WO2007070846A2

WO2007070846A2 - Method and apparatus for automatic detection and identification of broadcast audio or video signals

Info

Publication number: WO2007070846A2
Application number: PCT/US2006/062079
Authority: WO
Inventors: Stephen J. Brown
Original assignee: Mediaguide, Inc.
Priority date: 2005-12-15
Filing date: 2006-12-14
Publication date: 2007-06-21
Also published as: WO2007070846A3

Abstract

This invention relates to the automatic detection and identification of broadcast programming, for example music, speech or video that is broadcast over radio, television, the Internet or other media. 'Broadcast' means any readily available source of content, whether now known or hereafter devised, including streaming, peer to peer delivery or detection of network traffic. A known program is registered by deriving a numerical code for each of many short time segments during the program and storing the sequence of numerical codes and a reference to the identity of the program. Detection and identification of an input signal occurs by similarly extracting the numerical codes from it and comparing the sequence of detected numerical codes against the stored sequences. The system that operates the process can be distributed across a large territory with multiple broadcast sources whereby the system is a distributed system.

Description

IN THE UNITED STATES PATENT AND TRADEMARK OFFICE

SPECIFICATION

INVENTION: METHOD AND APPARATUS FOR AUTOMATIC DETECTION

AND IDENTIFICATION OF BROADCAST AUDIO OR VIDEO

SIGNALS.

METHOD AND APPARATUS FOR AUTOMATIC DETECTION AND IDENTIFICATION OF BROADCAST AUDIO OR VIDEO SIGNALS.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 : The data flow between components of the media broadcast monitoring system when operating the method.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

Background.

The present invention relates to the automatic recognition of widely disseminated programming, such as radio, television or digitally delivered content over the Internet.

Owners of copyrights in broadcast programming, including advertisers, need to measure when and where their programming has been broadcast in order to correctly compute performance royalties, confirm compliance with territorial restrictions or verify that certain advertising has been aired as scheduled. The traditional method for monitoring the radio or television has involved using humans to listen or watch and then record that which they hear or see, or alternatively, rely on the broadcast records of radio and television stations. This is a labor intensive process that has limited efficiency or accuracy. Tt is an object of the invention to use advanced computing systems to fully automate this process. In this manner, audio or video content is registered into the system, and then, in the case of audio detection, radio, the soundtrack from television or other sources of widely distributed audio content are input into the system. In. the case of video, the video signal is input into the system from whatever its source. By means of the invention, the detection and identification of registered programming content takes place automatically.

Prior Art:

A number of methods have been developed to automate the detection of broadcast programming. These techniques generally fall into one of two categories: cue detection or pattern recognition. The cue detection method is exemplified by U.S. Pat.

Nos. 4,225,967 to Miwa et. al.; 3,845,391 to Crosby and 4,547,804 to Greenberg.

These techniques rely on embedded cues inserted into the program prior to distribution. These approaches have not been favored in the field. In audio, the placement of cue signals in the program have limited the acceptance of this approach because it requires the cooperation of the program owners and/or broadcasters— thus making it impractical.

The pattern recognition method generally relies on the spectral characteristics of the content itself to produce a unique identifying code or signature. Thus, the technique of identifying content consists of two steps: the first being extracting a signature from a known piece of content for insertion into a database, and the second being extracting a signature from a detected piece of content and searching for a signature match in the database in order to identify the detected content. Tn this way, the preferred approach relies on characteristics of the broadcast content itself to create a signature unique to that content. For example, US Patent No. 4,739,398 to Thomas, et.al. discloses a system that takes a known television program and creates for each video frame, a signature code out of both the audio and the video signal within that frame. More recently, similar detection systems have been proposed for Internet distributed content, for example application PCT WO 01/62004 A2, filed by Ikeyoze et. al.

For audio by itself, U.S. Pat. No. 3,919,471 to Moon discloses an audio identification system where only audio signals are used, but it is of limited utility because it attempts to correlate an audio program represented by a limited time slice against the incoming broadcast signal. The disclosed method of matching in Moon is highly compute intensive because it relies on direct signal correlation. Further, this approach is unfavorable because it has been found to be limited in accuracy, especially if the program is time compressed or altered in other ways prior to detection. It is also prone to false positive identifications and is computationally uneconomic if the size of the time slice is expanded to improve its correct identifications. Lert, et. al. describes in U.S. Pat. No. 4,230,990 a way to mitigate the computational workload of the correlation method by combining it with the coding method of the first category: either an artificial code or some other naturally occurring marker is detected in the program indicating the beginning of a section of the program, and then a feature signature is measured at a pre-determined amount of time later. This method has limited utility in audio-only applications, where either an audible code has to be inserted into the audio to create the cue, thus degrading it or requiring cooperation of the content source, or reliance on natural markers indicating the start of a new audio program which is highly unreliable. In U.S. Pat. No. 4,677,466 Lert, et. al. further describes an improvement on the invention that waits until a "stability condition" has occurred in the signal before measuring and calculating a signature, but the reliability of the method is limited by the size of the sample time slice. U.S. Pat. No. 4, 739,398 to Thomas et. al. addresses the data processing load problem by randomly choosing portions of a signal to sample as input to the invention's signature generating process.

U.S. Pat. Nos. 5,436,653 to Ellis, et. al. and 5,612,729 to Ellis, et. al., disclose a more complex way of calculating a unique signature, where the audio signature corresponding to a given video frame is derived by comparing the change in energy in each of a predetermined number of frequency bands between the given video frame and the same measurement made in a prior video frame. However, the matching technique relies on a combination of the audio and video signatures or the use of a natural marker, in this case, the start or ending of a program. Thus, this method suffers the same problem as Lert with regard to audio-only programming.

In addition, U.S. Pat. No. 5,918,223 to Blum, et. al., discloses the use of audible features within audio programming to create a single signature value for each audio program, particularly the group of amplitude, pitch (i.e. fundamental), bandwidth, bass (i.e. rhythm analysis), brightness (i.e. shape of the frequency response of the program), and Mel-frequency cepstral coefficients. The aggregation of these detailed features across long periods in the audio produce highly variable results, and do not possess sufficient robustness in real-world broadcast situations. U.S. Pat. No. 5,210,820 and 4, 843, 562, both to Kenyon, discloses a digital circuit that uses the envelope (e.g loudness) features in the audio signal in order to create a signature. The approach is designed to address the time compression problem by application of time warping techniques. Reliance on loudness has other robustness problems that also make it difficult to use in real-world environments. U.S. Pat. Application No. 20030086341 filed by Wells, Maxwell, et. al., discloses a system where an audio signature is created using pre-determined numbers of digital samples counted from pre-determined locations from the start point of the music. This approach is much less reliable for broadcast or cases where the audio is detected in analog form, or in cases where the playback of the programming has changed speed, frequency equalization from the original track has been applied, or the audio dubbed into the programming segment.

The present invention describes a system and method whereby identification of known audio or video programming can be done without any reliance on a tandem video signal (in the audio case) or normative markers in the signal indicating a known time in the program and with unique and novel ways to calculate codes representing the characteristics of the audio program without requiring impractical computational capabilities. Benefits of this system and method are accuracy, speed, robustness to playback speed variation and the ability to perform the identification process in real time, without reliance on any embedded cue or watermark. In addition, the present invention takes advantage of the availability of low cost, high performance computing platforms in order to implement a high speed database searching methodology. In a further improvement over the prior art, the system is designed so that it can efficiently monitor multiple broadcast media outlets across a wide geographic territory.

Detailed Description. A. Overview

The broadcast monitoring and detection system embodying the invention works in two phases: registration and detection. During the registration phase, known programming content is registered with the system by sending the program, as digital data, into the system. A series of signatures, in the case here, a pattern vector and more generally in the art a "fingerprint" or "signature", are stored as a sequence of data records in a database, with the identity of the program content cross-referenced to them as a group. During the second phase, unidentified programming is input into the system. Such programming can include radio, television, internet broadcasts or any other source of audio or video programming, whether terrestrial broadcast, satellite, internet, cable television or any other medium of delivery, whether now known or devised in the future. While such programming is being monitored, the pattern vectors of the programming (or any other signature generating technique) are continually calculated. The calculated pattern vectors are then used to search for a match in the database. When a match is found and confirmed, the system uses the cross-referenced identity in the database to provide the identity of the content that is currently being played. In the preferred embodiment, the system is software running on a computer, however, it is envisioned that special purpose hardware components may replace parts or all of each module in order to increase performance and capacity of the system.

In the preferred embodiment, a computer containing a central processing unit is connected to a sound card or interface device into which audio programming is presented. During the registration phase, the CPU fetches the audio or video data from the sound card, calculates the pattern vector data, and then, along with timing data and the identity of the program, these results are stored in a database, as further described below. Alternatively, the data may be loaded directly from authentic material, such as compact discs, mp3 files or any other source of digital data embodying the signal. For non-audio applications, the source of material can be DVD disks, masters provided by movie studios, tapes or any other medium of expression on which the program is fixed or stored. Of course, for some material which may not have a readily available source, then the audio or other program signal is used in the following manner. If the system periodically detects an unknown program but with the substantially the same set of signatures each time, it assigns an aribitrary identifier for the program material and enters the data into the database as if the program had been introduced during the registration phase. Once the program identity is determined in the future, then the database can be updated to include the appropriate information as with authentic information while at the same time providing the owner of the programming the use data detected even when the identity of the program was not yet known. The database, which is typically a data file stored on a hard drive connected to the central processing unit of the computer by means of any kind of computer bus or data transmission interface, including SCSI.

During the detection phase, the CPU fetches the program data from the sound card or video card, or loads it from a data file that may be stored on the computer hard drive or external media reader. The CPU calculates the pattern vector data, and then, along with the timing data, submits database queries to the database stored on the hard drive. The database may be the same hard drive as in the computer, or an external hard drive accessed over a digital computer network. When matching data is found, the CPU continues to process the data to confirm the identification of the programming, as described further below. The CPU can then communicate over any of a wide variety of computer networking systems well known in the art to deliver the identification result to a remote location to be displayed on a screen using a graphical user interface, or to be logged in another data file stored on the hard drive. The program that executes the method may be stored on any kind of computer readable media, for example, a hard drive, CD-ROM, EEPROM or floppy and loaded into computer memory at run-time. In the case of video, the signal can be acquired using an analog to digital video converter card, or the digital video data can be directly detected from digital video sources, for example, the Internet or digital television broadcast.

Typically, the registration phase is performed at a central location where staff can supervise the process. Alternatively, multiple service centers can register items and share the data. The detection phase, however, is typically distributed across a wide territory. Pending PCT application PCT/US05/04802 to Cheung, is incorporated herein by reference. In that application, detail about how the registration creates a database of fingerprints is disclosed, as well as the process of matching incoming signal data with fingerprints in the database. Practitioners of ordinary skill will recognize that centralizing the detection phase, whereby a number of signals are transmitted to a central location or a single computer for detection, is inefficient for two reasons: first, the large amount of data that would have to be transported from a remote location to the central repository of data. Tn addition, re-transmission of signals from remote locations to a central location may implicate certain copyrights. In any case, the most efficient manner of conducting detections from a number of broadcast sources across a wide territory is to place a computer that executes the detection phase in each location, and centralizing the confirmed detections. This results in a substantial savings in processing time and data networking costs and requirements, albeit at the cost of deploying more computer hardware.

Computer hardware costs increase with the size of the database that has to be deployed at each detection location because of the size of storage media that has to be supported. Practitioners of ordinary skill will recognize that from a massive database of songs, only a few are played on a regular basis. Therefore, it is possible to save costs in deploying the distributed system by limiting each local detection computer to house only a subset of the entire database of registered songs. The system architecture is then a central database housing all of the registered songs and then distributed detection stations (also referred to herein as remote monitoring node) that receive updates to their database from the central database in such a manner that the local detection stations are only required to house a subset of the volume of data comprising the central database. This introduces a cost savings in deploying the system.

LMS

The sub-system that manages the local detection station databases is called the LMS or Library Management System. LMS was created to maximize detections within the confines of available time, CPU capacity and memory on the remote monitoring nodes and the N-tier nodes. Reiterating the points made above, LMS is necessary because content identification systems have three fundamental properties:

• The amount of content that is monitored is very large; • The amount of content that can be quickly monitored per box is finite; and

• Fingerprint patterns stored in the local detection database that never generate detections waste time and money.

LMS combines metadata, meaning song information including title, artist, publisher, release date, with heuristic rules and historical data about what songs were detected in which remote monitoring node and when, in order to choose a subset of the universe of possible song identification information (that is the fingerprint patterns associated with the song) into a given remote monitoring node .

In one case, the remote monitoring node is required to monitor for a particular song. In that case, the rule is applied that the central database delivers the fingerprint and metadata associated with that song to one or more remote monitoring nodes. In this disclosure, reference to delivering content to a node includes delivering the fingerprint patterns and metadata associated with the content.

In another case, when new content is registered in the central database, the database can include information about the song, advertisement or other content that recites the expected start and/or stop dates, or promotion dates of the content. This metadata is used by the LMS system to select which fingerprint data is sufficiently timely and should be sent to the remote monitoring nodes and which data residing on such node should be removed.

The LMS method is used to fill the balance of available space in each database. This is an optimization problem with the primary goal of maximizing the number of detections that can be generated from the mix of tuned stations feeding into that monitoring node.

Content that has already been registered in the central database but is not subject to either of the described rules is selected to be placed in the remote monitoring node if it is categorized as 'active.' The set of content that is 'active' will be filtered through the following ranking functions:

• Detected by any remote monitoring node during the most recent predetermined period of time; • Detected by any node within a defined region or market during the most recent predetermined period of time;

• Detected on a specific radio station or by a specific remote monitoring node during the most recent predetermined period of time;

• Detected anywhere within the same genre format as that being monitored by a particular remote monitoring node during the most recent predetermined period of time.

An example of a scoring function is a linear combination of Boolean results of the example tests. Different weights can be applied to each test and the sum calculated. Where the score is sufficiently high, the content (that is, the fingerprint patterns and metadata associated with the content) is housed in the corresponding remote monitoring node. In the preferred embodiment, the first test is such that if the Boolean result is true, then that content is considered active regardless of the other three rules. Alternatively, the time related rules be scored as some coefficient times the number of days from the last detection or time between the event in question, that is, the start date or end date of a song or advertisement promotion and the current date, or from the current date to when the last stated promotion event in the metadata. The point is a number of rules may be applied automatically while the system runs in order to determine the most efficient allocation of remote monitoring node data storage to the content set being monitored across the entire territory.

Practitioners of ordinary skill will recognize that radio stations and other broadcast sources typically select content for broadcast that fits one or more genres specific to that station. Therefore, the song metadata, which can contain genre format information, can be used to select which fingerprint data will be provided to a remote monitoring node.

The parameters in each rule may differ by region (usually a country or continent), content type (e.g. song, advertisement) or source (direct from broadcaster or clipped from television or radio). AU of the scoring functions are evaluated for all content, and a ranking of content that should be put into the remote monitoring node database is generated. The rules are expressed in computer programming code, which, in the preferred embodiment, is the SQL query language. The rules may be expressed as a set of boolean conditions and the domains involved. In the preferred embodiment, the domains are defined as follows: ents = the union of the subsets, that is, the set of all content in the entire system, ads = harvested content we found (content that has not been registered but is recognizable by the system) not_harvested (registered content) songs = remote_songs (global & per country) local_songs (global & per country) = lower detecting content stations = Audio sources monitored formats = Format/genre monitored markets = Geographical regions hosts = Hosts that will have a database built, the union of remote_tier_hosts (hosts out in the remote markets, i.e. remote monitoring nodes)

N_tier_hosts (dedicated N-tier hosts)

For all ents the system works through the other domains to evaluate if an ent will be in the database for a given host. The following functions are applied and Boolean results calculated:

Function MM: boolean true if for a given ent it must be monitored everywhere or on the given N-tier node or remote monitoring node, specific market or broadcast station associated with that node.

Function New: boolean for remote_tier_host$: true for remote_songs where the 'promote date' <= 1.9 months true for local_songs where the 'promote date' <= 1.5 months true for ads where the 'promote date' <= 4.5 days for N-tier hosts true if within local_songs false if in all remote monitoring node databases true for ads where the 'promote date' <= 2 weeks

Function Played: boolean true for songs detected anywhere <= 0.95 months true for ads detected anywhere <= 9.5 days true for songs played on the associated station <= 25 months true for ads played on the associated station <= 22 days true for ads not harvested played on the associated station <= 15 months

The above tests are then combined in the following logic rule. ! 'demoted' & ( MM | New | Played ) That is, a given N-tier node or remote monitoring node, keeps content that has not been demoted and meets either of the MM, New or Played criteria.

In the preferred embodiment, rule criteria keyed off of format are only run occasionally as a side process to determine if a song needs to be added back into the daily run of the rules.

Practitioners of ordinary skill will recognize that instead of Boolean logic rules, scores can be calculated using numeric values. For example a score based on the time since the recent play could be calculated. Those songs where that number is too high as compared to a threshold can be considered to be scoring outside the criteria associated with recent play. In addition, the time parameters associated with the time windows can be adjusted to achieve the most desired resource allocation result.

In the preferred embodiment, content is matched in the local remote monitoring node first, then sent to the next tier up if no detection match is found. The next tier up is called the N-tier, which is a set of detection nodes that service the remote monitoring nodes. From a system architecture view, the N-tier nodes stand between the remote nodes and the central system. Practitioners of ordinary skill will recognize that typical computer networking systems, including the internet, Ethernet and TCP/IP based networking can be used to connect these systems together in the desired network topology. The set of content that does not rank high enough to be housed in a remote node is aggregated and fed into the algorithm again to generate ranked content that is loaded into the N-tier boxes. In this step, slightly different variations of the rules may be applied to determine which content deserves an intermediate status. As a result, the N-Tier units also run detections, but house content that is less frequently detected.

In the preferred embodiment, the LMS system is operated by the remote monitoring nodes, the N-tier nodes and the central system on a regular basis. Each device acts on the rules based on their local context. The rules run at the N-tier do not have to be exactly the same as at the remote nodes. As a result, the system is self-correcting due to the combination of valuation functions across the systems. Because the evaluation functions are detection based, as soon is content detected in an N-tier box, sufficiently to pass the scoring function, it will 'bubble-up' and may score high enough on the next round to make it into remote databases, and content that is not producing sufficient detections in a remote monitoring node to meet the scoring requirements will 'fall-off into the N-tiers, possibly eventually off all active databases if it does not meet the N-tier level scoring. In addition, remote monitoring nodes in one region of a country will end up with different databases than those in another because of the local preference of certain genres. The LMS system maintains this automatically as a result of the continuous evaluation of detection data. This method not only puts content where it is most likely to be detected, but it improves the detection latency and overall performance of the distributed system by minimizing N-tier intercommunication.

Practitioners of ordinary skill will recognize that the N-Tier systems are not a necessary feature to implement the invention. Tt is possible to have the central servers directly manage the remote local nodes, such that infrequently detected content is processes in the central server. In addition, practitioners of ordinary skill will recognize that when the N-tier takes over a piece of content from the remote node, the remote node's series of incoming fingerprint data must be delivered to the N-tier node for database processing. Alternatively, the signal itself can be forwarded to the N-tier node. In another embodiment, demotion of content from a remote monitoring node does not result in the content or its fingerprint data being transferred to another node, but rather, that an index or other identifier is passed to the other node to indicate that such content and its fingerprint data is being removed from the remote monitoring node. The other node that will stand as the detecting system for the less frequently detected content can either already house the demoted content or upload the data from a central server.

Practitioners of ordinary skill will recognize that this system architecture and method is applicable to any kind of content identification technology, where signal characteristics of known content are stored in a database, and then the database searched using incoming signal characteristics to find a match. Content can include other kinds of detected data, for example, facial recognition characteristics that have to be searched in distinct locations. In this case, the "fingerprint" or pattern vectors are replaced with numerical data sets representing characteristics of the face, and song names with indicia of identity of the person. In the facial recognition application, there are video cameras or still cameras that are sources of facial images. Theses can be used to create a database of known people, whose facial features are stored in the database in the form of a numerical representation. Tn the distributed system, some of those facial feature data records are distributed to the remote nodes. At the remote node, a video camera or still camera takes images of unknown people as a source of data. This data is then distilled to the numerical representation of the facial features. In order to identify people at the remote node, the incoming numerical representations are used like the pattern vectors to create search queries of the database. The remote node holds those data records that meet the scoring requirement, which can be time periods since the prior identification and other criteria. Regardless of how the signal fingerprints are calculated or which matching algorithms are used, the same problem remains to be solved, that is, how to organize the system and remote monitoring nodes to store only the more frequently detected content and to dynamically adjust that storage.

Although the present invention has been described and illustrated in detail, it is to be clearly understood that the same is by way of illustration and example only, and is not to be taken by way of limitation. It is appreciated that various features of the invention which are, for clarity, described in the context of separate embodiments may also be provided in combination in a single embodiment. Conversely, various features of the invention which are, for brevity, described in the context of a single embodiment may also be provided separately or in any suitable combination. It is appreciated that the particular embodiment described in the Appendices is intended only to provide an extremely detailed disclosure of the present invention and is not intended to be limiting. It is appreciated that any of the software components of the present invention may, if desired, be implemented in ROM (read-only memory) form or stored on any kind of computer readable media, including CD-ROM, magnetic media, or transmitted as digital data files stored in a computer's memory. The software components may, generally, be implemented in hardware, if desired, using conventional techniques.

The spirit and scope of the present invention are to be limited only by the terms of the appended claims.

Claims

WHAT TS CLATMED:

1. In a system for detecting broadcast content occuranccs comprised of at least one remote monitoring node that monitors at least one broadcast source, a database of at least one signature associated with at least one item of content and where the system determines occurance data of said content items being broadcast by said broadcast sources, a method for determining which of the at least one signatures to store in the portion of the database controlled by the at least one remote monitoring node comprising: determining occurance data associated with the at least one content item; determining whether the at least one content item is active; and adding the signature associated with the at least one content item to the portion of the database controlled by the at least one remote monitoring node if the at least one content item is active.

2. The method of Claim 1 where the second determining step is comprised of calculating a score based on the occurance data; comparing the score to a predetermined value.

3. The method of Claim 1 where the second determining step is further comprised of any of the following steps:

Determining whether the occurance data indicates that the content was detected by the remote monitoring node during a predetermined period of time; Determining whether the occurance data indicates that the at least one content item was detected by the remote monitoring node within a pre-defined geographical region during a predetermined period of time;

Determining whether the occurance data indicates that the at least one content item was detected on a pre-determined broadcast source by the remote monitoring node during a predetermined period of time;

Determining whether the occurance data indicates that the at least one content item was detected on any broadcast source within the same genre format as that being monitored by a particular remote monitoring node during a predetermined period of time;

The length of time between approximately the last detection of the content item on any of the remote monitoring nodes and approximately the current date;

The length of time between the approximate stop date associated with the at least one content item and the approximate current date.

4. The method of Claim 1 where the score is calculated and the database updated on a regular basis.

5. The method of Claim 1 where signature data associated with the at least one content item is removed from the remote monitoring node database in dependency on the second determining step.

6. The method of Claim 5 where the removed signature data is passed to another database by either passing the signature data or by passing a reference to the content rather than the signature data itself or both.

7. A method of managing a database of signature data contained in a system comprised of at least one remote monitoring node and one central node comprising: Determining if at least one content item whose associated signature data is in the database meets a predetermined scoring criteria;

Promoting the signature data associated with said at least one content item.

8. The method of Claim 7 where the promotion step is comprised of any of: transmitting the signature data from the central node to a remote monitoring node, transmitting a reference to such signature data from a central node to a remote monitoring node to or deleting the signature data from the remote monitoring node.

9. The method of Claim 7 where the determining step is comprised of calculating a score as a function of occurancc data associated with the at least one content item.

10. The method of Claim 7 where the determining step is comprised of any one of:

Determining whether the at least one content item has been detected by any remote monitoring node in the system during a predetermined recent predetermined period of time;

Determining whether the at least one content item has been detected by any remote node within a predetermined geographic region during a recent predetermined period of time;

Determining whether the at least one content item has been detected on a predetermined broadcast source during a recent predetermined period of time; Determining whether the at least one signature has been detected on any broadcast source with the same genre format as that being monitored by a particular remote monitoring node during a recent predetermined period of time;

Determining the approximate time period between the last detection of the at least one signature on any of the remote monitoring nodes and the current date;

Determining the approximate stop date associated with the content and the current date.

1 J . In a system for detecting occuranccs of broadcast content comprised of at least one remote monitoring node, a method of determining which signature data associated with at least one content item to store in a remote monitoring node comprising:

For at least one content item, calculating a score based on historical data about the detection of the content item on at least one remote monitoring node.

12. In a system for detecting occurances of broadcast content comprised of at least one remote monitoring node, a method of determining which signature data associated with at least one content item to store in a remote monitoring node comprising:

For at least one content item, determining if the current time is after the start date associated with the content item.

13. In a system for detecting occurances of broadcast content comprised of at least one remote monitoring node, a method of determining which signature data associated with at least one content item to store in a remote monitoring node comprising:

For at least one content item, determining if the content is active.

14. The method of Claim 13 where the determination step is

Determining whether the at least one content item has been detected by any remote monitoring node in the system during a predetermined period of time.

15. The method of Claim 13 where the determination step is comprised of:

Determining whether the at least one content item has been detected by any remote monitoring node within a predetermined geographic region during a predetermined period of time.

16. The method of Claim 13 where the determination step is comprised of:

Determining whether the at least one content item has been detected on a predetermined broadcast source during a predetermined period of time.

17. The method of Claim 13 where the determination step is comprised of:

Determining whether the at least one content item has been detected on any broadcast source with the same genre format as that being monitored by a particular remote monitoring node during a predetermined period of time.

18. The method of Claim 13 where the determination step is comprised of:

Determining the approximate time period between the last detection of the at least one content item on any of the at least one remote monitoring nodes and the current date.

19. The method of Claim 13 where the determination step is comprised of:

Comparing the approximate stop date associated with the at least one content item and the current date.

20. A system for detecting broadcast content comprised of:

A central node containing a first signature database associated with at least one content item;

At least one remote monitoring node operatively connected to the central node, comprised of a second signature database further comprised of a subset of the first signature database;

A subsystem that selects which members of the first signature database are made members of the second signature database.

21. The system of Claim 20 where the selecting subsystem is further comprised of a determination subsystem that determines whether the at least one content item has been detected by any remote monitoring node in the system during a predetermined period of time.

22. The system of Claim 20 where the selecting subsystem is further comprised of a determination subsystem that determines whether the at least one content item has been detected by any remote node within a predetermined geographic region during a predetermined period of time.

23. The system of Claim 20 where the selecting subsystem is further comprised of a determination subsystem that determines whether the at least one content item has been detected on a predetermined broadcast source during a predetermined period of time.

24. The system of Claim 20 where the selecting subsystem is further comprised of a determination subsystem that determines whether the at least one content item has been detected on any broadcast source with the same genre format as that being monitored by a particular remote monitoring node during a predetermined period of time.

25. The system of Claim 20 where the selecting subsystem is further comprised of a determination subsystem that determines the approximate time period between the last detection of the at least one content item on any of the remote monitoring nodes and the current date.

26. The system of Claim 11 where the selecting subsystem is further comprised of a determination subsystem that compares the approximate stop date associated with the at least one content item and the current date.

27. A machine comprising a central processing unit, a digital data transceiver device and a data storage device comprised of any machine readable media, where the machine readable media contains a computer program that when executed by the machine, performs any of the methods of Claims 1 - 19.

28. A machine readable media of any type, which contains data that is a computer program that when executed by a computer, performs any of the methods of Claims 1 - 19.