FR3101451A1 - Method for identifying audio streams from a plurality of sources, system, receiver and program associated with the method - Google Patents

Abstract

Method for identifying audio streams originating from a plurality of sources, system, receiver and program associated with the method The invention relates to a method for identifying audio content transmitted by a plurality of identified audio sources, said sources transmitting an audio stream. having a denomination presented to users. The method is characterized in that it comprises the following steps executed at the level of a server (1): a step of collecting (4.1, 4.2, 4.3) the identifiers of sources and at least one name associated with the content sent. by each of these sources, a step of receiving (5.2) segments of the audio stream emitted by each of these sources using their source identifiers, a step of comparing (5.4) information representative of the audio content of each segment, in order to determining content identities, a step of associating (5.8) identifiers of audio sources transmitting the same content with at least one denomination. In this way, it is possible to identify each stream by its audio content and to make comparisons of this content with others to generate associations between sources emitting the same content. Abstract figure = figure 4

Description

Method for identifying audio streams originating from a plurality of sources, system, receiver and program associated with the method

1. FIELD OF THE INVENTION

The field of the invention is that of the identification of audio content accessible by a receiver capable of receiving content from both a broadcast network and a two-way network. The invention relates more particularly to the fact of comparing the audio signals originating from the various sources and to establishing identifiers according to the identities and similarities between the audio signals.

2. TECHNOLOGICAL BACKGROUND

In the field of the broadcasting of audio or audiovisual content, so-called "hybrid" receivers capable of receiving and reproducing both broadcasts from a broadcasting network, content transmitted from a site Web, and contents transmitted in podcasting or "podcast" according to the Anglo-Saxon terminology. These receivers are equipped with a user interface comprising a screen and a means for entering commands (keyboard, touch screen, voice recognition, etc.), radio reception means and means for transmitting sound signals to loudspeakers. The user sets his device to receive content broadcast on the radio or from a website. Some hybrid devices can enrich the reception of audio content broadcast by an Internet or "IP" type connection, for example by improving the coverage area of the content broadcast by combining their traditional receptions by broadcasting network according to the following broadcasting standards (list non-limiting):

• FM (Frequency Modulated) or AM (amplitude modulation),
• DAB (acronym for Digital Audio Broadcasting"), DAB+,
• DMB,
• DRM, DRM+,
• DVB (acronym for Digital Video Broadcasting"), DVB-T (the "T" stands for "Terrestrial"), DVB-S (the "S" stands for "Satellite"),
• …

with downloads or "streaming" according to the Anglo-Saxon terminology by IP network (Wifi, 3G, 4G and 5G in the near future).

These different sources can provide the same audio content. Indeed, it is common for a national radio station to broadcast FM radio signals from a terrestrial hertzian network, digital signals according to the DAB standard (acronym for "Digital Audio Broadcasting"), and has a website Web accessible by an IP address to transmit the audio content transmitted in real time. A user sets or programs his receiver to receive audio signals from any of these sources to receive the same content.

Databases are also known which list a large number of sources and provide identifiers for receiving such and such a program. These databases receive the names of broadcasting stations or audio programs as input, and provide radio station identifiers with the data allowing them to receive the broadcasts. The information relating to the radio stations they contain can be of a diverse nature: name of each station, genres, identifiers, logos, geographical location, parameters of the transmitters, programs of the day or week, details of the current program, URLs streaming. This service is accessible via the Internet and access to these databases is generally chargeable.

These databases are updated by radio station managers and more generally by audio content providers who have an interest in making themselves known to the public in order to increase their audience. These databases therefore receive information from certified transmitters but do not check whether the content transmitted by the sources identified by the information received does indeed correspond to the audio signals broadcast by these content providers. This lack of verification may lead to errors in the identification of content.

First of all, the different source databases do not necessarily have a common identifier for the same radio, each of these databases can identify the same audio content in several different ways. Take for example, the name of a radio station such as R.T.L., which is an acronym for "Radio Télé Luxembourg", this name can be written either in uppercase: "RTL", or in lowercase "rtl", or separate by dots "R.T.L." ". Although these three spellings identify the radio station well in the public's mind, the computer considers them to be three different identifiers.

The identifier of the same audio source can also depend on the way of receiving it, this identifier can for example be suffixed by "PI" (meaning "Program Identifier" which is an identifier of the RDS system implemented in FM radio ) or by "FM". Similarly, the same radio station can have different URLs in several databases.

Moreover, it happens that, either because of errors during the collection of these URLs, or because of their evolution, a URL associated with a radio station is no longer connected to a stream corresponding to the radio station with which it is associated. Due to computer or human error, a URL may point to another radio station's stream.

Finally, a content provider can transmit information to identify a certain source at a certain time and this is then rendered obsolete. Likewise, it is possible that the information transmitted is incorrect. In general, data consistency issues between databases change over time. Some errors are corrected and others appear following an update for example. It is therefore difficult at a given moment to be sure to enter the correct identifier of a source to obtain the means of receiving the associated audio stream.

3. OBJECTIVES OF THE INVENTION

There is therefore a real need for a method making it possible to associate at least one identifier of an audio source which is known to the users, with at least one means for receiving the audio stream thus identified.

The invention also proposes a server for collecting information on the availability of audio streams and for associating sources transmitting the same audio content with at least one denomination of this content which is known to users.

4. PRESENTATION OF THE INVENTION

In a particular embodiment of the invention, a method is proposed for identifying audio content emitted by a plurality of identified audio sources, said sources emitting an audio stream having a denomination presented to the users. The method is characterized in that it comprises the following steps executed at the level of a server (1):

- a step of collecting (4.1, 4.2, 4.3) source identifiers and at least one name associated with the content emitted by each of these sources,

- a reception step (5.2) of segments of the audio stream emitted by each of these sources using their source identifiers,

- a comparison step (5.4) of information representative of the audio content of each segment, in order to determine content identities,

- a step of association (5.8) of the identifiers of audio sources transmitting the same content with at least one denomination.

Thus, the proposed solution is based on a completely new and inventive approach, making it possible to identify each stream by its audio content, to collect information from various databases and to bring them together to create associations to provide identifiers of source when a user wishes to receive a certain content.

According to a first embodiment, the comparison step consists in comparing signatures calculated on the segments emitted by each of the sources. In this way, the comparison is facilitated.

According to another embodiment, the comparison step begins with a sub-step consisting in that the segments transmitted by the sources transmitting the audio streams having the same denomination are compared with each other, said sub-step triggering a sub-step association of this denomination with the identifiers of audio sources transmitting the same content. In this way, the process first compares the audio contents having the same denomination and if they are identical, the source identifiers are then gathered in the same association, which minimizes the processing time.

According to another embodiment, the comparison step determines a level of similarity between the contents, and in that the association step also associates information representative of the similarity between the contents. In this way, content with a high degree of similarity can be easily collected and offered for reproduction.

According to another embodiment, the comparison step detects redundancies of certain content at determined periodic times, the association step then also associates data signifying the periodic character of the similarities of the content transmitted by two different sources. In this way, users can be informed that certain sources provide identical content except at specific times, which is the case of national and local radio stations.

According to another embodiment, the identification method comprises the following subsequent steps executed at the level of a receiver: transmission to the server of a request comprising a denomination of an audio stream, reception from said server of a message containing at least one identifier of a source transmitting the identified audio stream, presentation of the at least one identifier on an interface for the purpose of selecting the source to receive said stream. In this way, the user is thus informed of the associations and can choose the source that suits him to receive a certain content.

According to another embodiment, the message transmitted to the receiver also contains information representative of the similarity between the contents, and in that the step of presenting each identifier also presents for each identifier said information representative of the similarity between the contents. In this way, the user is informed of the associations and can choose the source that suits him to receive a desired content.

According to another embodiment, different source identifiers are associated with different denominations, these sources transmitting the same audio content. In this way, if the same content is accessible under several denominations, it is easy to find the sources that issue them.

According to another embodiment, the identification method comprises a step of evaluating the validity of at least two denominations associated with the same content transmitted by two different sources, the association step associating the identifiers of these sources different with the denomination having the greatest validity value.

The invention also relates to a server for identifying audio content transmitted by a plurality of identified audio sources, said sources transmitting an audio stream having a denomination presented at the level of user receivers. The server comprises a means for collecting source identifiers and at least one denomination associated with the content emitted by each of these sources, a means for receiving segments of the audio stream emitted by each of these sources using their source identifiers, means for comparing information representative of the audio content of each segment, in order to determine content identities and a memory for recording the associations of the identifiers of audio sources transmitting the same content with at least one denomination.

The invention also relates to a receiver of audio content transmitted by a plurality of identified audio sources, said sources transmitting an audio stream having a denomination presented at the level of user receivers, said receiver comprises means for receiving the audio content transmitted in the stream, means for reproducing said content, means for transmitting a request to a remote server containing a content denomination in order to receive a list of identifiers of audio sources transmitting the content identified by this denomination, display means from the list received and means for selecting one of the sources displayed in order to launch the reproduction of the content emitted by this source.

5. DESCRIPTION OF FIGURES

Other characteristics and advantages of the invention will appear on reading the following description, given by way of indicative and non-limiting example, and the appended drawings, in which:

- FIG. 1 presents the structure of a server for producing associations of content and audio sources, according to a particular embodiment of the invention;

- : FIG. 2 illustrates the main components of a receiver apparatus for audio content, according to an exemplary embodiment;

- : FIG. 3 presents a general diagram of data processing in order to produce associations of content and audio sources;

- : FIG. 4 presents a flowchart of the search steps for collecting information on the audio sources, according to an exemplary embodiment;

- : FIG. 5 represents a flowchart of the processing steps with a view to producing content and source associations according to an exemplary embodiment.

6. DETAILED DESCRIPTION OF AN EMBODIMENT

The invention relates to a method for identifying audio content emitted by a plurality of identified audio sources, said sources emitting an audio stream having a denomination presented to the users. The method is characterized in that it comprises the following steps executed at the level of a server: a step for collecting source identifiers and at least one denomination associated with the content emitted by each of these sources, a step for receiving of segments of the audio stream emitted by each of these sources using their source identifiers, a step of comparing information representative of the audio content of each segment, in order to determine content identities, and a step of associating the identifiers of audio sources emitting the same content with at least one denomination. Thus, the proposed solution is based on an entirely new and inventive approach, making it possible to identify each stream by its audio content and to make comparisons of this content with other content to generate associations between sources emitting the same content.

6.2 Description of a particular embodiment

In all the figures of this document, identical elements (or steps) are designated by the same reference numeral.

We now present, in relation to FIG. 1 , a structure of a server 1, according to a particular embodiment of the invention. This server constitutes a service platform accessible by a network, the Internet for example, to obtain means for receiving a determined audio stream. This audio stream consists of a succession of data packets emitted by a certain source, the payload of the packets containing sound signals to produce a sound at the level of a receiver and make it heard by users. An audio source is identified by an identifier which depends on the type of source, if the signals come from an FM radio station then the identifier can be a carrier frequency, if the signals come from a computer site, the identifier is generally a URL.

The server 1 typically comprises a central unit 2 associated with a program memory 3, a module for receiving a radio stream 4 receiving emissions from a one-way broadcasting network, an FM network for example, and a communication module 5 for two-way communications using a cable. In a very particular case, where the server 1 is directly accessible by an operator, it then has local input/output means 6 for a connection with a terminal and/or a printer. The receiver also has a memory 7 capable of storing data coming from any one of the unidirectional or bidirectional reception means and of storing association data between audio sources and content denominations.

Fig. 2 illustrates the main components of a receiver device 10, according to a particular embodiment of the invention. The receiver 10 typically comprises a central unit 11 associated with a program memory 12, a module for receiving a radio stream 13 receiving transmissions from a one-way network, an FM network for example, and a communication module 14 authorizing two-way communications at short or long range by cable or by radio, using for example the Bluetooth, Wifi and/or GSM network. The receiver also has command input means 15 (keyboard, buttons, touch screen, etc.), and display means 16 (screen, indicator lights, voice synthesis, etc.). These means can be integrated into the device or remote. The device also has a means of reproducing audio content, composed for example of a remote screen, and/or an amplifier 17 integrated into the device emitting acoustic signals to remote speakers 18 According to a non-mandatory improvement, the receiver 10 also has a memory 19 capable of storing audio and/or visual content broadcast by a unidirectional or bidirectional network.

The receiver 10 receives an electrical power supply or has its own battery. It can be mobile and take on the appearance of a smart mobile phone (or "smartphone" according to Anglo-Saxon terminology. It can for example be embedded in a traveling vehicle, the receiver is then a car radio.

The receiver allows you to receive an audio stream from a remote source. The user introduces a denomination of an audio content and the receiver sends a request to the server 1 to receive at least one identifier of a source transmitting the stream of this audio content.

After having detailed the main devices making it possible to implement the method, we will now explain how these cooperate.

Fig. 3 presents a general diagram of data processing in order to produce associations of content and audio sources. A so-called “Data updater” server 30 is responsible for collecting data from a plurality of different sources 31, 32. The data collected associates a certain audio content identifier with a source. The collected data is gathered in a so-called “raw data” memory 33 . At a certain moment, an analysis module 34 is executed in order to read these data, to convert them into audio signals and to transmit them to a comparison module 35 of these audio signals in order to determine the similarities, the differences and the identities.

The analysis 34 and comparison 35 modules produce validated data consisting of a list of associations of content denominations with identifiers of audio sources of different types allowing receivers 10 to receive these audio contents.

We will now detail more specifically the different operations used to produce the validated data.

6.3 Information research phase

According to a first information search phase, which is illustrated by FIG. 4 , the data relating to audio source identifiers will be collected in a so-called “raw” database, in the sense that they are stored without prior processing. The server 1 collects data continuously and cyclically, the duration of each cycle depending on the number of radio stations to be scanned, and the databases to be interrogated.

According to a step 4.1, the server 1 sends requests to public or private databases, with the aim of receiving associations of identifiers and reception means from these databases. The server receives the association data and saves it in memory 7.

According to another step 4.2, the server 1 scans the FM frequencies with a view to discovering the radio stations which are currently broadcasting audio content. The identity of the transmitting station can be determined using the R.D.S. which consists in incorporating the identifier of the station into the radio signals. Another way of proceeding consists in analyzing the audio signals so as to detect jingles which are very often characteristic of the station, these jingles can be only musical, or contain the name of the station. After each scan of the frequencies, the server stores in the memory 7 associations between a transmitting station name and a given frequency.

According to another step 4.3, the server interrogates sites already referenced and which are supposed to transmit audio signals continuously. A short audio segment may be stored in memory for identification verification purposes.

These three steps can take place at the same time since they can use different material resources. At the end of each collection cycle, memory 7 contains raw data which is then used to generate verified lists associating at least one denomination of audio content with identifiers of audio sources (step 4.4).

6.4 Processing phase of the information collected

The processing of the raw data is schematized by the flowchart of FIG. 5 . All of the audio source identifiers recorded in the memory 7 are compared with each other in order to detect identities. In step 5.1, the server searches for the associations having the same audio stream names and the same audio source identifiers. In this case, there is no need to maintain duplicate associations, and the two associations are reduced to one. The server then calculates the signatures of the audio signals emitted by all the sources identified in the raw database (step 5.2), this calculation is carried out globally for all the signals during a certain period of time, or as background process. The calculation can also be performed punctually when the processing needs to calculate the signature of a content. The signature calculation is performed by receiving the audio content transmitted by an identified source (by connecting to an IP address, or by adjusting the radio reception means 4 with the identified frequency) and by using the data received in the stream.

In step 5.3, the server searches for associations comprising the same audio stream denominations and different source identifiers. The server then checks by comparing them whether the audio contents emitted by the sources thus identified are identical (step 5.4). For this, the server uses an audio signal comparison module which is a technique known per se. Patent No. EP 1,741,047 filed by the applicant on March 18, 2005 describes a method using an audio signal signature to determine the similarities, differences and similarities of two audio streams. If at the end of this comparison, the audio content is the same, then we can deduce that it is one and that it is accessible from at least two sources. In this case, the server creates an association between this identifier and the two sources discovered, and saves this association in a dedicated part of the memory 7 (step 5.5). Whenever, in the raw data, the server determines that the same audio content designated by the same identifier is transmitted by different sources, it then associates all the sources with this identifier.

If, on the other hand, the audio contents are not the same, then one of the two denominations may be incorrect. In this case and according to a first strategy, the server searches the stream for information on the denomination of the transmitted audio content. This information can be specific RDS type data sent in a unidirectional stream, or recurring data sent every hour and characteristic of certain audio content. For example, some radio stations play a jingle every time the clock changes that is recognizable and specific to the station. The server may have a library of jingles to better identify audio content. According to a second strategy, the server compares the audio signals of a content with the audio signals of the other contents accessible by the raw data recorded in the memory (step 5.6). If another content is found, then it is possible to rectify the denomination of the first content by using the second. The correction of a denomination is usually based on a confidence index, this index is maximum when the denomination is extracted from a stream (by RDS signals for example). If the denomination is not readable in the signals, the index can be proportional to the number of associations between the same content and the same source identifier. If, for example, the same audio content is accessible from five different sources and it is named four times according to a first name and another time with another, the first name is clearly the most plausible. In this case, and according to a first variant, the server considers that the least plausible denomination is erroneous, it must be deleted and replaced by the most plausible denomination. According to a second variant, the server aggregates all the denominations in a list, which is associated with the list of all the source identifiers. If a receiver sends a reception request to server 1 by transmitting a denomination of a content, it will receive the list of all the source identifiers associated with this denomination. According to an improvement, the confidence index associated with each denomination is also transmitted, in this way the user can find out which denomination is the safest to designate this content.

Step 5.6 also compares the data of the flows whose identifiers have been listed only once, these flows are said to be "unitary". If in step 5.6, the name of a content was found only once in the searches carried out during steps 4.1, 4.2 and 4.3, and this content is identical to that issued by another source under another denomination, then the two denominations as well as the two identifiers of these two sources are agglomerated and referenced in the same association (step 5.5). Take for example the radio station RTL which transmits radio waves on the carrier frequency of 104.3 Mhz and suppose that the content of this radio is also accessible under the name “R.T.L. to URL_1, then server 1 produces the association below and stores it in the dedicated part of memory 7:

[RTL or R.T.L. / 104.3 MHz or URL_1]

If later a receiver requests a source providing content from that radio station, the request may either contain "RTL" or "R.T.L." and the response will contain at least the full set of source IDs: 104.3 MHz or URL_1. In this way, a user of such a receiver can choose which source he wishes to use to receive this content. If he knows that his receiver has a low throughput via his wired communication network, he can choose the radio stream. On the other hand, if he knows that the radio conditions are not good, he can favor the wired network.

If, on the other hand, the analyzed content is not identical to any other, no other association will be produced and the data of this content recorded in the dedicated memory will remain the same as the raw data. It will only be possible to access this content by one means (step 5.7).

At the end of steps 5.5, 5.6 and 5.7, the associations of source identifiers and denominations are collected in the same so-called dedicated database, which is recorded in the memory 7 of the server 1 (step 5.8). In step 5.9, a user can use his receiver 10 to access the server 1 and ask it by a request containing at least one content name at least one identifier of a source to receive this content. The server 1 then reads in its memory 7 and searches for the associations containing the denomination indicated, it then returns the associated identifiers. Receiver 1 presents the data transmitted in a menu displayed on screen 16. This menu displays the list of source identifiers, but also the other denominations designating the same audio content. The user then selects one of the source identifiers displayed, which triggers the connection to this source and the reception at the level of the receiver of the requested content.

Once all the signatures have been compared at least once with another signature, and all the associations have been stored in the dedicated memory, the server restarts for a new processing, using new segments. In this way, the list of associations is constantly updated.

It turns out that many contents are issued by at least two sources, so it is advantageous to take advantage of these identities to group these sources in order to reduce the number of comparisons. The sequence of steps described above makes it possible to reduce the number of contents to be compared and therefore the number of signatures, by grouping the sources which emit the same content. The sequence of steps also makes it possible to verify the accuracy of source identifiers and denominations.

6.5 Determination of similarities between signatures

In some cases, the two-by-two signature comparison modules, which are implemented in steps 5.4 and 5.6, provide three results:

• the signatures are perfectly identical,
• the signatures are partially identical,
• the signatures are completely different.

The signatures are partially identical when, for example, the two contents are provided by a national broadcasting station and by a local station. In such a case, most of the time, the contents are identical but at certain times, the national content is replaced by local content, for example a regional newspaper, or a report on a local activity. According to another example, the station transmits advertising sequences during dedicated time slots, the content of these slots is modified to be replaced by more locally adapted content. The signature comparison module detects a high similarity between the data and can determine that over a large period of time the sound signals are identical.

According to the present invention, if the signatures are perfectly identical, then the contents are identical and the associated sources transmit the same audio content. These sources are associated with the same content name(s).

If the signatures are partially identical, a verification module detects the redundancies between the identities and the content differences. If the differences occur randomly, they may be events due to poor reception and/or the presence of interference, or even the result of chance. This is the case, for example, of two different radio stations broadcasting the same song at almost the same time. In this case, the verification module considers that these are two different contents. If, on the other hand, the redundancies are numerous and the differences appear very regularly and for a constant period from one time to the next, then we can consider that one of the contents is a variation of the other. In this case, the two sources, and possibly the two denominations (if they are different, for example names of stations comprising the name of the broadcasting region) are recorded in the dedicated memory according to the same association. An indication is added to indicate that the contents are identical except at certain times. In this way, a user can be warned that he can receive several local variations of the same national radio station, and can possibly select content that does not correspond to his region.

The verification module analyzes the content over a fairly long period of time, of the order of 12 or 24 hours, preferably starting at a particular time, at 00:00 for example.

According to an improvement, the method uses a module for evaluating the similarities between denominations, this module determines in step 5.4 that the denominations are similar and treats them as if they are identical. This is the case, for example, when the denominations are written one in uppercase and the other in lowercase, or when a spelling error has slipped into one of the denominations.

It should be apparent to those skilled in the art that the present invention permits embodiments in many other specific forms without departing from the scope of the invention as claimed. Accordingly, the present embodiments are to be considered illustrative but may be modified within the scope defined by the scope of the appended claims.

Claims (11)

A method of identifying audio content transmitted by a plurality of identified audio sources, said sources transmitting an audio stream having a denomination presented to users; characterized in that it comprises the following steps executed at the level of a server (1):
- a step of collecting (4.1, 4.2, 4.3) source identifiers and at least one denomination associated with the content emitted by each of these sources,
- a reception step (5.2) of segments of the audio stream emitted by each of these sources using their source identifiers,
- a comparison step (5.4, 5.6) of information representative of the audio content of each segment, in order to determine content identities,
- a step of association (5.8) of the identifiers of audio sources transmitting the same content with at least one denomination. A method of identifying audio content according to claim 1; characterized in that the comparison step (5.4, 5.6) consists in comparing signatures calculated on the segments emitted by each of the sources. Method for identifying audio content according to claim 1 or 2; characterized in that the comparison step begins with a sub-step consisting in that the segments transmitted by the sources transmitting the audio streams having the same denomination are compared with each other, said sub-step triggering an association sub-step (5.5) of this denomination with the identifiers of audio sources transmitting the same content. A method of identifying audio content according to any one of the preceding claims; characterized in that the comparison step (5.4, 5.6) determines a level of similarity between the contents, and in that the association step also associates information representative of the similarity between the contents. A method of identifying audio content according to claim 4; characterized in that the comparison step (5.4, 5.6) detects redundancies of certain contents at determined periodic times, the association step then also associates a datum signifying the periodic character of the similarities of the contents emitted by two different sources . A method of identifying audio content according to any one of the preceding claims; characterized in that it comprises the following subsequent steps executed at the level of a receiver (10):
- transmission (5.9) to the server (1) of a request comprising a denomination of an audio stream,
- reception from said server (1) of a message containing at least one identifier of a source transmitting the identified audio stream,
- presentation of at least one identifier on an interface for the purpose of selecting the source to receive said stream. Method for identifying audio content according to claims 4 and 6; characterized in that the message transmitted to the receiver (10) also contains information representative of the similarity between the contents, and in that the step of presenting each identifier also presents for each identifier said information representative of the similarity between the contents. A method of identifying audio content according to any one of the preceding claims; characterized in that different identifiers of sources are associated with different denominations, these sources transmitting the same audio content. A method of identifying audio content according to any one of the preceding claims; characterized in that it comprises a step of evaluating the validity of at least two denominations associated with the same content emitted by two different sources, the association step associating the identifiers of these different sources with the denomination having the largest validity value. Identification server (1) of audio content transmitted by a plurality of identified audio sources, said sources transmitting an audio stream having a denomination presented at the level of user receivers; characterized in that it comprises means (2, 3, 4, 5) for collecting source identifiers and at least one denomination associated with the content emitted by each of these sources, means for receiving (4, 5) of segments of the audio stream emitted by each of these sources using their source identifiers, means (2, 3) for comparing information representative of the audio content of each segment, in order to determine content identities and a memory (7 ) to record associations of identifiers of audio sources emitting the same content with at least one denomination. Receiver (10) of audio content transmitted by a plurality of identified audio sources, said sources transmitting an audio stream having a denomination presented at the level of user receivers, said receiver comprising means for receiving the audio content transmitted in the stream and means of reproducing said content; characterized in that it further comprises a means for transmitting a request to a remote server containing a content denomination in order to receive a list of identifiers of audio sources transmitting the content identified by this denomination, a display means from the list received and means for selecting one of the sources displayed in order to launch the reproduction of the content emitted by this source.