The explanations of terms as used
`herein are provided for exemplary purposes, and are not
`intended to be limiting, as appropriate other meanings for
`such terms may occur to those of ordinary skill in the art. In
`an advantageous embodiment, the thus-formed identifica-
`tion signal is modulated by a code signal, also known as a
`spreading signal, which is independent of the data and has
`a much wider bandwidth.
`
`The code signal is a pseudo-random signal which, after
`modulation with a broadcast segment, will be perceived, if
`at all, as a low-level white noise, generally referred to as
`hiss, and not as information. The code signal is mixed into
`the audio signal at a level sufficiently below the regular
`broadcast audio signal
`level to make it imperceptible as
`information, and in the alternative, may be mixed with the
`audio signal at lower levels depending on the manner in
`which the audio signal
`is acquired for decoding,
`for
`example, as a basebandsignal versus an acoustically repro-
`duced signal.
`One advantageous code is a sequence of tones added to
`the voiceband, which occupics approximately 300-3,000
`Hz,since all broadcast formats and all reception equipment
`provide for reproduction of voice information of at least
`reasonable quality.
`At each metering device, the audio signal portion of the
`broadcast segment is subjected to a correlation process, such
`as one of the processes described below, with a synchronized
`reference copy of the code signal to recover the identifica-
`
`

`

`5,579,124
`
`5
`tion signal, compared with valid information items (such as
`valid channels in the relevant geographic area), and subse-
`quently stored.
`Dueto the use of spread spectrum encoding,the identi-
`fying information may be successfully recovered despite the
`presenceof substantial ambient noise in the audio bandwidth
`in which the code signal is transmitted. Furthermore, the
`encoded identification signal can be made imperceptible to
`the audience.
`
`In certain embodiments, the audio signal portion, typi-
`cally 20-22,000 Hz, of a segment to be broadcasted is
`encoded with station, channel or other program source
`identifying information by mixing it with a code signal
`modulated with an information signal which convcys this
`information. The information uniquely identifies the particu-
`lar broadcasting source. The amount of information per
`broadcast segmentcan be keptshort, if only broadcasttimes
`and the source of the broadcast, thatis, the station or channel
`and notnecessarily the identity of the program segment, are
`transmitted.
`
`A passive meter, preferably worn by a selected memberof
`the audience on his or her person, recovers the source
`identifier and stores it in a local memory with a time and date
`stamp. At the end of each day, the meter is put into a base
`unit so it can be recharged,its recorded information can be
`extracted, and, if desired, new information can be loaded
`into the meter. The extracted information may be collected
`by a storage and transmission unit in the household, and
`either the base unit or the storage and transmission unit may
`be employed to transmit the information over a dial-up
`telephoneline to a centralized facility when the telephone
`linc is not used by a member of the household. Several
`passive meters can be servedby a single baseunit or storage
`and transmission unit. Alternatively,
`the meter may be
`physically sent to a centralized facility to extract its recorded
`data.
`
`Furthermore, additional information regarding the broad-
`cast segment, for example, identifying the particular pro-
`gram or commercial, is also encoded into the audio signal
`portion of the segment. This additional information may use
`a code signal having a frequency range substantially coex-
`tensive with the full range of the audio signal, or having a
`range above the voiceband but within the audio signal range,
`for example, 4,000-20,000 Hz. Alternatively, the additional
`information may be formed into an additional information
`signal which directly modulates the audio signal, that is,
`without spread spectrum encoding, below or above the
`voiceband, or which modulates anotherportion of a broad-
`cast segment, such as a video signal.
`A separate monitoring device receives the baseband
`broadcast segment and extracts therefrom the additional
`information regarding the broadcast segment,andsendsit to
`the centralized data processing facility whereit is matched
`with the source identification information from the personal
`monitoring devices, to provide a full audience record of who
`was exposed to what, and when.Alternatively, the separate
`monitoring device may be located at the broadcastsite, for
`example, at the headend of a cable system, and may monitor
`the signals immediately before they are cablecast.
`An advantageous methodfor spread spectrum encoding of
`the source identification informationutilizes direct sequence
`encoding in the frequency domain. Alternative methods
`include direct sequence encoding in the time domain, and
`frequency hopping. Each of these methods is further
`described below. However,
`the present invention is not
`limited to these methods, and other spread spectrum meth-
`
`20
`
`25
`
`40
`
`50
`
`35
`
`60
`
`65
`
`6
`ods using time hopping or pulse-FM systems, or a hybrid
`method,are feasible.
`An embodiment of the present invention will now be
`described in connection with FIG. 1, which shows an
`encoder, FIG. 2A, which shows a personal monitor, and
`FIGS. 3A-3K, which show frequency use charts.
`FIG. 1 shows an advantageous embodimentof an encoder
`100 according to the present
`invention. Encoder 100
`includes input
`terminals 105 and 110, modulator 120,
`inverse transformer 130, buffer 140, digital-to-analog (D/A)
`converter 150, low pass filter 160, mixer 170 and output
`terminal 175,
`
`Sourceidentification signal X(co), composedin bit form in
`the frequency domain, is supplied to the input terminal 105,
`while a frequency domain antipodal code signal G(w) also
`in bit form is supplied to the input terminal 110. An antipodal
`signal has only opposedvalues, such as “1” and “~1”. In this
`instance, the values of both X(w) and G(@) are composed of
`real numbers, and imaginary portions thereof are set of zero.
`These signals are described in detail below.
`As used herein, “bit” refers to a unit of data, such as a
`portion of a source identifier, and “chip” refers to an
`elementary unit of a code. One bit corresponds to many
`chips, since the bandwidth of the information signal
`is
`narrower than the predetermined bandwidth of the code
`signal. In the frequency domain, each chip is represented by
`a “point” which is essentially a data value.
`The codesignal can be changed, for example,on a daily
`basis, to meet a variety of needs, such as identification of
`taped replays, limiting the collected data to a predetermined
`survey time period, or discouraging unauthorized access.
`Codesignals can be provided to one or more encoders from
`a centralized facility via any of a number oftransmission
`techniques. For example, the code signals can be transmitted
`via the public switched telephone network, a local area
`network, satellite transmission, or as data encoded in a
`broadcast in the manner described below in connection with
`FIG. 9, Use of different codes for radio and television
`enables the same personal monitor to collect radio or TV
`only data. Alternatively, codes may be assigned based on
`geographic location,or to restrict audience exposure moni-
`toring to only commercial advertisements.
`The source identification signal, X(w), and the code
`signal, G(@), are supplied to modulator 120, which modu-
`lates these signals using, for example, direct multiplication,
`logical exclusive OR, or another combining technique for
`individual
`frequency components,
`to form a frequency
`domain encoded source identification signal.
`_A frequency domain encoded signal, when properly
`selected, has the property of matching its spectrum to the
`typical frequency response of the receiver circuitry and
`speaker in use by an audience member, as well as to
`compensate for the room or other acoustic environment in
`which monitoring will occur.
`The frequency domain encoded sourceidentification sig-
`nal is supplied to inverse transformer 130, which performs
`an inverse fast Fourier transform (FFT) or wavelet transform
`so as to produce a time domain encodedsourceidentification
`signal
`that
`is supplied to buffer 140, which holds, for
`example, 2,048 data items, and is shown as a random access
`memory used according to a first-in-first-out scheme. The
`contents of buffer 140 are fed to D/A converter 150, which
`is a 16-bit converter, for example, thereby providing about
`a 90 dB range oflevels in the analog encoded identification
`signal.
`In one embodiment,the converter 150 samplesat a rate of
`8,192 samples per second. The length of buffer 140 corre-
`
`

`

`5,579,124
`
`7
`spondsto onebit time at the selected sampling rate, thatis,
`(8,192 samples per second)/(4 bits per second)=2,048
`samples/bit. The corresponding FFT has a length of 1024
`points in the frequency domain, with each point correspond-
`ing to 4 Hz. The 676 points within the frequency range
`300-3,000 Hz are used, while the 75 points corresponding to
`the range 0-296 Hz and the 273 points within the range
`30044092 Hz are not used. The analog encoded identifi-
`cation signal
`is supplied to low pass filter 160, which
`removes spurious signals outside of the desired range.
`At the mixer 170, thefiltered encoded identification signal
`is combined with the audio portion of a segment in a ratio
`selected to maintain inaudibility and supplied to an output
`terminal 175 of the encoder 100, andis then broadcast with
`the other portions of the segment, if any, in a conventional
`manner such as by REF, satellite or cable broadcast, or is
`recorded on tape or other recording medium. Thelevel at
`which the encoded identification signal
`is combined is
`chosen to be approximately the normal noise level tolerated
`by most audio programs. Additional information, intended
`for a monitoring device distinct from the personal monitor,
`may also be separately supplied to mixer 170, for combi-
`nation with the encoded identification signal and audio
`portion.
`The modulating through mixing processing steps per-
`formed in the aforementioned elements of the encoder 100
`are repeated until the source identification information is
`fully encoded in the audio portion of the segment to be
`broadcast or recorded. These steps can be repeated to encode
`the source identification in various places or continuously
`through the audio portion of the segment. The succeeding
`identification information may be changed to reflect a
`change in the source of the segment, or as otherwise
`appropriate.
`FIG. 2A shows one advantageous embodimentof a per-
`sonal monitor 200 according to the present invention. Per-
`sonal monitor 200 includes a microphone 230, amplifier
`240, low passfilter 250, analog-to-digital (A/D) converter
`255, buffer 260,
`transformer 265, corrclator 270,
`input
`terminals 275 and 285, combiner 280, and memory 290. The
`outer dashedline in FIG. 2A generally indicates the enclo-
`sure of a metering device to be worn on the person, ¢.g.,
`clipped to a garment worn by the audience member.
`As shown in FIG. 2A, the encoded audio portion of the
`broadcast segmentis received at an input terminal 205 of a
`typical broadcast receiver 210, which acoustically repro-
`duces the audio portion using a speaker 220. Receiver 210
`and its speaker 220 represent devices normally used in
`households and elsewhere by audience members to acous-
`tically reproduce broadcast audio signals. Alternatively, a
`recorded segment containing an encoded audio portion may
`be reproduced, such as by a video cassette recorder, and the
`audio portion thereof acoustically reproduced by a speaker
`such as speaker 220.
`The acoustically reproduced audio portion of the broad-
`cast or recorded segmentis received by the microphone 230
`ofthe personal monitor 200, which transduces the acoustic
`energy into an electrical signal. The transduced electrical
`signal is supplied, via a physical line or wireless transmis-
`sion, to amplifier 240 shown as an automatic gain control
`amplifier, which produces an output signal at an increased
`powerlevel.
`In FIG.2A,the combination 235A of microphone 230 and
`amplifier 240 is shown as contained within the personal
`monitor 200 worn by an audience member. An alternative
`arrangementis depicted in FIG. 2B, showing a combination
`
`10
`
`20
`
`25
`
`30
`
`35
`
`40
`
`45
`
`55
`
`60
`
`65
`
`8
`235B which functionally corresponds to the combination
`235A. The combination 235B includes a first unit 241,
`intended to be worn by an audience member and physically
`separate from the remainder of monitor 200, and a second
`unit 242 contained within an enclosure containing the
`remainder of monitor 200. The arrangement shown in FIG.
`2B is intended especially for situations where the audience
`memberis a child, or other situations where miniaturization
`of the apparatus worn by the audience member is advanta-
`geous.
`The first unit 241 of combination 235B comprises micro-
`phone 230, transmitter 231 and antenna 232. The transduced
`electrical signal from microphone 230 is supplied to a
`transmitter 231 which is adapted for generating a signal
`suitable for wireless transmission from the transduced sig-
`nal, which is supplied to antenna 232. The antenna 232
`serves to produce a wireless transmission of the signal from
`transmitter 231.
`
`The second unit 242 of combination 235B comprises
`antenna 233 and receiver 234. The antenna 233 is operative
`to receive the wireless broadcast from antenna 232, and
`convert it into a received electrical signal which is supplied
`to receiver 234 which serves to produce an output signal at
`an increased powerlevel, corresponding to the output of
`amplifier 240.
`FIG. 2C shows anotheralternative combination 235C, for
`use when the audience memberis exposed to radio bro