Automatic Titling for International Sporting Events Jesús Martínez Barbero Eugenio Santos Menendez Abraham Gutiérrez Rodríguez ABSTRACT In sporting event transmissions of international significance, broadcasting in each country is carried out by the owner of the rights for that particular country. The output signal (program signal) in the original production center normally contains the graphic elements presented in the language considered the most international: English. The problem created by this dirty signal in the different broadcasters is double: On one hand, title adaptation produces the superposition of different graphics on a same image. On the other hand, those titles will remain in the channel s historical files. In order to solve the previous problem, different options are presently being taken, which involve increased program production costs. This is also aggravated by the introduction of high definition production. In this article, an inexpensive solution is proposed for the adaptation of these titles to each country. This solution is based on the use of the distribution and video production current infrastructure, for titling in the channel s natural language. In this manner we offer the audiences a more competitive product, without cost increase involved. Categories and Subject Descriptors H.5 [Information Interfaces and Presentation]: Group and Organization Interfaces - Organizational design. General Terms Management, Documentation, Design, Economics, Reliability, Theory. Keywords MPEG2-TS, Digital Television, Unidirectional, Cable T, Broadband Communications, Multicast, content management, content delivery, content distribution. 1. INTRODUCTION The different phases of the international production and broadcasting of a live event are the following: Production, channel distribution, packaging with the playout system in each channel, and broadcasting. Production is usually performed by mobile units in the same stadiums where the event takes place. In these units the different signals are mixed and the necessary titling is subsequently added. The signal is usually uplinked to a satellite so that the downlink can be performed by any of the channels owning the broadcasting rights for each territory. This is a one-to-many transmission. The channels downlink the satellite signal and they package it with the playout system of their channels to broadcast it through the Media owning the rights. Generally this transmission is one-to-many. In the signal production an international signal is generated with titles in English, in order to provide all the channels with a general and uniform titling service. This initial advantage at the production level creates future disadvantages. One of them is the impossibility of generating differing titles, since these would collide with those of the original signal when superimposing in the same image. In addition, the filing of such image in the general channel files entails another inconvenience: As the image is previously titled, its value is reduced because the type of production and its possibilities are limited. 2. CURRENT SYSTEMS To solve the problem of titling in the images different points of view are generally used. 2.1 Signal Production. The signal producer usually sends the satellite two different signals: a clean signal or clean feed and a signal with titles in the international language. Each channel is free to downlink one signal or the other. This alternative has the disadvantage of increasing costs, because an additional via has to be leased for signal distribution. Figure 1. Image with and without title
In fig (1) we can see the image with and without titles. The problem is aggravated by the introduction of contents in high definition, since the format is changed from 4/3 to 16/9, in order to distribute the titled signal. The signal producer has various alternatives: - 2 signals: HD with titling for 16/9 and SD with titling in 4/3. - 1 Signal in 16/9 with titling in 4/3. Apart from the clean signal without titles, the first alternative implies the generation and distribution of two added titled signals for the different formats (SD, HD). The second alternative implies the sending of an additional signal in the HD (16/9) format, with titles compatible for 4/3. The first option involves an added cost regarding titling resources and video link. The second compels the broadcasting channels to convert the HD signal into SD. 2.2 Event Broadcaster The channels owning the rights for event broadcasting in each country package it in the playout for its broadcasting in their territory. The kind of packaging in the playout differs from the significance given to the event, being able to create specific contents for a particular country. In this way various possibilities are offered. These range from the coverage using their own cameras mixed with the general program signal at the stadium, to the simple broadcasting of the downlink signal. One of the intermediate possibilities of the coverage consists of generating the titles in the specific language of each country with the introduction of particular data. This type of coverage requires a second production on the part of the broadcaster. This method was highly used in the past but it is currently falling into disuse for various reasons: Image/title incoherence: Given the fact that there is a second production (producer and broadcaster), when the broadcasting channel tries to include a title which refers to an image coming from the producer, this image may change with the title on air producing a desynchronization between image and titling. Economic: Depending on the broadcaster s facilities, and on the different categories of labour involved, various types of personnel will be needed to generate the titles included in each broadcast. In many cases this task entails the opening of various television studios. As it can be observed, there is an additional cost implied in the above mentioned. Furthermore, in many cases, results are not event satisfactory. 2.3 Title Production Graphics are composed at the graphic machines by means of templates. These templates or models have a fixed content and a variable content. Fixed information refers to the common elements present in a certain type of title. That usually consists of one or more TGA sequences with the key information needed for the video mixer to inject the title in the image. The variable part can be an image (static or dynamic), or a text containing the classical text characteristics regarding font, size, etc. Title generation in a sporting event involves the enumeration of the types of titles to be used, the creation of these titles common elements and the creation of the spaces occupied by the variable data of the mentioned models, their characteristics and types of contents (i.e. texts, images, etc.). Before the start of the event, titles are prepared based on the generated templates. They are stored in pages or addresses to be called from the control terminal, where the signal is generated. During the broadcast of the event, data are being updated and new titles are dynamically generated in pages for broadcast. 2.4 Program Production The move of a specific title to broadcast requires that this has to be ready in the output of the graphic machine. There the video mixer triggers the title machine and thus the title can be on air. Each input channel in the video mixer has been assigned one or various gpis s signals set up by software. One of the functions assigned to these signals is its activation when the video mixer sends this input on air. This function is normally used to warn the anchors (commentators) that they are visible by means of a bright signal in the camera. To adapt the titling to a specific language, the most inexpensive way is to use the clean feed, without titles. This feed or signal is usually distributed via satellite the same as with the titled signal- and it is sometimes used as a backup signal in case of equipment failure. With the signal in the studios, it is necessary the intervention of several professionals to introduce the content of the titles, to generate them and to inject them in the signal to be distributed. As it has been previously commented, this type of adaptation loses a large amount of data from the original production. Many scene sequences have been programmed with titles inserted) in a specific order. The producer at the adaptation station ignores this previous order, being unable to react in real time.
2.5 MPEG 2 Transpor Stream Despite the different encoding formats which have gradually appeared in the market, the most extended video transmission format for the contribution, distribution and broadcasting of professional quality video signals continues to be the MPEG2 standard [1]. In the said standard two types of formats are specified, the transport stream and the program stream. The first is used for transmission due to its greater robustness concerning noises in the channel and the second is used for production in environments with low error rates. The various errors which may occur during transmission of the transport stream are corrected at reception, so as to minimize the effect that these may produce in the image. Multiple methods have been developed for this purpose. A program comprises several types of data (video, audio, data) which are encapsulated into elementary streams (ES) and multiplexed in a data stream. Each elementary stream is packaged into packets PES(Packetized Elementary Stream). In order to maintain synchronization between the audio and video data, time stamps are inserted for a correct decoding and displaying of images and sound. Figure 2 shows the multiplexing of a video signal, audio signal and other data associated with a program stream. In each program stream multiple video, audio or data channels may be associated. The speed of the elementary stream may vary depending on the quality required for the images. To contribute to a central location, the speed may vary from 8 to 50 Mbps. The nature of the images and the transmission purpose will determine the selected quality. For signal broadcasting, 2.5 to 7 Mbps are generally used. Figure 2. A program multiplexed on a transport stream. For the distribution of monitoring channels, either DVB or VolP, the channels can be compressed to higher rates (i.e. from 4 to 8 Mbps) and several programs can be multiplexed as a single transport stream as shown in Figure 3.
Figure 3. N programs multiplexed on a transport stream In either case, previous data are accessible from the transport stream, which is thus generated at the source with the application of inverse operations from the transport stream as shown in Figure 4. 2.6 File Transfer Signal distribution is generally performed through satellite link. These types of video connections are data unidirectional links in charge of transmitting the transport stream. At the reception, it is installed the corresponding demultiplexer. Figure 4. Obtaining the components of a program. Figure 5. Data and signal distribution. For file distribution over unidirectional links, there are various file transfer protocols based on retransmission patterns of the same file. The Reliable Multicast Transport (RMT) IETF Working Group deals with the standardization of reliable one-to-many multicast transport protocols. In [2], a study discusses three types of transfer protocols which can be used in unidirectional networks. The Asynchronous Layered Coding (ALC) [3] does not require any type of feedback from the receivers, and the data are encoded using FEC codes. Repetitions of symbol transfer guarantee the integrity of the file at the expense of diminished effectiveness in the bandwidth. The Nack Oriented Reliable Multicast (NORM) [4] retransmits only the damaged parts from some of the receptors which send signals of Negative Acknowledgments (NACK) over damaged blocks. The File Delivery over Unidirectional Transport (FLUTE) [5], based on the ALC protocol, with the extension to be used in any type of transmission channel (unidirectional or not) presents metadata which complete the image signal itself (i.e. File name, codec, etc.). 3. PROPOSED SYSTEM This is based on the transfer of the data necessary for the signal adaptation through data PES in the streaming of MPEG2 transport stream, which is distributed to the different TV stations.
3.1 Data generation There will be two types of installations: Master and Slave installations. As Master figure shows, the mixer consists of 1 to n inputs (10...In) and a GPI s (On) output (General Purpose Interface). Figure 6. Video mixer. The system consists of a spy software in charge of watching over the operations performed in the titling station. The type of data generated will consist of: The model of title together with its text, the pages being prepared for their broadcast and the page located at the titler output. On the other hand it will be necessary to detect every time the titler for broadcast is punctured by means of the GPI s detector. Each time this signal is set to 1, the triggering detector sends an impulse to indicate the mixer has been set to air. The various data are multiplexed over the output transport stream, together with the video without titling and the international audio, for its distribution to the different televisions. Figure 6 shows the different blocks necessary for the data withdrawal from the titler, for testing the mixer s GPI output and for multiplexing the data in the transport stream. The transport stream is uplinked to the satellite for its distribution (see figure 5) to the different broadcasters. Within the broadcasters facilities, the signal must be multiplexed to obtain the different signals and data. By means of the titler s remote control, they are being executed the same generation and title-display commands as in the origin. The GPI signal regenerates again to command the mixer to puncture the titler, every time that, in its origin, a title is punctured in the mixer. 3.2 Content distribution and conversion The files to be transferred will have a structure XML for a better adaptation of the different results to current technologies. In this way it will be obtained a data stream over the video stream. In the target station (estación destino) the system must process the contents coming with the image, for the sending of commands to the titling stations. Two types of translation will have to be carried out: Message adaptation to the language, being able to include adapted publicity. Command adaptation to the titling machine, since the target broadcaster may have a machine with a different protocol from that produced by the signal. It will also be necessary a distribution of the texts which, predictably, are going to be used so that translations of every item can be carried out in each television. Before each event broadcast, it will be necessary a distribution of the texts which predictably are going to be used, so that translations of every item can be carried out in each television, generating XSLT s for the language and for the commands of the target machine.
Figure 7. Generation of commands over XML files. Figure 7 shows as XML incoming generates a continuous stream of data through the XML files. In order to provide the system with a higher modularity, this is divided into two processes: The translation to the country language and the translation corresponding to the commands of the machine. For each incoming message from the event production, there is an XSLT associated containing the translation of the messages. Once the conversion to the language has been carried out, it takes place the conversion to the specific commands for the type of machine needed. Additionally, there will an XSLT for each type of command to be executed. The translation is carried out by means of XSLT files. XSLT is a programming language to transform documents XML into other XML documents, or into other types of documents. The most relevant characteristics of the XSLT are: At least, it is necessary one XML input file. Parts of the input document are accessible by using the XML path language. XSLT-Stylesheet defines the input conversion to other documents. The principle of the transformation carried out in the processors is the following: Input XML are title command data bases. The first XSLT set defines the translation to the local language. The second XSLT set defines the translation of the commands to the local titler. The processor reads each input document and generates files according to the rules defined in the input XSLT. XSLT has the required functions for handling complex transformation requirementsand provides great flexibility to the system as the output data, as well as its configuration, can be simply modified by changing the XSLT document without the need to compile the program, since the document is read during execution time. 3.3 Different types of graphics Graphic rules imposed by the organization to each broadcaster vary amply from one another. Whereas in many competitions the type of graphics is very strict and each broadcaster has to set the titling with the graphics imposed by the organization, in other cases there is full freedom for every broadcaster to apply its corporate graphics. By treating, in the processes, each type of title as a message, it is easy to change the model of graphic title associated to a message. By this configuration, the operation in both cases is simplified. Thus, for the carrying out of the broadcast in either modality it will be necessary to initially define the linguistic translation. In case of different graphics, the assignment of the types of graphic messages required will be defined.
3.4 Files One of the most repetitive broadcasters discussions over time deals with the filing procedures which should be implemented in each enterprise. One of the aspects present in these discussions is the filing of material with or/and without graphics. Titled images are usually considered dirty, but they contain information about the test data. Clean images are perfect as far as filing is concerned because of their easy reutilization. When an image is titled, it is more complicated to carry out future titling since problems may arise concerning the composition used between titles. To not spoil the file signal and to facilitate its reutilization, images are stored without the titling graphics, but with the data provided by the transport stream which comes to the station through the FUTE protocol that has previously been commented.. In this way, there is access to the clean material and to the information present in the transport stream itself. 4. CONCLUSIONS This paper enumerates a flow of data which enables cost reduction in adapting sporting events, thus adding a series of advantages: To provide the said adaptation with (or of?) the original production, avoiding delay due to the users reaction time. To favour more advantageous use of the file, since simultaneous recording of the clean and titled image is not necessary. To assist in the content adaptation to the screen format used by each broadcaster. To adapt the titling to the types of graphics that each broadcaster wishes to use for whatever corporate image reasons. To enable the inclusion of local publicity, this may help to finance the cost of rights. To enable the reutilization of the data included in the transport stream, in order to carry out reproductions adapted to each user. As it can be observed, having the data embedded within the image signal itself, provides with a series of advantages which are very easy to make use of. By means of this configuration, a new via is opened for production, much better adapted to new reproduction environments and which adapts the sector to face future types of broadcast/distribution which may arise. 5. REFERENCES (1) ISO/IEC, 15444-1:2000 Information technology - JPEG2000 - image coding system - Part 1: Core coding system, International Standard ISO/IEC 15444-1 Std., 2000. [2] Neumann, C., Roca, V., and Walsh, R. 2005. Large scale content distribution protocols. SIGCOMM Comput. Commun. Rev. 35, 5 (Oct. 2005). [3] M. Luby, J. Gemmell, L. Vicisano, L. Rizzo, and J. Crowcroft. Asynchronous Layered Coding (ALC) protocol instantiation, Dec. 2002. Request For Comments 3450. [4] B. Adamson, C. Bormann, M. Handley, and J. Macker. Negativeacknowledgment (NACK)-Oriented Reliable Multicast (NORM) Protocol, Nov. 2004. Request For Comments 3940. [5] T. Paila, M. Luby, R. Lehtonen, V. Roca, and R. Walsh. FLUTE -File Delivery over Unidirectional Transport, Oct. 2004. Request For Comments 3926. [6] U. Reimers, DVB The Family of International Standards for Digital Video Broadcasting, Proc. IEEE, vol. 94, no. 1, pp. 173 181, Jan. 2006. [7] U. Reimers, DVB (Digital Video Broadcasting), Springer Verlag Berlin, 2nd Edition 2004. [8] Bürklin, H., Schäfer, R., and Westerkamp, D. 2007. DVB: from broadcasting to ip delivery. SIGCOMM Comput. Commun. Rev. 37, 1 (Jan. 2007), 65-67. [9] L. Liang, H. Cruickshank, Z. Sun, C. Kulatunga, G. Fairhurst, TESLA with FLUTE over Satellite, Int. Conf. Communications (ICC), Beijing, China, 2008. [10] Bürklin, H., Schäfer, R., and Westerkamp, D. 2007. DVB: from broadcasting to ip delivery. SIGCOMM Comput. Commun. Rev. 37, 1 (Jan. 2007)