The present invention relates to a terminal device and a method of displaying an image corresponding to a digital broadcast television signal on a display of the terminal device. In particular, the present invention relates to a decoding accelerator for enabling decoding and display of a digital television signal on a mobile terminal.

The use of digital, as opposed to analogue signals, for television broadcasts and transmission of other types of video and audio signals has been proposed as a way of allowing improved picture quality and more efficient use of spectral bandwidth. The International Standards Organization (ISO) has set a standard for video data compression for generating a compressed digital data stream which is used for digital television. This standard is referred to as the ISO MPEG (Moving Picture Expert's Group) standard. In accordance with the MPEG standard, video data is encoded using discrete cosine transformation (DCT) encoding and is arranged into variable length coded data packets for transmission. What is commonly referred to as "MPEG video" actually consists of several standards, MPEG-I and MPEG-2, which are similar in basic concepts and based on motion compensated block-based transformation coding techniques, while a third standard, MPEG-4 deviates from these more traditional approaches in its usage of software image construct descriptors for target bit rates in the very low range, smaller than 64Kb/s. In DCT based video compression systems, a 2-D DCT is applied to decorrelate the spectral energy of a group of pixels such that compression can be achieved by applying quantization as well as removing the irrelevant spectrum. In typical systems, the coefficients are assigned a one-dimensional ordering to allow for sequential transmission. MPEG allows for two different orderings, generally called zigzag scan and alternate scan. Both scans aim to cluster the energy starting at the DC value up to the highest frequency of that particular transcoded block. This allows the removal of frequency components that are not visible (irrelevance information) and to reduce (quantize) the amplitudes of visible frequency information. As a result a large part (the lower right area) of the 2D-DCT matrix becomes zero. This allows the abortion of the 2D-DCT coefficients with aid of an End-Of-Block

(EOB) symbol. The scanned coefficients are efficiently coded using a run- level variable length code word.

In corresponding applications such as set-top-boxes (STB), MPEG-2 parsing and decompression is executed on a media engine supported by hardware accelerators or in hardware only. For the situation of standard definition digital TV, the target resolution matches that of Dl, which is a digital resolution standard and means 720x480 pixels in the NTSC (National Television Standards Committee) television system and 720x576 pixels in the PAL (Phase Alternating Line) television system.

However, in TV-on-Mobile applications, where digital television signals are intended to be received and displayed at a mobile telephone device, it is an objective to have the MPEG-2 decoding algorithm in software, running on a platform generic to any mobile phone architecture, exploiting the fact that the display resolution in a mobile phone or other terminal devices is usually much lower than Dl. Moreover, parsing of the DCT coefficients in the zigzag scan is a time consuming function causing a considerable amount of computational load.

It is therefore an object of the present invention to provide an improved decomposition of functionality such that digital television signals can be displayed on a conventional architecture of terminal devices at reduced processing load.

This object is achieved by a terminal device as claimed in claim 1, by a display method as claimed in claim 8, by a computer programme product as claimed in claim 11, and by a computer-readable medium as claimed in claim 12.

Accordingly, processing load in the terminal device can be reduced by extracting matched transformation coefficients of predetermined significance and reassembling the extracted transformation coefficients to a new data stream used for generating the image at the display of the terminal device. In this way, only most significant coefficients are left which match with the resolution of the mobile display, leaving out all non-relevant coefficients. Incoming energy is thus adapted to the target display resolution to aid in lowering the number of decoding operations. The benefit of this approach, which can be implemented as a software approach, is that any platform of terminal devices, e.g. mobile terminals, can be used for displaying the digital television signal. Thereby, a very flexible solution is provided, which can be implemented for any kind of terminal device and any kind of digital television broadcast signal. Moreover, data throughput between the receiver front-

end and the terminal back-end can be reduced to the absolute minimum amount of data per second.

The transformation coefficients may be extracted for each block of a compression format of the digital broadcast television signal. Then, the extraction may comprise truncating a zigzag scan of transformation coeffients that represent the block of the compression format. The reassembled stream can be obtained by inserting an end-of-block code at a position that matches the number of extracted transformation coefficients. To achieve this, the received digital broadcast television signal is parsed and the parsed signal is used in the decoder acceleration process. Thereby, processing load can be reduced by truncating the series of transformation coefficients, e.g. DCT coefficients, in the zigzag scan of the video compression format, e.g. MPEG format. As a result, the decoding procedure can be implemented on common platforms, such as conventional mobile phone architectures. Preferably, a scalable decoding may be used after reassembling the stream. This allows to decode and sub-sample the 2D signal in one processing step. In view of the fact that the suggested solution to the above problem can be implemented based on a pure software approach with almost no hardware modifications, implementation may be based on a computer programme product which can be distributed to terminal devices as a pure software download or stored on a computer-readable medium, such as an optical or magnetical disc or the like. Further advantageous modifications are defined in the dependent claims.

The present invention will now be described based on a preferred embodiment with reference to the accompanying drawings in which: Fig. 1 shows a schematic block diagram of a mobile terminal with a television receiver front-end and a mobile phone back-end, according to the preferred embodiment; and

Fig. 2 shows a schematic block diagram of a decoding accelerator according to the preferred embodiment.

The preferred embodiment will now be described on the basis of an exemplary combination of a DVB-T (Digital Video Broadcast - Terrestrial) receiver in a mobile phone architecture.

Broadcasting technologies like DVB should not become a competitor to a mobile telephone standard such as GPRS (General Packet Radio Services) or UMTS (Universal Mobile Telecommunications System). There are significant advantages in trying to exploit the benefits of both to enhance services provided to the consumer. The DVB core system is designed to carry a flexible combination of MPEG-2 video, audio and data using a common MPEG-2 transport stream multiplex. The DVB-T system is specified in the ETS 300 7XX standard family and defines a digital terrestrial television system designed for 8MHz and 7MHz terrestrial channels. Transmission is based on digital modulation of an RF carrier in accordance with the ISO/IEC 13818-1 standard and comprises an elementary stream (ES) of compressed video data, such as MPEG frames.

Fig.l shows a schematic block diagram of the preferred embodiment in which a mobile digital television (DTV) receiver front-end 10 according to the DVB-T system is coupled to a mobile phone back-end or architecture 20 which comprises a MPEG-2 decoder 202 and a mobile display 204. The DVB-T radio signal is received via a receiving antenna and supplied to a

VHF/UHF (Very High Frequency / Ultra High Frequency) receiver 102 for channel selection to receive a desired DVB-T signal. For modulation of the DVB-T carrier, Coded Orthogonal Frequency Division Multiplexing (COFDM) has been chosen, where the data is modulated onto a large number of carriers using FDM (Frequency Division Multiplexing) technique. The key features which suit the modulation well to terrestrial channels include orthogonality, addition of a guard interval, and use of error coding, interleaving and channel-state information. The carriers are modulated by complex numbers which change from symbol to symbol. Inter-symbol interference and inter-carrier interference may happen if carrier phases and/or amplitudes change due to multipath propagation and corresponding relative delays. This problem can be alleviated by adding a guard interval, which ensures that all information integrated comes from the same symbol and appears constant during it. By addition of the guard interval, the symbol period is extended so as to exceed the receiver integration period. As long as the delay of any path with respect to the main (shortest) path is less than the guard interval, all signal components within the integration period come from the same symbol and the orthogonality criterion is satisfied. The guard interval length is chosen to match the expected level of multipath.

The received and filtered channel signal is supplied to a COFDM demodulator 104 where the digital data stream is obtained by demodulating the carriers. This may be achieved by multiplying the carriers by a carrier of the same frequency and then integrating

the result. Any other carriers will give rise to beat tones which occur at integer multiples of the carrier frequency and thus integrate to zero. Hence, without explicit filtering, all carriers can be separately demodulated without any mutual cross talk, just by a particular choice for the carrier spacing. The carriers can be closely packed so that they occupy the same spectrum in total as would a single carrier if modulated with all the data and subject to ideal sharp-cut filtering.

In a following demultiplexer stage 106, the demodulated data stream is demultiplexed to extract the elementary stream ES-V of video data which is intended to be decompressed at the MPEG-2 decoder 202 and displayed at the mobile display 204. In the preferred embodiment, a variable run- length decoder (VLD) accelerator

108 is provided to relieve the parsing load in the mobile phone architecture 20 for MPEG-2 streams. By adding this functionality to the television receiver front-end 10, the mobile phone architecture 20 does not need any digital television-specific functions, while at the same time the communication bandwidth between the two terminal portions can be reduced. From the viewpoint of the mobile phone architecture, the external VLD accelerator 108 of the television receiver front-end 10 provides for lowering the number of decoding operations in the MPEG-2 decoder 202.

The VLD accelerator 108 may be implemented as a first software routine in a dedicated processing device of the mobile phone architecture 20 or in a central processing device of the terminal device. The VLD accelerator 108 serves to reduce the amount of DCT coefficients in the elementary video stream ES-V in order to match the energy such that it matches the display resolution of the mobile display 204. Thereby, computational load of the MPEG-2 decoder 202, which may be implemented as a second software routine in a dedicated processing device of the mobile phone architecture 20 or in the central processing device of the terminal device, can be reduced so that a conventional stressing platform is sufficient.

Fig. 2 shows a schematic block diagram of the VLD accelerator 108 according to the preferred embodiment. The elementary video stream ES-V is supplied to a VBV (Video Buffer Verifier) buffer iunctionality 1081 which serves as a buffer for temporal storage of at least one picture. Then, the video data is supplied to a header prefix parser functionality 1082 which generates a control signal for a parser control functionality 1086 and which brakes or decomposes data into smaller elements according to a set of rules that describe the data structure. In particular, the header prefix parser functionality 1082 acts to extract header prefixes provided in the elementary video stream ES-V. The extracted header

prefix portions and/or corresponding control information are supplied to a video stream multiplexer functionality 1087 where a new elementary video stream ES-V with reduced number of DCT coefficients is reassembled at the output of the VLD accelerator 108. After header prefix parsing, the video data stream is supplied to a slice processor 1089 where the variable run- lengths are parsed, e.g. every bit is examined to determine boundaries between variable length codes and table look-ups are executed at speeds up to the input bit rate. The parsed variable length codes are supplied to a truncate zigzag scan functionality 1084 where the zigzag scan of DCT coefficients which represent a block in the MPEG-2 format is truncated or shortened in such a way that only the most significant coefficients are left, which match with the resolution of the mobile display 204. To achieve this, the incoming and parsed run-length coded stream is subjected to the truncated zigzag scan. Then, in an EOB insertion functionality 1085, an EOB code is inserted in the zigzag scan on a position that matches the required number of DCT coefficients.

In the slice processor 1089 the VLD parser functionality 1083 and the truncate zigzag scan functionality 1084 are controlled by the parser control functionality 1086 based on control signals derived from the higher video layers supplied by the header prefix parser functionality 1082 to the parser control functionality 1086. Thereby, the processing of the variable length code can be adapted to the format of the received elementary video stream ES-V. A refinement functionality may be added such that field pictures that are not used by the scalable MPEG-2 decoder 202 are replace by dummy field pictures to reduce the bit rate even further. This would require corresponding modification of the current slice processor behaviour. To achieve this, the original field picture information can be flushed and simply replaced by a pre-defined field picture.

At the output of the slice processor 1089, the video stream with the reduced number of DCT coefficients and inserted EOB code is supplied to the multiplexer functionality 1087 where the elementary video stream ES-V is reassembled together with the header prefixes and supplied to an output buffer functionality 1088 for rate matching purposes. Thus, all non-relevant DCT coefficients are left out from the elementary video stream ES-V, so that the delivered signal or stream contains only the energy required by the MPEG-2 decoder 202 to generate a video signal for the mobile display 204.

In the mobile phone architecture 20, the variable run- length coded stream is decoded again, but only for the remaining number of DCT coefficients and thus at reduced processing load.

In summary, the amount of DCT coefficients is reduced in the VDL accelerator 108 so as to just match the spatial energy to display resolution of the mobile display 204. Thereby, the mobile phone architecture 20 is relieved from the parsing task, and processing load is reduced, provided that the sub-sampling process to derive the final display resolution is done in a smart way, e.g. using a scalable MPEG-2 decoder. The series of DCT coefficients is truncated in the zigzag scan of the MPEG-2 format. As a result, the MPEG-2 decoding algorithm can be implemented in any common processing platform of the mobile phone architecture 20.

It is to be noted that the present invention can be used for any TV-on-Mobile applications where television signals based on a stream of transformation coefficients are to be processed by a resource constraint architecture such as a mobile phone. In particular, the present invention can be implemented in connection with any digital television standard and mobile phone architecture to reduce decoding effort. The gain of removing irrelevant energy by reducing the number of transformation coefficients becomes more important when the input bit rate is high and the target display resolution is low. The transformation coefficients may be any frequency amplitudes obtained by any kind of transformation. Furthermore, the truncation or shortening can be performed in a way that only coefficients with at least a predetermined significance, e.g. higher than a predetermined level of significance, are left, which match with the resolution of the mobile display. The preferred embodiments may thus vary within the scope of the attached claims.

It will further be noted that the above-mentioned embodiment illustrates rather than limits the invention, and that those skilled in the art will be capable of designing many alternative embodiments without departing from the scope of the invention as defined in the dependent claims. In the claims, any reference signs placed in parenthesis shall not be construed as limiting the claims. The words "comprising" and "comprises", and the like, do not exclude the presence of elements or steps other than those in any claim or the specification as a whole. The singular reference of an element does not exclude the plural reference of such elements and vice versa. If certain measures are recited in mutually different dependent claims, this does not indicate that a combination of these measures cannot be used the advantage.