WO2009083635A1

WO2009083635A1 - Method and apparatus for protecting non real time content

Info

Publication number: WO2009083635A1
Application number: PCT/FI2008/000147
Authority: WO
Inventors: Ville Ollikainen
Original assignee: Envault Corporation Oy
Priority date: 2007-12-28
Filing date: 2008-12-29
Publication date: 2009-07-09
Also published as: FI20071024A0

Abstract

A drawback of applying DRM to a filecast program resides in that it is difficult to find out a user who has stored a program and then distributed illegal copies thereof. This can be prevented by extracting portions of data from a stream of a digital media object and then mass distributing the remaining media object. The extracted portions are fingerprinted and then transmitting via a point-to-point channel to a user terminal, which inserts the fingerprinted portions to the media object received from a mass distribution channel. Viewing experience of the file-cast program becomes disturbed in the absence of the extracted portions.

Description

Method and Apparatus for Protecting Non Real Time Content

Field of the invention

The invention relates in general to the delivery of digital media objects, such as audio and video, via a mass distribution network to terminals and in particular to how the protection against unauthorized reception and copying can be implemented in a uniform manner as regards filecasting. Particularly the invention relates to transmission of digital content via a hybrid network to a mobile terminal capable of storing media objects into its memory device for future playback.

Background of the invention

In this description, all digital sound and image recordings and transmissions that are handled as one entity will be called "digital media objects". Thus, an object may be a picture, sound effect, portion of music, film, animated program, radio program, multimedia program or other corresponding entity which can be transmitted, stored and reproduced to a user as such and/or together with other corresponding objects. Computer programs that are broadcast in file format can also be considered objects. By filecasting it is meant broadcasting objects to terminals for storing the objects persistently. By transmission it is meant not only filecasting over a broadcast network, where a transmitting station electrically distributes objects to a great number of receivers to be stored to the local memory of receiving equipment, but also other means of distributing files to a number of receivers simultaneously, such as multicasting in IP (Internet Protocol) networks. Also a peer-to- peer network may be considered as a filecast network, in which case there are intermittent storages in the distribution. By making a stream it is meant conversion of a digital media object or a part of it into a stream of successive data elements supplied over time. The elements may be packets or frames, for example.

Future portable terminals will be equipped with high-resolution displays, high capacity memory devices and improved audio and video capabilities. Accordingly, future services require that the amount of data to be transmitted is far higher than needed for currently used services for handheld terminals. Although the mobile network of the third generation (UMTS) is more powerful than the network of the second generation (GSM/GPRS), it still encounters limitations in the transmission of larger files or streams having audio/video content to a number of simultaneous users. Thus, it appears that the capability of existing communication systems is not satisfactory for new services.

An example of above-mentioned services is mobile digital television that is expected to ^' become the next high growth consumer technology. Mobile TV is intended for a number of different devices and is not restricted to television on mobile phones. Naturally a mobile television receiver can be a mobile phone, but it may also be a dedicated terminal placed in a car, or a laptop computer or a PDA.

Although 3G operators offer today video on mobile phones in the form of video images (music videos, animated sequences, programme excerpts), which can be downloaded or viewed in streaming on a mobile phone, such video is not mobile television in the strict sense because mobile television involves accessing continuously broadcast television channels by means of a portable receiver. By using a hybrid network consisting of both a mobile and a broadcast network (for example UMTS/DVB-H) capacity limitations of 3G networks could be overcome. While video on mobile terminals may make use of the 3G or 2G telecommunications network on the basis of a point-to-point bidirectional connection, mobile digital television uses the same broadcasting network as fixed television, which is dedicated to "point-to-multipoint" transmissions. This is because, in order to broadcast continuously the same programme to a wide audience, it seems at present to be more competitive to cover an area with a television transmitter than to make use of individuali connections via the telecommunications network. A television transmitter can cover large areas without gaps or disturbances in viewing experience even if the user is moving.

Many multimedia services will be asymmetric and interactive, wherein users receive large amounts of data. Therefore, broadcast systems would be better prepared to serve these users. However, broadcast networks provide broadband access serving many users simultaneously but they do not support interactivity. The combination of unicast (bidirectional point-to-point) and broadcast (unidirectional point-to-multipoint) networks to form a hybrid network could provide both broadband accesses to many users and also individual interactive channels. Such hybrid mobile communication networks, including unicast and broadcast delivery systems, may consist of UMTS (Universal Mobile Telecommunications System) and DVB-H (Digital Video Broadcasting - Handheld), respectively. DVB-H is an open standard based on broadcast system DVB-T (Digital Video Broadcasting - Terres- ^' trial). It is optimized for small devices such as mobile phones by using time slicing for less power consumption and an additional forward error correction. Other mobile digital TV transmission standards are also used in addition to DVB-H. MediaFLO (Media Forward

Link Only) is a proprietary standard developed by Qualcomm of the US. T-DMB, which is

■ an updated version of the DAB (Digital Audio Broadcast) system, is currently used in

South Korea. DAB-IP used in United Kingdom carries broadcast signals over DAB IP encapsulation. Other standards may also emerge; for example China's broadcast industry regulator has announced that it will require mobile phone service providers in China to use a domestic standard for broadcasting television signals to mobile phones.

• Broadcast services have a benefit of being cost efficient in delivering large amount of content to large audiences simultaneously. Whereas communications capacity required for point-to-point streaming services increase in a linear manner when the amount of users increase, the capacity required for point-to-multipoint broadcast services is not depending on the amount of users. This makes broadcast more scalable.

^• Broadcast, on the other hand, may not be considered suitable for mobile use: In contrast to an average television viewer who may want to adjust his living to match ordinary TV program schedules, a user will watch mobile TV when there is consumable time and a suitable place available. Mobile users have thus a tendency of using the mobile terminal in a sporadic manner. Excluding some special occasions, such as live events, a mobile user is not in general expected to wait for a specific programme to begin. Suitable content has

^• to be available whenever the terminal is turned on, which is contradicting the way broadcasted programmes are delivered in real time.

In addition to broadcast and streaming, media objects can also be filecasted to the termi- nal. Filecast is typically using broadcast communication to transfer files to many users simultaneously. In the case the files would be digital media objects, they are not supposed

^• to be played back in real time, but to be stored to a local memory of the mobile terminal for future use. The media object can be used at any time later on, being thus more suitable for sporadic mobile use. Additionally, as the delivery takes place in point-to-multipoint manner, the scalability of broadcast services is maintained. There are IP (Internet Protocol) based standardized protocols, such as FLUTE (File Delivery over. Unidirectional Transport, RFC 3926) which can be used for filecast, and candidates for standards, such

^• as FCAST. Some companies, such as Digital Fountain, Inc. pecialize in filecast technologies. As far as a conceptual framework is concerned, filecast relates to other mass distribution methods, such as peer-to-peer networks. In a wide sense, any mass distribution method delivering files can be considered to be filecast.

It should be noted that while a large number of files may be transmitted over a filecast enabled delivery network, only a small portion of them may be of interest to an individual user. Storing unnecessary files to the local memory of the terminal would not only consume storage space but also increase energy consumption and reduce battery life in wireless devices. This is why there are built-in methods in prior art filecast technologies for ordering particular files of groups of files.

Most above-mentioned broadcast networks can carry IP (Internet Protocol) data. Some of them, such as DVB-H are completely based on IP, although the underlying network handles Transport Stream datagrams. Therefore, in DVB and DVB-H any kind of digital con- tent may be encapsulated in DVB datagram (or any other broadcast standard datagram) as IP packets, which are the same format used to transfer digital content on the Internet. The combination of a broadcast radio transmission technology with the Internet Protocol (IP) is called IP Datacast. IP Datacast architecture over DVB-H networks is well presented in ETSI Technical Report TR 102 469. The IP layer common with the Internet and the broadcast network means that applications, content format and middleware technologies developed for the Internet can be applied in IPDC terminals as well. This makes DVB especially suitable for filecast.

FIG. 1 an architectural model for a hybrid network. A broadcast operator maintains and manages a broadcast TV network that is DVB-T/H in this example, whereas a mobile operator maintains and manages a cellular network comprising of a core network and a radio access network. The cellular network may be a 2G network like a GSM/GPRS network or a 3G network like UMTS. A service platform refers generally to broadcast service providers and internet service providers which provide various contents to be played out into the broadcast channel. The broadcast operator transmits content data (for example games, video and audio files or computer programs) simultaneously via transmitters to multiple mobile television receivers that are mobile phones, dedicated terminals, laptop computers etc. Depending on the structure of the hybrid network there may be a number of cells of the mobile network coinciding with just one cell of the hybrid network. A subsystem, a mediation platform, ensures control and signalling between the parties involved in the provision of the services on the hybrid network.

A broadcast program terminal is provided with an electronic service guide (ESG) for de- tecting broadcast programs being currently broadcast and displaying broadcast program information, using ESG data. ESG may also contain information about digital media objects available in datacast services, such as filecast files.

Irrespective of underlying mobile TV transmission standards users will watch their favour- ite programmes. However, several studies indicate that a session of watching mobile digital television is short, typically between 5 and 15 minutes. This is quite contradictory when compared to a program length in ordinary television. However, program brands in ordinary television are strong and probably people would like to see them also in their mobile TV, although they in general represent a longer format. Doing a separate short format production of a program brand for mobile TV is expensive and not feasible as long as the number of viewers remains low. Sending the programmes as filecasted media objects and storing the received media objects into terminal memory allows viewing to take place in arbitrary small chunks.

From a content provider's angle there are some problems concerning mobile TV. One problem relates to making a profit from a broadcast transmission whereas another problem relates to copyright, i.e. how to prevent mobile TV users from spreading copies of broadcast transmissions such as movies.

Since service providers will not want to give content away for free, a straightforward solution to the first problem is to encrypt the transmission and acquire subscribers to an encrypted digital channel wherein a subscriber must purchase and install a decryption module or software into the mobile TV device, and in addition, pay a monthly subscription fee. That kind of solution is well known conditional access (CA) technology. There are numer- ous digital video broadcast (D VB)-com pliant CA systems available for a broadcaster to choose from.

A drawback of applying conventional CA for broadcasts resides in the nature of protecting the transmissioins by using plain encryption Because of encryption, none of the protected content can be viewed by the user, neither can decryption keys be disclosed for a user wishing to briefly peek the content, A solution to the other problem is known as Digital Rights Management (DRM), which refers to technologies used by publishers to control access to and usage of digital data. DRM schemes are built on numerous technologies, such as using cryptographic, restricting owners' use of purchased content, using product activation or certificate-based en- cryption, or applying digital watermarking or fingerprinting i.e. placing hidden data on the media.

At its simplest, fingerprinting can be made by changing one bit in Discrete Cosine Transform (DCT) blocks of compressed images in such a way that the particular bit change can be monitored later on. Monitoring the bit change can be done by saving the information of all changed bit positions for tracking purposes. Each user in this simple fingerprinting method has a unique pattern, according to which the bit is changed or not. When detecting the fingerprint, bit changes in the tracked positions are monitored over a period of time and the unique pattern is detected, thus disclosing the particular user.

A drawback of applying DRM to a broadcast program resides just in the nature of the broadcast; for example, a broadcast movie is intended for all users or a group of users having subscription to the program. Thus, it is almost impossible to find out a user who has stored a movie on the hard disk of the mobile TV and then distributed illegal copies thereof.

Therefore, an objective of the present invention is to provide a filecast compatible method which enables mobile TV viewers to pay for a program either prior to, or at the beginning of, or in the event of the program by using a bi-directional connection to the operator's filecast server. The method should be applicable also to pay-per-view (PPV) applications.

.A further objective is to provide means for users to have controlled peek to transmitted content.

A further objective is to provide a filecast server for handling a digital media object to be protected against unauthorized use in such a way that the user can reproduce the media object with full quality only after subscription or registration.

A still further objective is to provide a filecast server capable of adding user-specific fin- gerprints to filecasted digital media content. A still another objective is to devise a mobile TV terminal capable of receiving the digital . media object as filecast and storing it persistently for later consumption, reproducing it with full quality only after subscription or registration.

Summary of the invention

The objective is achieved by a method in which such portions of data from a stream of an

. original digital media object are selected by a filecast server, the absence of which will prevent flawless reproduction. Then these selected portions of data are extracted from the digital media object wherein an incomplete digital media object is formed. The incomplete digital media object is delivered via a mass distribution channel to terminals as filecasts and further stored into a local memory device of terminals, which have ordered the media object. The extracted portions are stored on a memory medium and conveyed via a bidi-

. rectional communication channel to a terminal requested of having to get the complete digital media object.

The mass distribution channel could be any channel that is suitable for distributing digital media objects to public as an electronic signal. In the preferred embodiment the mass distribution channel is a terrestrial broadcast digital TV channel and the bidirectional . communication channel is a mobile network channel, wherein the incomplete digital media object is converted into a filecast format for a broadcast transmission and the extracted portions of data are transmitted via a mobile network.

The terminal receives the incomplete filecasted digital media object and stores it onto its local memory for later use.

Because a terminal is not able to reproduce the incomplete digital media object it has received from a mass distribution channel, at a time of launching the reproduction of the digital media object a user sends via a bidirectional point-to-point channel a request of having to get the extracted portions. In response to the request the extracted portions are sent along the bidirectional channel to the terminal, which then inserts the extracted por- • tions into the incomplete digital media object and so reproduces the complete digital media object.

Preferably the extracted portions are modified to contain digital fingerprints prior to transmission to the terminal. The connection or user specific fingerprints are stored in a mem- ory, wherein a terminal that is the origin of unauthorized copies of the digital media object is traceable.

According to the method of the present invention, a terminal receives from a mass distri- bution channel a streaming incomplete digital media object. The terminal examines the incomplete digital media object for finding a position where a portion of data has been extracted from the complete digital media object and, upon detecting such a position, sends a request via bidirectional communication network to the filecast server that responds by sending the extracted portion of data via the same communication network. The terminal inserts the extracted portion into the position of the missing portion of data so regenerating the complete digital media object.

A filecast server of the invention comprises means for reading a digital media object or part of it as a stream, an analyzer adapted to receive the stream of the digital media object and to analyze the stream for finding a predefined portion of data thereof, and responsive to said portion of data to produce a selection signal, an extractor for extracting the predefined portion of data from the stream responsive to the selection signal, wherein an incomplete digital media object stream is constructed, a mass distribution data interface adapted to transmit the incomplete digital media object to the filecast enabled mass distribution system for further delivering to terminals, a packetizing block operatively connected with the extractor for receiving the ex- tracted portion of data therefrom, and adapted to create a unicast packet for conveying the extracted portion of data, a unicast data interface adapted to receive through the bidirectional communication system a request for the extracted portion of data and responsive to the request to transmit the unicast packet addressed to the terminal through the ^• bidirectional communication system.

In the preferred embodiment the filecast server further comprises a fingerprinting unit for adding a mobile TV terminal specific fingerprint to the extracted portion of data.

Optionally the filecast server further comprises a replacement forming block operatively ^• connected to the extractor and adapted to form a replacement of the extracted portion of data, and an insertion block adapted to receive the replacement and to insert it in place of the extracted portion of data, wherein the incomplete digital media object stream is formally conformed with the complete digital media object stream.

A terminal of the invention comprises a memory device for storing those filecasted media objects, which have been ordered by the user, a missing data detector operatively connected to the network receiver of the terminal for a receiving a streaming digital media object therefrom, and for produc- ing a notification signal in response to detection of a position of an extracted portion of data, a requester having access to a bidirectional communication network, for transmitting a request for the extracted portion of data to a filecast server in response to the notification signal, an insertion block for inserting the extracted portion of data received from the file- cast server through a bidirectional communication network into the position detected, wherein the complete digital media object is formed.

The missing data detector may detect missing portions of data by keeping track of the consecutive numbers of the packets received, wherein a missing packet number triggers the notification signal. Alternately, the missing packet detector may identify predefined replacement packets, wherein the replacement packet identified triggers the notification signal.

The detector is preferably launched at the time of launching the reproduction of the digital media object. When launched, it detects all missing data in the incomplete digital media object as quickly as it can and produces notification signals for the whole digital media

• object. Having accomplished this task the requester requests all missing data and saves it onto a local memory of the terminal for the real time insertion.

The proposed methods and devices are applicable for networks where data is transmitted in packets, wherein the digital media content arriving to the filecast server is streamed in packets and the portions of data to be extracted and transmitted via the mobile network

• are packets. The mass distribution network may be a filecast enabled broadcast network, for example a digital television network or a mobile TV network, but also any IP network or physical media. Further, a peer-to-peer network within an IP network or the IP network may be used for mass distribution.

For bidirectional communication any network offering bidirectional packet transmission is usable, such as the Internet and a mobile digital network.

Brief description of the drawings

In the drawings,

Fig. 1 depicts an architectural model for a hybrid network,

Fig. 2 illustrates the principle of the invention,

Fig. 3 is a flow chart of method steps of an embodiment of the invention, Fig. 4 depicts functional blocks of the filecast server,

Fig. 5 depicts functional blocks of a mobile TV terminal,

Fig. 6 illustrates proposed encryption method. .

Detailed description of the invention

FIG. 2 illustrates the general principle of the invention. Extracting device 21 in a filecast server removes selected fragments from a digital media object that can be a movie, for instance. Prior to extracting the object must be in or it must be transformed into a digital data stream. Removing takes place more or less periodically as the stream is progressing and extracted fragments are stored on storage medium 22 that can be a volatile or nonvolatile memory. Due to a small amount of data that is removed from the digital media object a need for the memory size is low. Extracted fragments are only a minor part of the whole data, typically a few percents only. Major part of the data of the digital media objects adapted in unit 23 to a suitable form for a mass distribution channel 24. This part may be called as an incomplete digital media object.

Preferably the form is an electrical signal converted to a suitable form for broadcasting in a digital TV network, wherein the one-directional mass distribution channel is a digital TV channel. Thus, any terminal 25, hereinafter referred as a terminal, capable of receiving filecast transmissions can receive, store, decode, and present the digital media object for a user. But because the digital media received is incomplete, quality of the presentation even at its best is low. It may be high enough to get a reasonable conception of the content of the media but low enough to make watching unpleasant to the user.

The terminal 25 is provided with a unit for wireless bidirectional communication and if a user wants to watch presentation of the digital media object, stored to a local memory 26, with high quality, the user sends a request of having extracted fragments to the filecast server via a communication network that may be a mobile network, for example. In response to the request a stream of the extracted fragments is sent via the bi-directional communication channel 27 to the terminal that then inserts the fragments into their original location. Thereafter the complete digital media content can be presented with high quality.

FIG. 3 illustrated method steps of an embodiment of the invention. Assuming a filecast server receives from a broadcast service provider (see Fig. 1) a digital media object for further delivering for mass distribution, the filecast server first selects portions of data to be removed from the stream; step 301. To make the delivery more secure, the data may be pre-processed 315 for instance by encrypting it using CBC mode block decryption on a frame basis. Selection is based on an analysis of the content of data received. The purpose is to extract data from the data stream only such an amount that although a user may be able to some extent listen or watch the presentation of the media object in a ter- minal, perception is unpleasant enough to encourage the user to pay for a high quality presentation. Thus, as to the music it might be enough to extract data representing a certain bandwidth; low, medium, or high tones, for instance. However, the extracted data remarkably lowers quality of the reproduced media object but does not make it impossible to listen or watch it. Proper selection of the analysis criterion allows adjustment of quality of the reproduced media object to a desired level.

It is advantageous if the data stream to be analyzed has some clear structure. For instance, MPEG-2 an elementary video stream is simply a continuous set of video frames. It is known from the art that the elementary stream is split into packets in order to make the multiplexing process easier at a broadcaster¹ site. At the start of each packet is a package identifier (PID) informing whether a packet belongs to a video stream or to an audio stream. At the broadcaster's site the resulting packetized elementary stream is packetized again by storing the data from the elementary streams in transport packets. Each transport packet has a length of 188 bytes which is smaller that the length of a packet of the packetized elementary stream. Keeping in mind the explanation above, it is clear that the analysis can be applied either the stream prior to its packetizing or to packets. In the first case and if the media object is video then removing now and then at least a part of a P-frame makes it impossible in a receiver to flawlessly decode subsequent frames until the next l-frame arrives. In this • sense the first P-frame in GOP structure can be considered most vulnerable. An amount of data in a P-frame is small thus resulting a very small amount of data to be extracted in comparison to the amount of whole data of the movie. Removing packets that include a whole P-frame makes it impossible to decode subsequent frames in the receiver.

Referring back to FIG. 3, applying the selected criterion in the analysis results in the discovery of a portion of data to extract, step 302. If the digital media object is packetized then each packet is analysed in real time as the packet arrives, wherein successive extractions result in a minor stream of extracted packets whose total amount may be around 2 % of the original stream, for example, whereas the rest of the original packet stream, which is hereafter denoted as an incomplete digital media object, is consequently about 98 % of the original packet stream. This proportion is an example, and it should be noted that even a tiny fraction missing from the original packet stream may cause adequate disturbances to motivate users to pay for the full quality, The extracted packets are stored temporarily on a storage medium. The incomplete digital media object is directed to a broadcaster's system for packetizing in transport packets to insert into a filecast enabled broadcasting multiplex and then for mass distributing via a broadcast channel, step 304.

Optionally, the extracted packets can be substituted with replacement packets, step 303. Each replacement packet has the same payload or an empty payload but a replacement packet gets at least the same packet number as the extracted packet in question has. Thus, the incomplete digital media object comprises the same number of packets as the original digital media object. The replacement packet may include a unique identifier referring to the particular piece of extracted data,

It is worth noting that the incomplete digital media object, which display apparatuses are receiving through a broadcast channel, step 305, is not necessarily encrypted. The files are stored to a local memory of the terminal, step 316. Therefore, any terminal can decode packets and present the object, step 318. However, pursuant to the extracted packets quality of presentation is low and far from being enjoyable to a watcher. Anyhow, ^' quality may be high enough so that the user can get a reasonable conception of the content. If the program is interesting, the viewer can ask the filecast server to "open" the program. This can be done so that the user selects a "subscribe" button on a menu, for example. In response to the selection, the terminal automatically creates and sends via a mobile network a message addressed to the filecast server, step 306. A person skilled in the art may also implement a partial subscription for providing only a short duration of full quality viewing.

In receipt of the message the filecast server authorizes the terminal to decide if the user or the terminal is allowed to have the digital media object with full quality, step 309. Any known authentication process is applicable and subscriber information stored in a sub- scriber database may be used.

After the user's request has been accepted, a bidirectional connection from the filecast server via the mobile network to the terminal is established. The mobile network sets up the connection and receives from the filecast server packets that has just extracted from the digital media object. When forming a unicast packet, step 311, the target address is the same as the address of the terminal in the mobile network and is obtained from the address field of the request message. Henceforth, every time as a packet is extracted from the stream of packets forming the digital media object, it is inserted into a unicast packet that is transferred via the mobile network to the terminal, step 312.

The terminal receives the unicast packet from the mobile network, step 313, discovers the extracted packet therefrom and inserts said packet into the stream of packets of the incomplete digital media object obtained from the broadcast network. Following the packet numbering the extracted packet is inserted to its original location in the stream of packets, step 314.

In summary, the terminal is receiving a major flow of packets of the digital media object from the broadcast network and a minor flow comprising of extracted packets from the mobile network. Upon receipt an extracted packet it is inserted to its original location wherein a complete packet stream of the digital media object is gained. If the data was pre-processed, step 316, the inverse procedure has to be done as post-processing, step 317. In the example of encrypting the data using CBC mode decryption on frame basis, the inverse process is decrypting the data using CBC mode encryption on frame basis. As a result, the media object can be presented in full quality on the terminal, step 318. Using CBC mode decryption for encryption and CBC mode encryption for decryption is explained in Fig 6. A person skilled in the art recognises that plaintext 601 is converted to ciphertext 602 using block cascaded block ciphering, but the processing order in encryption 60 is the same as usually implemented in CBC block mode decryption. Respectively, ciphertext 603 is turned back to plaintext by using CBC block mode familiar to encryption. The motivation of doing this is that all errors caused by extracted portions propagate to the end of the plaintext when decoding the content without extracted packets inserted back to their original positions.

Because the digital media content is a broadcast transmission a great number of terminals may simultaneously receive the same digital media content but extracted packets are transmitted individually via a bidirectional channel of the mobile network only to those terminals which have authorized to receive these packet. There is a serious risk that an authorized user may store the high quality digital media object, such as movie, on the hard disk of the terminal and later distribute illegal copies of the movie. Tracking a source of the illegal copies is impossible without using an additional feature of an embodiment of the invention.

This feature is based on the insight to fingerprint the extracted packets individually prior to submitting to a terminal. Thus, after authentication of a user having requested extracted packets, it is created a unique fingerprint, i.e. an individual fingerprint is assigned to each user, step 310. Information relating to the fingerprint and the digital media object is stored in the subscriber database for that user.

Any fingerprinting technique known in the art can be used. One option is to use a limited number of watermarks or fingerprints, a so-called fingerprint pool. Each packet in a defined number of successive extracted packets in a bidirectional connection gets an individual fingerprint wherein extracted packets travelling via a bidirectional connection are distinguishable from packets travelling via another bidirectional connection by different combinations of subsequent fingerprints.

The fingerprints are quite transparent to a terminal that has received the extracted packets. Then, if illegal copies of a movie are found, an origin of the copies can easily found by comparing fingerprints in the copies with the fingerprints stored in the subscriber data- base. It may be enough to fingerprint one single packet only because it is sufficient to find one fingerprint only to proof the source of a copy. As stated previously, there is an option to substitute the extracted packets with replacement packets, step 303. If that step is taken then the replacement packets must be removed in a terminal, step 307. A replacement packet to be removed may be recognized from a predefined constant bit pattern in the payload or a predefined token in a header field.

It is also feasible to attach the information of the original location to the replacement packets themselves. Insertion is then based on this information, with not requirement of disclosing the positions in the filecast data.

Then, after the terminal has sent the request for extracted packets, it starts monitor the packet stream arriving from the broadcast channel for detecting a missing packet. The location of an extracted packet is detected by monitoring either replacement packets if the replacement option is used or by monitoring packet numbers for a missing packet number, step 308. The missing packet identifier may also be inserted into the replacement packet data.

There are two alternatives to proceed. Either every time as a missing packet is detected in step 308, a request for the extracted packet is sent to the filecast server. Or the first re- quest causes that the filecast server will send extracted packets automatically as a continuous stream as long as the mobile terminal sends a "close" message to the filecast server. Such a message is sent automatically in response to a channel change or shutting down the terminal, for example.

Every time when a packet has been received from the bidirectional channel, it will be inserted to its original location in the stream. Then the digital media object is presented in its original quality on the display.

Reference is now made to FIG. 4, which illustrates functional blocks of a filecast server. Filecast server 40 operates in conjunction with a mobile operator's network 42 and a broadcast operator's network 41, using in communication an intermediate computer network such as the Internet. Optionally, the filecast server can also be incorporated into the broadcast operator's network. The filecast server can communicate via a bi-directional connection with mobile TV device 400 that is provided with a wireless network unit allow- ing access to the mobile network. Filecast server 40 receives various digital media objects from content provides 43, which first store the objects in content servers 44 or on other delivery media.

A digital media object, which is converted into the form of a packet stream, is first ana- lysed in packet analyser and selector 401. The purpose is to find from each packet a predefined data fragment. For instance, the data packet is analysed for detecting P -frame data thereof. When a packet including such a frame has been found, a packet selection signal is generated. Responsive to the packet selection signal the packet extractor 402 removes from the packet stream the selected packet that is temporarily stored in memory means 402 .

An amount of the extracted packets in comparison to the total amount of the streamed digital media object is adjusted by proper selection of the analysis criterion to be very low, only a few percents. The major part of the stream, hereafter the incomplete digital media object, is transmitted through filecast data interface 403 to the filecast enabled broadcast system 41 for further filecasting to receivers, such as mobile TV terminals. Filecast packets of the incomplete digital media object are put into transport packets, which are broadcast in the multiplex. Steps taken at the broadcast operator's site depend on the broadcast system in question and are outside of the scope of the present invention. The opera- tor may use DVB-H system, for example.

In some implementations it might be advantageous to replace the extracted packets with replacement packets. To that end, the filecast server includes optionally replacement- packet forming block 404. Responsive to the packet selection signal from packet analyzer 401 the replacement-packet forming block 404 creates a replacement packet including the same header as the extracted packet has or at least the packet number remains. At least a part of the payload of the replacement packet is a predefined bit pattern and every replacement packet has this repeating part of payload. Therefore, replacement packets are easily detectable for a packet stream. Rest of the replacement packet may contain indexing information identifying the extracted portion. Replacement packet insertion block 405 inserts the replacement packet into the location of the extracted packet in the stream wherein the packet stream of the incomplete digital media object has same number of packets as the original packet stream has. Then the incomplete digital media object including replacement packets are transmitted to the broadcast operator' site. The filecast server 40 has also unicast data interface 406 for communicating with mobile network 42. In practice, the intermediate network between the filecast server and the broadcast operator' site and the mobile network is a computer network, such as the Internet, wherein interfaces 403 and 406 may be merged into a single network interface. How- 5 ever, for clarity reasons the interfaces are handled here separately.

Through the unicast data interface 406 the filecast server receives from the mobile network a message carrying a request of having extracted portions. The origin of the message is mobile TV device 400 that has sent it to the mobile network for further delivering to

10 the filecast server. In response to the request the filecast server first authenticates the user of the mobile TV device. Any known authentication method can be used and possibly several queries and replies are exchanged between the filecast server and the mobile TV depending on whether the user is a new subscriber or an old one. In the latter case subscriber information is already stored in subscriber information database 407. When the

15 authentication is completed, information about subscriber and a digital media object requested by the subscriber is stored in the database. That information with additional information is used for billing purposes.

Next, unicast packetizing block 408 starts to generate packets for the mobile TV, each

20 including in its payload an extracted packet fetched from temporary memory 402. The target address of each unicast packet is obtained from the request message initiating the user authentication. As a unicast message is ready it is transmitted through the unicast data interface to the mobile network 42 that transports the packet further to the mobile TV device 400, which then inserts the extracted packet into its original location in the packet

25 stream received from the broadcast network.

From now on, whenever packet extractor 402 removes a packet from the original packet stream, the packet may be persistently stored 412, after which the unicast packetizing block 408 incorporates the extracted packet into a unicast packet, which is transported 30 immediately through the mobile network to the mobile TV device 400. It worth noting, that unicast transmission of the extracted packets is synchronized with broadcast transmission wherein the mobile TV device is able to present the digital multimedia object in real-time.

Optionally the filecast server includes a fingerprinting unit 410. The extracted packets are

3.5 transmitted at the same time perhaps to hundreds or thousands of mobile TV devices.

Digital technology makes it possible for a user to record the digital media object in full quality and then spread unlawful copies of other copyrighted work. Fingerprinting the digital media object can effectively prevent distributing of said copies. Moreover, fingerprinting each digital media object with unique, subscriber specific fingerprint or watermark offers a powerful means to track the source of unlawful copies.

To that end, some or all extracted packets, more accurately their payloads may be fingerprinted prior transmission. Advantageously fingerprints applied to the packets are connection-specific, i.e. each mobile TV receiving the extracted packets through a bidirectional connection also receives connection-specific fingerprints therein. Thus, upon completed authorization of a user but prior to transmission of the extracted packets the filecast server selects from fingerprint storage 411 a fingerprint or a combination of fingerprints to be applied only to those extracted packets that are sent to the user. Information of the applied fingerprints is stored in subscriber information database 407 in connection with the previously mentioned information about the subscriber and the digital multimedia content.

As a result of connection-specific fingerprinting, a specific user whose mobile TV device is the source of unlawful copies can always be identified by comparing fingerprints in a copy with the fingerprints stored in database 407.

FIG. 5 illustrates additional unit 50 of a terminal of an embodiment of the invention to present a digital media object. The unit operates in conjunction with receiver unit 501 adapted to receive filecast over broadcast transmissions, for instance DVB-H transmissions, and with wireless radio unit 502 enabling access to a network that offers bidirectional bidirectional communication channel. Preferably, the network is a mobile net- work enabling point-to-point packet transmission. Thus, the network could be a 2.5G network like a GSM/GPRS or a 3G network but also a Wi-Fi network where wi-fi refers to the underlying technology of wireless local area networks (WLAN).

Receiver unit 501 receives a packet stream of an incomplete digital media object from a broadcast channel, said packets being embedded in transport packets of the broadcast system considered. Receiver unit decodes transmission packets and forwards resulting packet stream of the incomplete media object to the unit 50. The incomplete digital media object is stored into a local memory of the terminal 505. Upon playout of the digital media object, missing packet detector 503 examines every packet in order to find those locations in the packet stream where packets have been extracted. Missing packet can be detected in two ways. The simplest way is to monitor packet numbers of successive packets, wherein always when missing packet detector 503 detects that a number in the number order is missing it produces a detection signal. The signal included at least the number of the missing packet.

Optionally, the missing packet detector may detect replacement packets if such packets are used to substitute the extracted packet at the transmission end. A replacement packet has the same packet number as the original packet, so detection based on the number order is not usable. But a replacement packet may have a special tag in its header field, wherein detection of said tag produces the detection signal. Alternatively, each replace- ment packet may have the same predefined payload pattern wherein upon detection of such a pattern the detection signal is produced. A part of the replacement packet payload may contain unique index referring to the particular data extracted. In all cases the signal includes at least the number of the missing packet.

In case replacement packets are used, replacement packet remover 509 removes such a packet from the packet stream prior to guiding the packet stream to a buffer 510 for pres- entation 511.

The detection signal that missing packet detector 503 produces is an input signal to miss- ing packet requester 506. Responsive to the detection signal the requester forms a request message addressed to the filecast server. The message contains information that should be sufficient for the filecast server to identify the digital media content. For that purpose information included into a package identifier (PID) at the start of each transport packet may be used. In addition, the number of the missing packet is told in the request message. After the content of the request message has completed the unit for wireless bidirectional communication 502 sends it to the filecast server.

The missing packets received from the bidirectional communication unit 502 may be buffered to wait for playout an addition or instead or buffering the complete packet stream in 510.

According to the mechanism describe above, the filecast server sends each extracted packet via the bi-directional communication channel whereupon the packet is inserted to its original location. The result is a the complete digital media object, i.e. a complete packet stream that is then transferred for further processing and is finally presented for the user. It is stated in the description above that every time as the missing packet detector 503 has detected that a packet has been extracted, a request for the extracted packet is sent. Alternatively, the request may be sent only once; when the filecast server has received the first request, it will transmit an extracted packet every time when extraction has been made. In addition, the filecast server adds to each extracted packet a portion of information about how long the terminal must wait until the next extracted packet is sent. Transmission of extracted packet is broken off in response to a stop request of the terminal. Alternatively, transmission may be continued during a predefined time or till the stored digital media object ends.

Although DVB-H has been presented as a filecast enabled mobile broadcast network, other networks such as T-DMB and DAB-IP as well as MBMS can be used. It should be noted that MBMS will be a technology for 3G networks, whereas the missing packets can be retrieved from 2.5G (e.g. GPRS) with better coverage. In Internet, peer-to-peer and multicast networks can be used as the filecast enabled network.

As mentioned earlier, a typical prior art filecast system comprises of an arrangement of ordering particular files or a group of files. When ordering the files or group of files, it may be feasible to order also the respective extracted packets whenever they become avail- able. The packets can be received by the terminal and buffered for later playback. Alternatively the resulting complete packet stream may be stored after inserting the extracted packets, instead of storing the incomplete media object.

Claims

1. A method for handling in a filecast server a stream of a digital media object to be protected against unauthorized use, wherein the digital media object is distributed via a mass distribution channel as filecast to a plurality of terminals, the method comprising the steps of: selecting such portions of data from the stream, the absence of which cause disturbances when the digital media object is reproduced in a terminal, forming an incomplete digital media object stream by extracting said selected portions of data from the stream, storing said extracted portions on a memory medium, converting the incomplete digital media object stream into a format suitable for a mass distribution channel, distributing for persistent storage the converted digital media object stream via the mass distribution channel.

2. The method as in claim 1 , characterized by the further steps of: responsive to a terminal's request received from a bidirectional communication channel, transmitting at least one extracted portion through the bidirectional communication channel to the terminal, whereupon the terminal is able to insert said at least one extracted portion into the stream of incomplete digital media object and so reproduce the complete digital media object.

3. The method as in claim 2, characterized in that responsive to the first request all extracted portions are transmitted.

4. The method as in claim 1 , characterized by the further steps of: fetching from a subscriber database information about users authorized to reproduce the complete digital media object, and transmitting the extracted portions to the terminals of the authorized users.

5. The method as in claim 1 , characterized in that each of the extracted portions is substituted with a replacement portion.

6. The method as in claim 5, characterized by incorporating into the replacement portion information about the location of at least the subsequent replacement portion in the stream.

7. The method as in claim 2 or 3, characterized in that extracted portions are transmitted prior to mass distribution of the incomplete digital media object.

8. The method as in claim 1 , characterized in that the digital media object is encrypted prior to extracting portions of data.

9. The method as in claim 1 , characterized by the further steps of: converting the incomplete digital media object stream into a format suitable for mass distribution to peers of a peer-to-peer network, distributing to the peers the incomplete digital media object converted, transmitting extracted portions to authorized peers only.

10. The method as in any one of claims 2-4, characterized by the further step of: digitally fingerprinting the extracted portions prior to transmission to the terminal, storing terminal -related fingerprints on the memory medium, wherein a terminal that is the origin of unauthorized copies of the digital media object is traceable.

11. The method as in claim 1, characterized in that each extracted portion of data is a packet of the stream of the digital media content.

12. The method as in claim 8, characterized in that said packet comprises of data from the first forward prediction (P) frame of the MPEG -compressed digital media object GOP structure.

13. A method of reproducing a digital media object received by a terminal as a streaming incomplete digital media object from the local storage, the digital media object being protected against unauthorized use, the method comprising the steps of: monitoring the streaming incomplete digital media object for detecting a position where a data portion has been extracted from the original stream of the digital media object, responsive to detection of the position, inserting an extracted portion of data received via a bidirectional channel and addressed to the terminal into said position so reproducing the original stream.

14. The method as in claim 13, characterized in that detection is based on tracing a predetermined replacement pattern in the streaming incomplete digital media object, wherein said replacement pattern is an indication of the position of the extracted data.

15. The method as in claim 13, characterized in that detection is based on keeping track of packet numbers in the streaming incomplete digital media object, wherein a missing number indicates a position of extracted data.

16. The method as in claim 13, characterized in that detection is based on examining packet header information, wherein a data portion extracted is indicated in the header information of the packet.

17. The method as in claim 13, characterized in that the location of each extracted portion is received from the bidirectional channel.

18. The method as in claim 13, characterized in that, upon detection of the position of missing data, sending to a digital media server via a bidirectional communication channel a request for the extracted portion of data, and responsive to the request, receiving the extracted portion of data from the bidirectional communication channel.

19. The method as in claim 13, characterized by the further steps of: responsive to ordering a file or group of files from the mass distribution channel, sending a subscription message via the bidirectional communication channel to a filecast server, receiving extracted portions of data whenever they become available.

20. A filecast server for handling a stream of a digital media object to be protected against unauthorized use, the filecast server operating in conjunction with a mass distribution system distributing the digital media object to terminals and in conjunction with a bidirectional communication system providing a point-to-point connection between the filecast server and a terminal, characterized in that the filecast server comprises an analyzer adapted to receive a stream of a digital media object and to analyze the stream for finding a predefined portion of data thereof, and responsive to said portion of data to produce a selection signal, an extractor for extracting the predefined portion of data from the stream responsive to the selection signal, wherein an incomplete digital media object stream is constructed, a mass distribution data interface adapted to transmit the incomplete digital media object stream to the mass distribution system for further delivering to terminals as file- casts, a packetizing block operatively connected with the extractor for receiving the extracted portion of data therefrom, and adapted to create a unicast packet for conveying the extracted portion of data, a unicast data interface adapted to receive through the bidirectional communication system a request for the extracted portion of data and responsive to the request to transmit the unicast packet addressed to the terminal through the bidirectional communication system.

21. The filecast server as in claim 20, characterized in that the filecast server further comprises a replacement packet-forming block operatively connected to the extractor and adapted to form a replacement portion, a replacement insertion block adapted to receive the replacement portion and to insert it in place of the extracted portion of data, wherein the incomplete digital media object stream is formally conformed to the complete digital media object stream.

22. The filecast server as in claim 20, characterized in that the packetizing block further comprises a fingerprinting unit for adding a terminal -specific fingerprint to the extracted portion of data.

23. The filecast server as in claim 22, characterized in that the fingerprinting unit selects the fingerprint from a group of fingerprints, wherein the unicast packets sent to the terminal are distinguishable from the unicast packets sent to other terminals by different combinations of the fingerprints.

24. The filecast server as in claim 20, characterized in that it comprises a subscriber authentication unit for authentication of the terminal requesting the extracted portion.

25. The filecast server as in claim 20, characterized in that the analyzer is adapted to analyze packets for finding predefined packets thereof, wherein a portion of data is a packet.

26. The filecast server as in claim 25, characterized in that the predefined packets are determined to consist of around 2 % of the packet stream.

27. The filecast server as in claim 20 characterized by a fingerprint storage for storing terminal-specific information about fingerprints added into the extracted portions of data sent to the terminals.

28. A terminal having a mass distribution network receiver for receiving a transmission carrying a streaming digital media object and a packet transceiver for bidirectional communication with a filecast server, wherein the terminal is capable of reproducing the digital media object, characterized in that the terminal comprises local memory connected to the receiver for persistently storing the digital media objects received as streaming digital media objects therefrom, a missing data detector operatively connected to the local memory, and for producing a notification signal in response to detection of a position of an extracted portion of data, a requester operatively connected to the packet transceiver, for transmitting a request for the extracted portion of data to a filecast server in response to the notification signal, an insertion block for inserting the extracted portion of data received from the file- cast server through a bidirectional communication network into the position detected, wherein the complete digital media object is formed.

29. The terminal as in claim 28, characterized in that the missing data detector is adapted to keep track of the consecutive numbers of the packets received, wherein the notification signal is triggered by a missing number.

30. The terminal as in claim 28, characterized in that the missing data detector is adapted to identify predefined replacement portions having the same content, wherein the notification signal is triggered by an replacement portion identified.

31. The terminal as in claim 28, characterized in that the missing data detector is adapted to examine packet header information, wherein the notification signal is triggered by a predefined tag found in the header.

32. The terminal as in claim 30 or 31 , characterized by a remover for removing the replacement portion prior to applying the stream received to the insertion block.

33. The terminal as in claim 28, characterized in that the information of position to where the extracted portion of data is to be inserted is read from the data received from the bidirectional channel.