FR2832582A1 - Multimedia data coding process for domestic networks selects parameters using quality assessment - Google Patents

Abstract

A multimedia data coding process evaluates coding memory (E403) requirements and unit quality (E407) for different coding parameters from comparison of coded (E405) and decoded (E406) data.

Description

<Desc / Clms Page number 1>

 The present invention relates to a method for determining at least one encoding parameter of multimedia digital data organized in elementary data units to be coded according to at least one coding mode.

 The technological advancement of digital communication devices such as televisions, pocket PCs, conventional microcomputers or any other miniaturized device is such that, in the coming years, it will most likely be possible to exchange data from one another. multimedia type and, for example, audio and / or video between these different devices.

 Market research also shows that users are willing to buy such communication devices as long as the price is not too excessive and, above all, the services provided are of good quality.

 Following the identification of this new market, numerous studies have been conducted on the miniaturization of electronic components, in order to obtain devices having high memory and computing capacity.

 In addition, many standardization bodies have established standards to unify the exchange of data between these devices with different characteristics. These standards, such as those of the IEEE 802 family, allow both the implementation and management of data transfer via a particular network.

 The IEEE 802.11 standard and its European counterpart HYPERLAN are dedicated to wireless LANs. They are particularly studied now

<Desc / Clms Page number 2>

 because they are broadband networks for transmitting audio and / or video data between two machines that are geographically close.

 These standards are therefore particularly suitable for developing wireless LANs inside homes.

 The technical solutions currently studied in the context of the aforementioned standards applied to the local home networks make it possible to transmit digital data either between a so-called server machine and a so-called client machine (point-to-point transmission), or between the server and a group of clients, again between the server and several clients (multi-point communication) without any wires.

 In this environment, the data is stored on the server that centralizes all communications with the client (s). This server can also serve as a gateway to the outside world (Internet, television, camcorder ...).

 It should be noted that home local area networks may be heterogeneous in nature, that is, they may, for example, be partly wireless and partly wired.

 While many projects are studying the implementation of interactive television services and the exchange of information on the Internet, few of them address issues related to the implementation of local home networks, namely implementation of solutions to obtain an acceptable quality of service.

 Since the server is connected to the outside world, the data stored is of various sizes. Transporting this data over a local network is not an easy service to set up.

 A communication system is known from the article "Adapting Multimedia Internet Content for Universal Access" by R. Mohan, J. Smith, C-S. Li, IEEE Transactions on Multimedia, March 1999, which plans to build a database listing the different categories of communication devices that can connect to a server.

<Desc / Clms Page number 3>

 For each category of device, the system takes into account the size of the display screen of this category of device and the color depth that the device can support (number of bits per color component).

 Then, the system prepares in advance, for each category of device and for each of the above characteristics, a coded version for each of the videos that one or the other of the devices could ask the server.

 These coded versions are then stored by the server.

 Thus, when one of the devices sends a request to the server, the system taps into the database and selects the precoded version of the data adapted to the category of the device and the aforementioned characteristics.

 Such a system has several disadvantages.

 Indeed, the system must perform many calculations to develop the different coded versions, thus requiring a large computing power and monopolizing the central processing unit for a relatively long time.

 Moreover, the constitution of such a database requires the use of memories with high storage capacity.

 In addition, the system described above is fixed in that it can not take into account a communication device of a new category that is not listed in its database.

 Taking into account a request from a new category of device would require for example to list in the database this device and, to do this, to develop coded versions of the different videos, adapted to the characteristics referred to above (screen size and color depth).

 However, this consideration would still be in the sense of an increase in the memory capacity of the system.

 The present invention provides for remedying at least one of the aforementioned drawbacks by proposing a new method and a new device for determining at least one multimedia digital data coding parameter which are such that the coding parameter is determined in such a way that

<Desc / Clms Page number 4>

 adapted according to a predetermined memory capacity which serves as threshold or reference.

 By determining one or more coding parameters appropriately, the data encoded with this or these parameters require for their decoding a total memory capacity less than or equal to the predetermined memory capacity.

 The subject of the present invention is thus a method for determining at least one coding parameter for multimedia digital data organized in elementary data units intended to be coded according to at least one coding mode, characterized in that said method comprises the steps following: - determination of an elementary memory capacity necessary for the decoding of each elementary data unit coded according to a different coding mode, - determination of a total memory capacity necessary for the decoding of a given number of elementary units of coded data as a function, on the one hand, of at least one elementary memory capacity determined for the decoding of an elementary unit of data encoded according to a coding mode and, on the other hand, the number of elementary data units coded according to each of the coding modes, - decision as to the determination of at least one coding parameter of the data s according to the total memory capacity previously determined and a predetermined memory capacity.

 Correlatively, the invention relates to a device for determining at least one coding parameter of multimedia digital data organized in elementary data units intended to be coded according to at least one coding mode, characterized in that said device comprises: means for determining an elementary memory capacity necessary for the decoding of each elementary data unit coded according to a different coding mode,

<Desc / Clms Page number 5>

 means for determining a total memory capacity necessary for the decoding of a given number of elementary data units coded as a function, on the one hand, of at least one elementary memory capacity determined for the decoding of a unit coded coded data element and secondly the number of elementary data units coded according to each of the coding modes, - decision means for determining at least one coding parameter of the coding methods. data according to the previously determined total memory capacity and a predetermined memory capacity.

 Thus, by determining the total memory capacity necessary for the decoding of a plurality of elementary units of coded data, the invention makes it possible to decide on the choice of the coding parameter or parameters adapted to a predetermined memory capacity.

 As a result, the coding parameter or parameters thus determined make it possible to code the data so that, for example, these coded data require, for their decoding, a total memory capacity less than or equal to the predetermined memory capacity.

 For example, the encoding parameter that was used for encoding the data before determining the total memory capacity may or may not be satisfactory in view of the determined total memory capacity. If this parameter is not satisfactory, it will then be necessary to determine a parameter which will, this time, be adapted to the predetermined total memory capacity.

 Note that if an encoding parameter is adjusted according to the predetermined total memory capacity, it is because a modification of this parameter has an influence on the total memory capacity which is necessary for the decoding of the data encoded with said parameter. .

 The invention does not require a high storage capacity as in the aforementioned prior art since it does not memorize, in advance, several versions of coded data according to different categories of communication device.

 In addition, the processing unit of the device according to the invention is not monopolized by long calculations for a relatively long time, as it is the

<Desc / Clms Page number 6>

 case with the system of the prior art which must constitute a database of pre-coded versions.

 More particularly, when several coding modes are used for the coding of the given number of elementary data units, the step of determining the total memory capacity is more particularly performed as a function, on the one hand, of the elementary memory capacity determined for the decoding of each elementary data unit encoded according to a different coding mode and, secondly, the number of elementary data units encoded according to each of the coding modes.

 According to another approach, when several coding modes are used for the coding of the given number of elementary data units, said method comprises an additional step of selecting the highest elementary memory capacity among the different elementary memory capacities determined for each one. coding modes.

 According to this approach, the step of determining the total memory capacity is more particularly performed as a function, on the one hand, of the highest elementary memory capacity previously selected and, on the other hand, of the given number of elementary units. coded data.

 Thus, instead of taking into account, for each elementary unit, the corresponding elementary memory capacity which has been determined in relation to its own coding mode, it is placed in the most unfavorable case by adopting each time the elementary memory capacity the highest.

 This makes it possible, on the one hand, to simplify the determination of the total memory capacity and, on the other hand, to ensure that there will be no problem of data storage capacity during the decoding operations.

 In the case where the steps of determination of the elementary memory capacity for each mode of coding, determination of the total memory capacity and decision are made in a first communication device connected to a second communication device by a communication network, the invention makes it possible to adapt easily and dynamically to communication devices not yet connected to the first device, but which can be connected thereafter.

<Desc / Clms Page number 7>

 According to one characteristic, prior to the decision step, the method comprises a step of comparison between the total memory capacity determined for decoding the given number of coded data elementary units and the memory capacity available in the second communication device for decoding of these data.

 The predetermined memory capacity thus corresponds to that of the second communication device.

 The method according to the invention thus makes it possible to adapt the coding of the data carried out in the first apparatus to the memory capacity available in the second apparatus.

 According to another characteristic, prior to the comparison step, the method comprises a step of obtaining by the first communication device the available memory capacity in the second communication device.

 Thus, the information on the memory capacity available in the second communication device is transmitted to the first device before the comparison step referred to above.

 If one or more communication devices are later connected to the first device, it is sufficient that the information on their respective memory capacity is transmitted to the first device so that it can adapt the data encoding appropriately for each device.

 According to an alternative embodiment, the method further comprises the following steps: determining the basic number of operations to be performed to decode each elementary unit of data encoded according to a different coding mode, determining the total number of operations to be performed during a predetermined time interval for decoding a plurality of elementary units of data coded according to, on the one hand, the previously determined elementary number and, on the other hand, the number of elementary units coded according to each of the coding modes .

<Desc / Clms Page number 8>

 Thus, the decision as to the determination of said at least one coding parameter is also made based on the total number of operations previously determined and a predetermined total number of operations.

 According to one characteristic, the steps of determining the basic number of operations, determining the total number of operations during a predetermined time interval and decision are performed in the first communication device.

 According to one characteristic, prior to the decision step, the method comprises a step of comparing the total number of operations determined during the time interval with the total number of operations that could be performed, during the same interval of time. time, in the second communication device.

 The predetermined decoding complexity thus corresponds to that of the second communication device.

 The method according to the invention thus makes it possible also to adapt the coding of the data carried out in the first apparatus to the computing capacity of the second apparatus in addition to its total memory capacity.

 According to another characteristic, prior to the comparison step, the method comprises a step of obtaining by the first communication device the total number of operations that can be performed by the second communication device during the predetermined time interval.

 Thus, the information on the calculation capacity of the second communication device is transmitted to the first device before the comparison step referred to above.

 If one or more communication devices are later connected to the first device, it is sufficient that the information on their respective calculation capabilities are transmitted to the first device so that it can adapt the coding of the data appropriately for each device.

 Furthermore, it is also possible to make the decision as to the determination of the coding parameter or parameters taking into account characteristics of the communication network such as, for example, the bandwidth available on this network.

<Desc / Clms Page number 9>

 According to one characteristic, the method comprises a step of estimating the quality of at least one elementary unit of data by comparing said at least one elementary unit of coded data with the determined coding parameter and said at least one elementary unit of data. not coded.

 It is thus possible to develop a version of the elementary data unit adapted to the predetermined memory capacity and possibly to the predetermined decoding complexity and which makes it possible to obtain data, once decoded, of good quality. that is, true to the uncoded elementary data unit.

 It should be noted that in the case of data transmission between a first and a second communication apparatus, it is thus possible to provide the second apparatus with a version of the elementary unit of data of good quality which is adapted to the available memory capacity in the second unit. the second device and possibly the computing capacity of the second device.

 According to another characteristic, when a plurality of coding parameters have been determined, the method comprises the following steps: - estimation of the quality of at least one elementary unit of data for each combination of coding parameters determined, - selection of the best quality among the different qualities estimated for the different coding parameters.

 It is thus possible to select, using another criterion, a combination of coding parameters from among several possible combinations, in each of which one or more coding parameters have been determined in a manner adapted to the predetermined total memory capacity and, optionally, to the predetermined decoding complexity.

 According to yet another characteristic, the method comprises the following steps: coding of at least one elementary unit of data with the determined coding parameter; decoding of said at least one coded elementary data unit;

<Desc / Clms Page number 10>

 estimating the quality of said at least one elementary unit of data by comparing this at least one elementary unit of uncoded data with that previously decoded.

 This method of estimating the quality is particularly simple to implement.

 The invention also relates to a communication apparatus comprising a device as briefly described above.

 According to another aspect, the invention also aims at: a computer-readable information storage means or a microprocessor comprising code instructions of a computer program for executing the steps of the method according to the invention as briefly set forth above, and - a partially or fully removable computer-readable information storage medium or microprocessor having code instructions of a computer program for executing the steps of process according to the invention such as that briefly described above.

 According to yet another aspect, the invention relates to a computer program loadable in a programmable apparatus, comprising sequences of instructions or portions of software code to implement steps of the method according to the invention as briefly discussed below. above, when said computer program is loaded and executed on the programmable apparatus.

 The characteristics and advantages relating to the device, to the communication device comprising such a device, to the information storage means and to the computer program being the same as those described above concerning the method according to the invention, they will not be recalled here.

 Other features and advantages of the present invention will emerge more clearly on reading the description which follows, made with reference to the appended drawings, in which: FIG. 1 schematically represents a client-server type communication architecture in which the invention can be implemented;

<Desc / Clms Page number 11>

 FIG. 2 represents a video data processing algorithm according to the invention; FIG. 3a represents a coding algorithm for video data; FIG. 3b represents a decoding algorithm for video data; FIG. 4a represents an algorithm for determining at least one video data coding parameter according to the invention and which corresponds to step E208 of the algorithm of FIG. 2; FIG. 4b represents an alternative embodiment of the algorithm of FIG. 4a; - Figure 5 is an embodiment of a programmable apparatus embodying the invention.

FIG. 1 shows a communication architecture of the client-server type in which the invention is advantageously implemented;
In this figure, a first communication device 1 acting as the server is connected to a second communication device 2 which is the client machine, via a communication network 3 and a connection which is considered as being established.

 This network is for example a wireless local communication network inside a home.

 More specifically, this network complies with the IEEE802 standard. 11.

 In the example under consideration, this is a point-to-point transmission of data between the server 1 and the client machine 2.

 However, the principle remains applicable in the case where the server is simultaneously connected to several client machines that may require the digital multimedia data server.

 In this case, however, it is necessary to take into account, on the one hand, the management of the adaptation and parallel transmission processes of the different adapted versions of the requested data and, on the other hand, the multiple accesses to the or to the server memories.

<Desc / Clms Page number 12>

 It should be noted that the data stored in the server may have been received from the environment outside the dwelling, for example, by another communication network such as the Internet.

 The data grouped under the term multimedia data may be, in a nonlimiting manner, still images, videos, sound, text data (eg graphic documents ...), documents in HTML, signals from a fax machine or a printer ...

 In the exemplary embodiment considered, only video data will be considered and these will be transmitted without loss on the network 3.

 The device according to the invention makes it possible, in a general manner, to adapt the content of the data requested by a client machine to the characteristics of this machine and, more particularly, to the memory capacity available therein.

 The memory capacity of a machine is the amount of memory available in this machine for storing data.

 In the context of the invention, it is more particularly the memory space available for storing data during the decoding operations of these data.

 Furthermore, it will be noted that the device according to the invention also allows, as a variant, also to adapt the content of the data to the memory capacity and the computing capacity of the client machine.

 By computing capacity of the client machine is meant the number of operations that this machine can perform in a predetermined time interval corresponding, for example, to 1 second.

 The device 4 according to the invention is for example integrated with the server 1 of FIG. 1 and may comprise electronic elements and / or software elements.

 However, the device according to the invention could, for example, be confused with the server 1.

 The device 4 comprises units 11 and 12 which respectively make it possible to encode and decode a video.

<Desc / Clms Page number 13>

 The server includes a digital data storage unit 5 such as videos.

 The algorithms of the coding and decoding methods will be described respectively with reference to FIGS. 3a and 3b.

 It should be noted that the videos are created, for example, by a digital video camera or any other means of data acquisition or are transmitted by an external communication network and are stored in the unit 5 which may or may not be included in the device. device 4.

 This unit can be, for example, a local or distributed database.

 In addition, the videos are stored either in uncompressed form, for example according to a YUV format, according to a preferred embodiment of the invention, or in compressed form, by using the coding method implemented in the coding unit 11. .

 It should be noted that a YUV video format means that the video concerned has three components, the Y component for the luminance and the U and V components for the chrominance.

 The device 4 also includes a storage unit 13 for temporarily storing digital data and coded video, and a data recovery unit 14 characteristic of the videos from the storage unit 5.

 The data recovered by the unit 14 are either the values associated with each of the pixels of each image of a video, or the information specific to this video, namely the size of the video file, the path to access this video. video and, in the case where the video is stored in compressed form, the encoding parameters that were used to encode that video.

 The device 4 also comprises a unit 15 for recovering the characteristics of the client machine 2, namely the memory capacity and the resolution of its display screen.

 It should be noted that these characteristics can be obtained as a result of the transmission by the device 4 of a specific request.

<Desc / Clms Page number 14>

 The device 4 further comprises a unit 16 for recovering the characteristics of the network 3, a unit 17 for determining the coding parameters, and a unit 18 for transmitting / receiving digital data.

 The unit 17 implements the method for determining the coding parameters which will be described later with reference to FIG. 4.

 The unit 18 makes it possible for it to transmit data to the client machine 2 through the network 3, as well as to receive data from the client machine.

 The device 4 according to the invention also comprises a control unit 19 of the various operations performed.

 It will be noted that the different units that make up the device according to the invention and, more generally, the server, reside on the same machine.

 However, it may be envisaged to distribute these units on several server machines and to establish communications between these different units through the communication network 3.

 This distribution of units does not change the implementation of the invention, we will consider later, for reasons of simplicity, all units are located on the same machine that we call server.

 The client machine 2 comprises a digital data transmission / reception unit 20.

 This unit 20 makes it possible to transmit data to the device 4, through the network 3, and to receive data from the device.

 The client machine 2 also comprises a digital data storage unit 21, a decoding unit 22 identical to the unit 12 of the device 4 and a display unit 23 for displaying decoded videos.

 FIG. 2 illustrates an algorithm comprising different instructions or portions of software code corresponding to steps of the method according to the invention.

<Desc / Clms Page number 15>

 The computer program noted "Progr" which is based on this algorithm is stored in the temporary data storage unit 13 of Figure 1 and executed by the unit 17 under the control of the control unit 19, which allows thus to implement the method according to the invention.

 This program is also stored in the apparatus of Figure 5.

 The algorithm of FIG. 2 comprises a first step denoted E200 during which the device 4 of FIG. 1 and, more particularly, the reception unit 18 receives a request sent by the client machine 2 from the unit transmission 20.

 For example, this request is in the form of a character string for uniquely identifying video data stored in the storage unit 5 of the server 1.

 In the next step E201, the request from the client machine 2 is transferred to the temporary storage unit 13 for storage.

 Step E201 is followed by a step E202 during which a test is performed to find out if the video data stored in the storage unit 5 and subject to the request of the client machine are available in the form compressed.

 If so, the step E202 is followed by a step E203 during which the decoding unit 12 of the device of FIG. 1 decodes the video data according to the coding parameters provided, for example, by the header of the bitstream stored in the storage unit 5.

 Once the decoding of the video data has taken place, the decoded data present in the YUV format as well as the aforementioned coding parameters are stored in the storage unit 13.

 Returning to step E202, when the test performed in this step is negative, step E204 provides for recovering the values of the available pixels in the storage unit 5 and storing them in the storage unit 13 .

<Desc / Clms Page number 16>

 The steps E203 and E204 which have just been described are both followed by a step E205 during which the unit 14 proceeds to a video data recovery.

 It is a matter here of obtaining characteristics of this video data, namely the coding parameter or parameters used if this video data was present in the storage unit 5 in compressed form.

 Thus, the coding parameters may be, for example, the spatial resolution of the video, the number of frames per second of this video (temporal resolution), the number of images coded in Intra mode or in Inter mode.

 In the case where the video data was not available in compressed form in the storage unit 5, the information associated with the video data and which is retrieved in the step E205 is, for example, the information that the video was not compressed and its resolution.

 In the next step E206, the characteristics of the client machine 2 that issued the request mentioned in step E200 are obtained.

 The characteristics of the client machine are for example recovered following the transmission of a request to obtain these characteristics, from the device 4 of FIG. 1 to the client machine 2, through the communication network 3.

 These characteristics are mainly the memory capacity available in the client machine 2 and the resolution of its display unit 23 (screen).

 However, as an alternative, these features may also include the computing capacity of the client machine, i.e., the total number of operations it can perform.

 These characteristics are stored temporarily in the storage unit 13 of the device 4 of FIG.

 Preferably, the computing capacity of the client machine is equal to the speed of the processor thereof, expressed in number of elementary operations per second.

<Desc / Clms Page number 17>

 Step E206 is followed by a step E207 in which the device 4 of FIG. 1 obtains the characteristics of the communication network 3 such as, for example, the bandwidth available for transmitting the video data of the device 4 to the device. client machine 2, packet loss, wrong bit rate ...

 Preferably, the characteristics of the network taken into account will be limited to the available bandwidth.

 These characteristics are recovered by the unit 16 which then stores them in the temporary storage unit 13.

 During the next step E208, the encoding parameter or parameters of the video data are determined in a manner adapted to the memory capacity of the client machine 2.

 The details of the operations performed during this step will be provided later in the description made with reference to FIG. 4a.

 Furthermore, as a variant, during this step E208, it is also possible to determine the data encoding parameter or parameters in a manner adapted to the memory capacity as well as the computing capacity of the client machine.

 This variant will be described later with reference to FIG. 4b.

 This step is performed by the unit 17 of the device 4 of FIG. 1 under the control of the control unit 19 and from the information available in the storage unit 13.

 The determination of this or these coding parameters can also be performed in a manner adapted to the characteristics of the communication network 3, in addition to the memory capacity of the client machine and possibly its computing capacity.

 During the next step E209, the device 4 of FIG. 1 and, more particularly, the transmission unit 18 of the latter proceeds to the transmission of the video data which has been previously compressed in step E208 to the customer machine 2.

 Note that the transmission of compressed video data is preferably done on the fly (known in English terminology under

<Desc / Clms Page number 18>

 the term "streaming"), that is to say that the information necessary for the user to reconstruct an image is stored in the storage unit 13, then are transmitted to the client machine 2 before all the video data are compressed.

 Thus, the user receives at the level of the client machine 2 the packet information, in each of which are information allowing this user to decode a current image.

 Figures 3a and 3b respectively illustrate the coding and decoding of a color video in YUV format.

 According to a preferred embodiment of the invention, the video decoder used for the implementation of this invention is in accordance with the visual part of the MPEG-4 standard (information technology-Generic coding of audiovisual objects = part2: visual, ISO / IEC JTC1 / SC 29 / WG11 No. 3056, December 2000).

 In the example discussed here, the video will be considered as a single rectangular object.

 Therefore, only a limited set of compression tools that allow texture decoding and motion compensation will be taken into account. The tools needed to decode the arbitrary form of an object are therefore not used here.

 It should be noted that the texture of an image is given by the values of the pixels of the image which are expressed on three components.

 The videos are stored on the server 1 either in compressed form MP4-compliant mp4 format or uncompressed YUV format.

 For example, the 4: 2: 0 YUV format is used, which means that the chrominance components (U, V) have four times fewer samples (pixels) than the luminance components (Y).

 FIG. 3a illustrates the coding method implemented by the coding unit 11 of FIG. 1.

 Each image of the video requested by the user is divided into 8x8 pixel data blocks.

<Desc / Clms Page number 19>

 Each group of four blocks forms a macro-block of 16x16 pixels.

 The coding of an image is performed on each of the macroblocks according to two distinct coding modes: one denoted 1 for the Intra coding mode and the other coded P for the predictive coding mode or Inter.

 A sequence of 1 second images will be coded as follows: IPPIPPIPPIPPIPPIPPIPPIPPIPPIPP ...

 This means that the first image is coded according to the Intra mode, the next two according to the Inter mode, the next according to the Intra mode, the following two according to the Inter mode, etc.

 Thus, for a sequence of 30 frames per second, 10 images will be encoded in Intra mode and 20 images in Inter mode.

 Note that a sequence of images will be called later a segment.

 In intra mode, the coding is carried out block by block according to the steps E310 to E313 of FIG. 3a.

 In this encoding mode, each block constitutes an elementary unit of data which is encoded independently of the other blocks.

 It is recalled that the values of the pixels of each image of the image segment to be encoded are stored temporarily in the temporary storage unit 13 of FIG. 1 and constitute a video which is called the original video.

 In the case where the video is stored in the storage unit 5 in compressed form, the decoding method is applied to the step E203 of FIG. 2 so as to obtain a video in the YUV format.

 The values of the video thus reconstructed are stored in the temporary storage unit 13 and in the storage unit 5, also called the original video.

 The method for encoding an Intra mode block begins with a step E310 during which the discrete cosine transform is calculated according to the known method described in the article entitled "Fast Algorithms for the Discrete

<Desc / Clms Page number 20>

 Cosine Transform, E. Feig and S. Winograd, IEEE Trans., On Signal Proc., Vol.

40, No. 9, September 1992.

 Step E310 is followed by a step E311 which provides for quantifying the transformed coefficients obtained during step E310.

The step E311 is followed by a step E312 in which the quantized coefficients are ordered so that the variable length coding method used in the subsequent step E313 is as effective as possible.

This is done by using the scanning method described in the Information technology-Generic coding of audio-visual objects = part2: visual, ISO / IEC JTC 1 / SC 29 / LVG11 N3056, December 2000.

 Step E312 is followed by a step E313 in which the statistical coding of the ordered coefficients is carried out according to the variable length coding method described in the standard Information technology-Generic coding of audio-visual objects = part2: visual, ISO / IEC JTC 1 / SC 29 / WG11 N3056, December 2000.

 At the end of step E313, the statistically coded coefficients are stored in the temporary storage unit 13 and are concatenated with the other coding information of the current image.

 In predictive or Inter mode, the coding is macro-block by macroblock where each macroblock constitutes an elementary unit of data which is coded independently of the others.

 Each of the macroblocks can be coded either in Intra mode or Inter mode.

 The coding mode is chosen according to the activity present inside a macro-block.

 For example, this activity is measured by the variance between a macroblock of a current image and the macroblock of the previous image located at the same spatial position.

 If the variance is greater than a predetermined threshold, then the macroblock is coded in Inter mode.

 This macro-block will be coded in Intra mode otherwise.

<Desc / Clms Page number 21>

 If the macroblock is encoded in Intra mode, then the coding is performed according to the steps E310 to E313 described above.

 If the macroblock is coded in Inter mode, the coding is performed according to the steps E315 to E317 which will be explained below.

 During step E315, a conventional motion estimation method is applied by comparison of macroblocks.

 The algorithms used are based on the simplifying assumption that the pixels of the same macroblock are animated by the same movement.

 The motion vector of a macroblock of a current image is then determined by finding the most resembling macroblock in the previous image in a preselected search window.

 If the dimensions of the search window have an influence on the calculation time of the motion estimation, the major problem related to the motion estimation which is carried out by comparison of macroblocks remains the determination of the distance between the macroblocks.

 A motion estimation method described in the article "Online motion estimation", L. Difdier, R., will preferably be used.

 Kamdem, LIM University of Provence, Dec. 96.

 This method makes it possible to obtain a motion vector with two components of translational movement.

 In order to obtain a binary train conforming to the MPEG-4 standard, this motion vector is coded differentially with respect to the motion vector of the macroblock of the preceding image disposed at the same spatial position.

 Step E315 is followed by a step E316 in which the previously determined differential motion vector is statistically coded.

 The coefficients thus obtained are stored in the temporary storage unit 13 and are concatenated with the other coding information of the current image.

<Desc / Clms Page number 22>

 Once the movement for a macro-block has been estimated during the execution of step E315, then, according to step E317, a predicted macro-block is constructed from the macro-block of the previous image to which the motion vector associated with this macroblock is applied.

 Then, a so-called difference macro-block is calculated between the macroblock of the current image and the predicted macroblock.

 Step E317 is followed by a step E318 in which the difference macro-block is coded according to the Intra mode described above.

 The coefficients resulting from the Intra coding mode are stored in the temporary storage unit 13 and are concatenated with the other coding information of the macroblock of the current image.

 The parameters used during the quantization and variable length coding steps are defined for each macroblock, whether in the Intra or Inter mode.

 For this purpose, a conventional flow control method is used during a step E314.

 This method is used to control the rate allocation for each macroblock so as not to exceed the total bit rate allocated to the transmission of the compressed video.

 In the example described, the maximum flow value is set by the value BR.

technology-Generic coding of audio-visual objects = part2 : visual, ISO/IEC JTC 1/SC 29/WG11 N3056, December 2000, avant d'être transmises sur le réseau 3 de la figure 1 vers la machine client 2. The data coded and stored temporarily in the unit 13 are put in the MPEG-4 format, as indicated in the standard Information

technology-Generic coding of audio-visual objects = part2: visual, ISO / IEC JTC 1 / SC 29 / WG11 N3056, December 2000, before being transmitted on the network 3 of Figure 1 to the client machine 2.

 It should be noted that the encoding mode used to encode each macro-block is part of the information that is transmitted.

 FIG. 3b illustrates the decoding method implemented by the decoding unit 12 of the device 4 and by the decoding unit 22 of the client machine 2 of FIG. 1.

<Desc / Clms Page number 23>

 The decoding method will be described when it is implemented in the client machine 2 and, more particularly, at the level of the units 20, 21 and 22.

 The transposition of this description to the decoding method implemented by the device 4 of the server 1 is easily performed by replacing the units 20,21 and 22 respectively by the units 18,13 and 12.

 The client machine 2 comprises a storage unit 21 which is fed by the digital data coming from the device 4 via the connection established between the two machines through the network 3.

 This data is received by the reception unit 20 in the form of packets which contain the information necessary to decode one or more macroblocks of the current image.

 The storage unit 21 is capable of extracting information from the bitstream generated by the coding unit 11 of the device 4 and supplying them to the decoding unit 22 of the client machine 2.

 FIG. 3b illustrates the decoding of a coded macroblock either in Intra mode, as indicated by steps 322 to 326, or in Inter or predictive mode as indicated by steps 327 to 332.

 The decoding algorithm begins with a step E319 during which the coding mode of the current macro-block is recovered.

 During step E320, a test is made to know if the macroblock has been coded in Intra mode.

 If so, the step E320 is followed by a step E321 which provides for recovering, from the storage unit 21, the statistical coefficients (variable length codes) related to this macroblock.

 Step E321 is followed by a step E322 which decodes these coefficients according to the coefficient table described in the standard Information technology-Generic coding of audio-visual objects = part2: visual, ISO / IEC JTC VSC 29 / WG11 N3056, December 2000.

 The coefficients thus obtained form a set of two-dimensional data denoted QF [u] [v].

<Desc / Clms Page number 24>

 In the next step E323, an inverse quantization operation is performed on the values of the data set QF [u] [v] to obtain the two-dimensional array F "[u] [v] according to the quantization tables described in the standard Information technology- Generic coding of audiovisual abjects = part2: vssual, ISO / IEC JTC 1 / SC 29 / WG11 A / 3056, December 2000.

 In the next step E324, a saturation method is applied to the values of the coefficients of the two-dimensional array F "[u] [v] according to the standard Information technology-Generic coding of audio-visual objects = part2: visual, ISO / IEC JTC 1 / SC 29 / WG11 N3056, Oecember 2000, to obtain the two-dimensional array F '[u] [v] whose values belong to the range [-2bits per pixel + 3, 2bits per pixel + 3 - 1].

 The number of bits per pixel is preferably 8.

 During the next step E325, a control operation of the saturated coefficients is performed so as to obtain the DCT coefficients F [u] [v].

 In the next step E326, the pixel values of the current macro-block are reconstructed as a function of the DCT coefficients F [u] [v] using the decoding method described in the article Algorithms for the Discrete Cosine Transform " , E. Feig and S. Winograd, IEEE Trans., On Signal Proc, Vol 40, No. 9, September 1992.

 These reconstructed values are stored in the storage unit 21 for display on the display unit 23.

 The decoding algorithm of a macro-block is given below.

The values of the coefficients dc ~ scaler, quantiser ~ scale and the matrix] [u] [v] used in this algorithm are values defined by the standard Information technology-Generic coding of audio-visual objects = part2: visual, ISO / IEC JTC 1 / SC 29 / WG11 N3056, December 2000.

<Desc / Clms Page number 25>

else if ( (u==0) & & (v==0) & & (macroblockjntra)) { F'lu] = dc~scaler* QF [u] [v] ; } else { if (macroblockintra) { F'lu] [v] = QF [u] [v] * MO] [u] [v] * quantiser~scale)/32 ; } else { F'lu] [v] = ( ( (OF [ [] * 2) + Sign (QFMM)) * ] [u] [v] * quantiser~scale)/32 ; } } } } sum = 0 ; for (v=0 ; v < 8 ; v++) { for (u=O ; u < 8 ; u++) { if (FU] [v] > 2 bitsper~pixel + 3 F1u ] ; [ ;v ] ; = 2 bits~per~pixe + 3 - 1 ; } e ! se { bits-per * 1+3 if (F] M < -2--'") { if (F &verbar;u] [v] < -2 bits~Per~Pixel + 3) { F] [t/] =-2 Msperpixe ! + 3 } else { else { Flu] [v] = F'lu] ; } } sum = sum + Flu] [v] ; F [u] [v] = Fluez ; } } if ( (sum & 1) == 0) {

else if ((u == 0) && (v == 0) && (macroblockjntra)) {F'lu] = dc ~ scaler * QF [u] [v]; } else {if (macroblockintra) {F'lu] [v] = QF [u] [v] * MO] [u] [v] * quantize ~ scale) / 32; } else {F'lu] [v] = (((OF [[] * 2) + Sign (QFMM)) *] [u] [v] * quantize ~ scale) / 32; }}}} sum = 0; for (v = 0; v <8; v ++) {for (u = 0; u <8; u ++) {if (FU] [v]> 2 bitsper ~ pixel + 3 F1u]; [; v]; = 2 bits ~ per ~ pixe + 3 - 1;} e! se {bits-per * 1 + 3 if (F] M <-2-- '") {if (F &verbar; u] [v] <-2 bits ~ Per ~ Pixel + 3) {F] [t /] = -2 Msperpixe! + 3} else {else {Flu] [v] = F'lu];}} sum = sum + Flu] [v]; F [u] [v] = Flue;}} if ((sum & 1) == 0) {

<Desc / Clms Page number 26>

} }
La méthode de transformation en cosinus discret d'un pixel f (x, y) d'un bloc de taille N x N est donnée par l'équation suivante :

avec v=0, 1, 2,... N-1, où x, y sont les coordonnées spatiales, u, v sont les coordonnées dans le domaine transformé, et

}}
The method of discrete cosine transformation of a pixel f (x, y) of a block of size N x N is given by the following equation:

with v = 0, 1, 2, ... N-1, where x, y are the spatial coordinates, u, v are the coordinates in the transformed domain, and

Each pixel is represented by n bits (to represent one of the components Y, U or V) and the transformed coefficients are represented by (n + 4) bits. The amplitude of the DCT coefficients is in the range [- 2n + 3: + 2n + 3 ~ 1].

 Preferably, a fast method of DCT coding / decoding is used.

Discrete Cosine Transform", E. Feig and S. Winograd, IEEE Trans. On Signal Proc., vol. 40, No. 9, September 1992, permet de coder/décoder des macro- blocs de taille 2m (ici m=3) selon l'équation suivante :

où Yest le vecteur des coefficients DCT de dimension 8 etCJa matrice 8*8 de coefficients DCT. This method, described in the article "Fast Algorithms for the

Discrete Cosine Transform, E. Feig and S. Winograd, IEEE Trans., On Signal Proc, Vol 40, No. 9, September 1992, allows coding / decoding of macroblocks of size 2m (here m = 3). according to the following equation:

where Y is the vector of DCT coefficients of dimension 8 and CJa matrix 8 * 8 of DCT coefficients.

 This matrix can be factorized and, after some algebraic manipulations, this matrix is equal to Cg = * jSg, where / is a matrix of permutation of the signs of dimension 8 * 8, Kg and Bg are matrices of

<Desc / Clms Page number 27>

 dimension coefficients 8 * 8 which have been defined to obtain a fast encoder / decoder.

 Subsequent to step E326, step E333 provides for storage and visualization of the decoded macroblocks from units 21 and 23. The current image is reconstructed once all the macroblocks of this image are decoded and viewed.

 Returning to step E320, if the result of the test is negative, it means that the macroblock has been coded in Inter or predictive mode.

 During the next step E327, information is then retrieved from the digital data transmitted by the device 4 and stored in the storage unit 21 of the client machine 2.

 This information is specific to the coding of the macroblock in Inter mode and is, for example, the statistical coefficients (variable length codes) related to this macroblock.

 Step E327 is followed by a step E328 which provides for decoding these statistical coefficients so as to obtain the differential motion vector.

During the next step E329, the motion vector is reconstructed from the differential motion vector and the motion vector (Px, Py) of the macroblock of the previous image disposed at the same spatial position.

if ( (f == 1) Il (horizontal~mv~data == 0)) MVDx = horizontal~mv~data ; else { MVDx = ( (/4bs (horizontaLmvdata)-1) * f) + horizonta ! mvresidua) + 1 ; The algorithm used is the following: r ~ size = vop ~ fcode-1 f = 1 rs / ze high (32 * f) - 1; low = ((-32) * f); range = (64 * f);

if ((f == 1) Il (horizontal ~ mv ~ data == 0)) MVDx = horizontal ~ mv ~ data; else {MVDx = ((/ 4bs (horizontaLmvdata) -1) * f) + horizonta! mvresidua) + 1;

<Desc / Clms Page number 28>

MVDy = vertical~mv~data ; else { MVDy = ( (Abs (vertical~mv~data) - 1) * f) + vertical~mv~residual + 1 ; if (vertical~mv~data < 0)
MVDy =-MVDy ; } MVx = Px + MVDx ; if (MVx < low)
MVx = MVx + range ; if (MVx > high)
MVx = MVx-range ; MVy = Py + MVDy ; if (M\/ < low)
MVy = MVy + range ; if (MVy > high)

MVy = MVy-range ;
Les paramètres du train binaire sont tels que les composantes MVDx et MVDy du vecteur de mouvement différentiel, appartiennent à l'intervalle [low; high]. if (horizontal ~ mv ~ data <0)
MVDx = -MVDx; } if ((f == 1) Il (vertical ~ mv ~ data == 0))

MVDy = vertical ~ mv ~ data; else {MVDy = ((Abs (vertical ~ mv ~ data) - 1) * f) + vertical ~ mv ~ residual + 1; if (vertical ~ mv ~ data <0)
MVDy = -MVDy; } MVx = Px + MVDx; if (MVx <low)
MVx = MVx + range; if (MVx> high)
MVx = MVx-range; MVy = Py + MVDy; if (M \ / <low)
MVy = MVy + range; if (MVy> high)

MVy = MVy-range;
The parameters of the bitstream are such that the MVDx and MVDy components of the differential motion vector belong to the interval [low; High].

 The extreme values of this interval are chosen during the implementation of the decoder.

 Moreover, the components of the reconstructed motion vector, MVx and MVy, also belong to this interval. The values that define this interval for motion vectors are defined by the

<Desc / Clms Page number 29>

standard Information technology-Generic coding of audio-visual objects = part2 : Msua/,/SO//FCJ7'C /SC 29AVG M3056, Decem & e/-2000.

standard Information technology-Generic coding of audio-visual objects = part2: Msua /, / SO // FCJ7'C / SC 29AVG M3056, Decem & e / -2000.

horizontal~mv~residual et vertica! mv~residual sont extraites du train binaire. Note that r ~ size, f, MVDx, MVDy, high, low and range are variables, while the data horizontatmvdata, verticatmvdata,

horizontal ~ mv ~ residual and vertica! mv ~ residual are extracted from the binary train.

The variable vop ~ fcode refers to the prediction mode used which is extracted from the bit stream.

 During this same step E329, the macroblock of the preceding image is compensated in motion, that is to say that the macroblock is displaced from the motion vector that has just been reconstructed.

 During the next step E330, the information specific to the coding of the macroblock in Intra mode is retrieved and, in particular, the statistical coefficients that represent the difference macroblock.

 During the next step E331, the difference macro-block is reconstructed from these statistical coefficients according to the Intra decoding method described above, during the execution of the steps E322 to E326.

 Step E331 is followed by a step E332 which calculates the sum of the compensated macroblock of step E329 and the difference macroblock obtained in step E331.

 The step E332 is followed by a step E333 which organizes the storage of the decoded macroblocks in the unit 21.

 The image is then displayed on the display unit 23 once all the macroblocks have been decoded.

 FIG. 4a illustrates the algorithm detailing the various operations performed during step E208 of FIG. 2.

 More particularly, this algorithm comprises different instructions or portions of software code corresponding to steps of the data encoding parameter determination method according to the invention.

 The computer program that is based on this algorithm is stored in the temporary data storage unit 13 of Figure 1 and executed by the unit 17 under the control of the control unit.

<Desc / Clms Page number 30>

 This program is part of the program "Progr" above with reference to Figure 2 and is also stored in the apparatus of Figure 5.

 The execution of the computer program based on this algorithm will implement the method for adapting one or more coding parameters that will be used for coding the data transmitted to the client machine, according to the memory capacity of the latter. .

 Moreover, this method can also make it possible to adapt this or these coding parameters to the resolution of the screen of the client machine 2 as well as to the resources available on the communication network 3, namely the bandwidth.

 The algorithm of FIG. 4a begins with a step E400 during which a recovery of the context in which the data will be transmitted from the device 4 of the server 1 to the client machine 2 (FIG. 1) is carried out.

 More particularly, the recovery of this context consists in obtaining, from the temporary storage unit 13: the total memory capacity M available in the client machine 2, the resolution of the screen of this machine, the band BP pass-through available on the communication network 3 for transmitting the video data, and - the encoding parameter or parameters used to encode the video data stored in the storage unit 5.

 The coding parameters used are, for example, the size of the file containing the video data, denoted BRo, the resolution of the image ro and the number of images per second fro.

* Note that when video data is stored in YUV format, encoding settings such as file size, image resolution, and frame rate are necessarily retrieved during this step E400.

 The other encoding parameters used to encode this video data are then values that are chosen by default.

 During the next step E401, the size of the video file BR that will be transmitted is determined.

<Desc / Clms Page number 31>

 In this step, two cases are envisaged depending on whether the size of the file BR0 stored in the storage unit 5 of FIG. 1 is greater or smaller than the available bandwidth on the communication network 3.

 To do this, the value of the maximum authorized bit rate BR in the communication network 3 and which is allocated to the rate control module during the execution of step E314 of FIG. 3a is taken into account.

The size of the video file BR that can be transmitted is then provided by the following equation:
BR = min (BRo, BP).

 During the next step E402, the values contained in two tables are initialized: a first table, of dimension R, in which are stored the admissible resolution values of an image, and a second table, of dimension FR, in which different values of the number of frames per second are stored.

 More particularly, the resolution values of the image that are contained in the first table are admissible in the sense that these resolution values are less than or equal to the resolution of the display unit 23 of the client machine 2 (FIG. ) and which is stored in the storage unit 21.

 For example, a first table whose first value is equal to the resolution of the video data stored in the storage unit 5 of FIG. 1 and constituting the original version of the video mentioned above, in the case, of course, where this value is considered eligible.

 The other allowable values in this table are selected beforehand and, in the case where the dimension of the array is three, the last two values correspond to a resolution of 352 columns and 288 rows, for one, and 176 columns and 144 columns. lines, for the other.

 The second array containing different values of the number of frames per second is, for example, also of dimension three and the first value of this array is equal to the number of frames per second of the original version of the video stored in the storage unit 5 of FIG.

<Desc / Clms Page number 32>

 The other two values are respectively 20 and 10 frames per second.

 More particularly, during this step, we proceed to the initialization of two index variables, fr and r, respectively corresponding to the second and the first table above.

 Each index variable is initialized to the first value contained in the corresponding table.

 During the next step E403, the total memory capacity ME which is necessary for the decoding operations of a given number of coded data elementary units is determined, operations which are described with reference to FIG. 3b.

 This memory capacity is in particular that which is necessary for storing decoded macro-data blocks for display, as explained above with reference to step E333 of FIG. 3b.

 For example, the given number of macroblocks will correspond to the number of macroblocks that make up an image.

 In the exemplary embodiment, the macroblocks constituting an image are coded in Intra mode or in Inter mode.

 Since the decoding of a macro-block which has been coded in Inter mode requires the decoding of a coded difference macro-block in Intra mode, the so-called basic memory capacity necessary to decode a coded macroblock in Inter mode is no longer necessary. higher than that required to decode a coded macroblock in Intra mode.

 In this case, the total memory capacity needed to decode an image will be equal to that needed to decode an Inter coded image in which all macroblocks are coded in Inter mode, which is the worst case.

 This total memory capacity is provided by the following equation: ME = MB * MElnter, where MB is the number of macroblocks of an image and MElnter the estimate of the elementary memory capacity necessary to decode the coded macroblock into Inter mode

<Desc / Clms Page number 33>

 The estimation of the elementary memory capacity MEnter of a macroblock is performed by determining the memory capacity necessary to construct the compensated macroblock and the macroblock of difference mentioned with reference to FIG. 3b.

 The decoded operations of the compensated macroblock described in steps E327, E328 of FIG. 3b begin first with the statistical decoding of the differential motion vector.

 The VLD statistical decoding (step E328) requires the storage of the table defined in the standard Information technology-Generic coding of audio-visual objects = part2: visual, ISO / IEC JTC 1 / SC 29 / WG11 N3056, December 2000.

 In order to be able to build the compensated macroblock during step E329, it is necessary to have previously stored in memory the values of the macroblock of the previous image.

 The decoding of the difference macro-block (E330) firstly provides for statistically decoding the coefficients according to step E321 of FIG. 3b (Intra mode decoding).

 VLD decoding, step E322 requires the storage of two tables defined in the standard Information technology-Generic coding of audio-visual abjects = part2: visual, ISO / IEC JTC 1 / SC 29 / WG11 N3056, December 2000.

 One of the tables corresponds to the luminance coefficients and the other table to the chrominance coefficients.

 The inverse quantization described by the step E323 involves, for its part, a weighting matrix W [1] [u] [v].

 In addition, the reconstructed macroblock needs to be memorized until all macroblocks of the image are decoded.

 The table below details the elementary memory capacity needed to decode a macro-block that has been encoded according to the most memory-intensive encoding mode, namely the Inter mode:

<Desc / Clms Page number 34>

<Tb>
<tb> Processing <SEP> memory
<tb> VLD <SEP> 1112
<tb> Macro-block <SEP> previous <SEP><SEP> 16 * 16 * 8 <SEP> = <SEP> 2048
<tb> Table <SEP> VLD <SEP> luminance <SEP> 218
<tb> Table <SEP> VLD <SEP> chrominance <SEP> 237
<tb> Matrix <SEP> of <SEP> weighting <SEP> 16 * 16 * 8 <SEP> = <SEP> 3072
<tb> Macro-block <SEP> backup <SEP> 16 * 16 * 8 <SEP> = <SEP> 2048
<tb> Total <SEP> 8735
<Tb>

In the example just described, the total memory capacity of an image was determined as a function, on the one hand, of the highest elementary memory capacity among the different elementary memory capacities determined for each of the modes of memory. coding and, secondly, the number of elementary units (macroblocks) constituting the image.

 The calculations making it possible to arrive at the total memory capacity ME are thus simplified.

 However, it is also possible to determine this total memory capacity ME as closely as possible, taking into account, on the one hand, the different elementary memory capacities that have been obtained respectively for the various coding modes used and, on the other hand, the number of elementary data units (macroblocks) coded according to each of these coding modes.

 It should be noted that the number of different coding modes may be greater than two, or even be unique.

 Returning to the algorithm of FIG. 4a, since the total memory capacity ME has been determined, the step E403 is followed by a step E404.

 During this step, a test is performed to determine whether the total memory capacity ME thus obtained for decoding an image that is part of a video requested by the user and encoded with the current coding parameters is less than or equal to a predetermined memory capacity (M).

<Desc / Clms Page number 35>

 It will be noted that the current coding parameters are, during the first execution of the loop of the algorithm, those initialized in step E402.

 Subsequently, the parameter values are provided by steps E409 and E411.

 In particular, the memory capacity ME that has just been determined is compared with the memory capacity M of the client machine 2 to ensure that the memory capacity M of this client machine is sufficient to store video data. required by the user and which must be decoded.

 If so, the step E404 is followed by a step E405 during which the coding of the video data is carried out according to the coding parameters recovered in step E400, as well as the current values of the number of images per second (fr) and the resolution of the image (r), these values having been initialized in step E402.

 In the next step E406, the coded video data is decoded during the execution of step E405, so as to obtain a version of the video in YUV format.

 In the next step E407, the quality of the video data thus compressed is quantized.

où q, désigne un pixel de l'image décodée, p, un pixel de l'image originale, N le nombre de pixels de l'image et Y la dynamique du signal original, correspondant dans le cas présent à huit bits. For example, the difference in quality between two videos will be estimated using the notion of signal-to-noise ratio (PSNR) and provided by the following equation:

where q, denotes a pixel of the decoded image, p, a pixel of the original image, N the number of pixels of the image and Y the dynamics of the original signal, corresponding in this case to eight bits.

<Desc / Clms Page number 36>

 The PSNR provides a quantitative measure of the quality of the reconstructed image. It is determined from the ratio of the maximum amplitude of the signal to the squared error.

 It should be noted that other methods of estimating a quality can of course be envisaged.

 For example, it is possible to establish a model by determining the quality of different test videos for which the coding parameter or parameters are varied.

 Then, for each video to be transmitted, the coding parameter or parameters according to the invention are adapted and, depending on the type of the video and the aforementioned model, the quality of the video coded with this or these parameters is deduced therefrom.

 Regarding the PSNR, it should be noted that the number of pixels of the original image must be the same as that of the decoded image.

 However, this is not always true insofar as the method according to the invention makes it possible to modify the coding parameter that is the resolution of the image.

 Therefore, it will be necessary if there is a difference between the number of pixels of the original image and that of the decoded image to introduce a spatial subsampling process, in order to reduce the resolution of the image of the image. origin to that of the decoded image.

 The PSNR value is determined for each frame of the video.

 We will thus have a good quality image by basing the choice of the version of the data to be sent on the value of the PSNR.

 It should be noted that it would be possible to determine the value of the PSNR for each elementary unit of data constituting an image and then to sum it to deduce the PSNR value of the image considered.

 By calculating the value of the PSNR on each elementary unit of data (macro-block), one will find macro-blocks which will be of good quality and others of less good quality and it will then be necessary to average the PSNR values of the macro-blocks. blocks of the image.

<Desc / Clms Page number 37>

 This method, however, causes more calculations than calculating the PSNR value for the image overall.

 Since we are interested in the determination of the quality on a given video segment, we will calculate an average value of PSNR on this segment.

 Therefore, a time sub-sampling process is used to bring the number of frames per second of the original video back to the number of frames per second of the decoded picture.

 Once the average quality of a given video segment has been estimated by comparing the decoded images of this segment which have been previously coded with the above coding parameters with respect to the corresponding uncoded images, it is memorized the temporary storage unit 13 of FIG. 1, the value of the average PSNR, the digital data constituting the video compressed with the coding parameters.

 In the next step E408, a test is made to determine whether the current resolution of the image, r, is the last value in the first table of acceptable resolutions of the image.

 If not, step E408 is followed by a step E409 in which the current resolution of the image is incremented to the next value in the table under consideration.

 It will be noted that during the execution of the first loop of the algorithm of FIG. 4a, the value r initialized during the step E402 corresponds to the first value of the first array and thus this step E409 is executed.

 The step E403 described above is then performed again in order to determine the total memory capacity ME necessary to decode the video data required by the user with the determined coding parameters, taking into account the above.

 The coding parameter corresponding to the resolution of the image has just been modified, while the other coding parameter corresponding to the number of images per second remains unchanged.

<Desc / Clms Page number 38>

 It is thus understood that, during the execution of this loop of the algorithm, all the resolution values of the image indicated in the first table mentioned above are selected and lead, in each case, to the determination of a particular memory capacity.

 Note that in the example described here, only one coding parameter is changed each time although it is also possible to influence the other encoding parameter that is the number of frames per second.

 For this combination of coding parameters in which only one parameter varies, the quality of each picture and, more generally, of a given video data segment is estimated.

 However, it would also be possible to simultaneously modify several different coding parameters within one and the same combination of coding parameters, and for each combination of coding parameters thus modified, the quality of the image is estimated.

 Note that, when performing step E404, if the test performed is negative, then step E404 is immediately followed by step E408 described above.

 In general, the algorithm of FIG. 4a makes it possible to take a decision as to whether an encoding parameter can be used to code the video data required by the user, taking into account: the total memory capacity determined during step E403 according to this parameter and - the comparison of this memory capacity with a predetermined memory capacity (total memory capacity of client machine 2), during step E404.

 Returning to the algorithm of FIG. 4a, when the test performed in step E408 is positive, this means that all the image resolution values provided by the first aforementioned table have been selected, then this step is followed by a step E410 during which another test is performed.

 The purpose of this test is to determine if the current value of the number of frames per second (fr) is the last value of the second array

<Desc / Clms Page number 39>

 in which are stored different values of the number of frames per second.

 If not, this step is followed by a step E411 in which the current value of the number of frames per second is incremented to select another value of the second array.

 Step E403 is then executed again and it is possible to determine, for this new coding parameter, the total memory capacity necessary for decoding the video data required by the user with this coding parameter and the other current coding parameters mentioned above.

 Note that in this case, the value of the resolution of the image corresponds to the last value of the first array and the different values of the number of images per second appearing in the second array will successively be selected, while the value of the resolution of the image will remain the same.

 However, it is possible to modify the algorithm of FIG. 4a so that different other combinations of coding parameters can be envisaged.

 For example, the different successive values of the number of frames per second in the second table can be combined with the first and then the second value of the resolution of the image in the first table.

 Note also that the first loop consisting of steps E403 to E409 could also take into account the encoding parameter that is the number of frames per second of the video requested by the user instead of taking into account the resolution of the picture.

 Returning to step E41 0, when the test is positive, this means that all the values of the number of frames per second in the second table have been selected, and then the next step E412 is selected.

 During this step, the various versions of compressed video data are sorted in descending order according to their value of PSNR.

<Desc / Clms Page number 40>

 For example, the compressed video data version that will be transmitted from the device 4 according to the invention to the client machine 2 of FIG. 1 will be the one that has the highest PSNR.

 Step E412 makes it possible to select the best quality among all those that have been determined.

 This ends the algorithm of Figure 4a.

 It will be noted that the step E209 of the algorithm of FIG. 2 is then executed to transmit the video data required by the user and which has been adapted to the memory capacity of the client machine 2.

 It should be noted that the algorithm of FIG. 4a can also take into account a single coding parameter or, on the contrary, take into account other coding parameters such as, for example, the number of images that have been coded in the same mode. inter.

 Thus, one or more coding parameters are adapted for encoding the video data or, more generally, multimedia data required by the user, so that the client machine 2 is able to decode this compressed data, taking into account the capacity memory available to the user.

 The coding parameter (s) on which the invention (s) intervene must of course have an influence on the memory capacity necessary for the decoding of the multimedia data requested by the user.

 It should be noted that if two modes of coding intervene in the coding of the data, the invention makes it possible to be interested only in one of the coding parameters which exerts an influence on only one of the modes of coding.

 FIG. 4b illustrates an alternative embodiment of the algorithm of FIG. 4a.

 The algorithm of FIG. 4b differs from that of FIG. 4a by the steps E430, E434 and E435 which will be detailed below.

 The steps E431, E432, E433 and E436 to E443 corresponding, in turn, respective steps E401, E402, E403 and E405 to E412, they will not be described again.

<Desc / Clms Page number 41>

 The algorithm of FIG. 4b comprises various instructions or portions of software code corresponding to steps of the method of determining a data coding parameter according to the invention.

 The computer program which is based on this algorithm is stored in the temporary data storage unit 13 of FIG. 1 and executed by the unit 17 under the control of the control unit 19.

 This program is part of the program "Progr" above with reference to Figure 2 and is also stored in the apparatus of Figure 5.

 The execution of the computer program based on this algorithm will implement the method for adapting one or more coding parameters that will be used for the coding of the data transmitted to the client machine, as a function, on the one hand, of the memory capacity of the client machine and, on the other hand, the computing capacity of the latter.

 Moreover, this method can also make it possible to adapt this or these coding parameters to the resolution of the screen of the client machine 2, as well as to the resources available on the communication network 3, namely the bandwidth.

 During the execution of the first step E430, the recovery of the context in which the data will be transmitted from the device 4 of the server 1 to the client machine 2 (FIG. 1) comprises, in addition to that described in step E400 of FIG. FIG. 4a, obtaining the computing capacity C of the client machine 2.

 During the step E434, the computation complexity, denoted 0, of the decoding process of a sequence of images, also called image segment, which corresponds to a number of coded images for an interval of predetermined time.

 In the example described here, the temporal resolution being a coding parameter, the time interval is fixed for example at 1 s and the number of coded images during this time interval is counted.

 The number of images is for example equal to thirty.

 This calculation complexity 0 is also called decoding complexity, and its determination corresponds to the determination of the total number

<Desc / Clms Page number 42>

 operations that are to be performed during a predetermined time interval in order to decode a plurality of elementary units of coded data. These elementary units are, for example, the macroblocks defined above.

 To determine this decoding complexity, we will look at a segment composed of Nlntra + N, nter images.

où : est lue nombre d'images codées en mode Intra pour le segment considéré, le segment correspond à un nombre donné d'images par seconde, - NInter est le nombre d'images codées en mode Inter ou prédictif pour le segment considéré, - MBIntra est le nombre de macro-blocs codés en mode Intra dans une image codée en mode Intra, - OIntra est la complexité de décodage d'un macro-bloc codé en mode Intra, - MBest le nombre de macro-blocs codés en mode Intra dans une image codée en mode Inter, - MB'Inter est le nombre de macro-blocs codés en mode Inter dans une image codée en mode Inter, et - OInter est la complexité de décodage d'un macro-bloc codé en mode Inter. The complexity of such a segment is provided by the following equation:

where: is read number of images coded in Intra mode for the segment in question, the segment corresponds to a given number of images per second, - NInter is the number of coded images in Inter or predictive mode for the segment considered, - MBIntra is the number of macroblocks encoded in Intra mode in an Intra mode encoded image, - OIntra is the decoding complexity of an Intra mode encoded macroblock, - MB is the number of Intra mode coded macroblocks. in an Inter mode coded picture, - MB'Inter is the number of Inter mode encoded macroblocks in an Inter mode coded picture, and - OInter is the decoding complexity of a coded macroblock in Inter mode.

 It can thus be seen that the determination of the decoding complexity of the image segment defined above passes through a preliminary step of analysis of the coding mode used to code each elementary unit of data (macroblock). Next, it is a question of determining, for each elementary unit of data which has been coded according to a different coding mode, the decoding complexity of this elementary unit of data, that is to say the basic number of operations. that it is necessary to perform to proceed to its decoding.

<Desc / Clms Page number 43>

 Determining the decoding complexity of the abovementioned segment then goes through a step during which the number of elementary data units coded according to each of the different coding modes is identified and, knowing the decoding complexity for each of these modes of coding, we deduce the total number of operations that are necessary to perform during the predetermined time interval to decode the plurality of elementary units of coded data constituting the segment defined above.

 In the example described here, we are interested in a first step in the encoding mode of an image then, depending on whether it is encoded in Intra mode or Inter mode, we proceed as follows: - If image is encoded in Intra mode, then it is enough to take into account the decoding complexity of a macro-block and the number of macroblocks contained in this image, then to multiply the obtained figure corresponding to the decoding complexity of a image encoded in intra mode by the number of N1ntra images coded according to this mode.

 If the image is coded in Inter mode, then it is necessary to search inside each image for the number of macroblocks coded according to the Intra or Inter coding mode, then obtain for each image, on the one hand, the decoding complexity for macro-blocks coded according to the Intra coding mode and, on the other hand, the decoding complexity for the coded macro-blocks according to the Inter coding mode.

 It is then necessary to sum these decoding complexities in order to obtain the total decoding complexity of an image coded in Inter mode. It is then necessary to multiply this number by the number of images coded in Niter mode.

 It should be noted that the determination of the decoding complexity of an elementary data unit such as a macroblock varies according to whether it is coded in Intra mode or in Inter mode.

 As described below, the determination of the decoding complexity of an elementary data unit will be detailed depending on whether the coding takes place in Intra mode or Inter mode.

<Desc / Clms Page number 44>

1) Encoding in Intra mode
Decoding being a block of data by data block The decoding complexity of a macroblock is equal to the complexity related to the decoding of the four blocks that compose it.

 More particularly, the complexity of decoding a block is determined by counting the number of elementary operations involved in the various algorithms used in the decoding process.

 The complexity values related to DCT decoding are provided by the article "Fast Algorithms for the Discrete Cosine Transform", E. Feig and S.

Winograd, IEEE Trans. On Signal Proc., Vol. 40, No. 9, September 1992.

Table 1 below details the determination of the decoding complexity of the different algorithms involved in the decoding of a block from the different elementary operations.

<Tb>
<Tb>

Processing <SEP> Test <SEP> Multiplication <SEP> Addition <SEP> Assignment
<tb> VLD <SEP> 64
<tb> Inverse <SEP> Q <SEP> 192 <SEP> 256 <SEP> 192
<tb> Saturation <SEP> 128 <SEP> 64
<tb> Control <SEP> 2 <SEP> 64 <SEP> 65
<tb> Inverse <SEP> DCT <SEP> 94 <SEP> 454 <SEP> 64
<tb> Total <SEP> 322 <SEP> 350 <SEP> 582 <SEP> 385
<Tb>

Table 1
Preferably, all these operations are considered elementary except for the multiplication which is considered equivalent to two elementary operations.

 Thus, the total complexity of decoding a macro-block (luminance components) coded in Intra mode is 4 * 1989 = 7956 elementary operations.

2) Coding in Inter mode
In the case where the macroblock is coded in Inter mode, it is necessary to add to the decoding complexity in Intra mode, the linked decoding complexity,

<Desc / Clms Page number 45>

 on the one hand, to the reconstruction of the motion vector from the differential motion vector which will be transmitted to the client machine 2 of FIG. 1 and, on the other hand, to the reconstruction of the macro-block (reconstruction of the macro difference-block and addition of this macro-block with the predicted macro-block).

The determination of the decoding complexity of these processes is detailed in table 2 below which provides, for each process, the number of elementary operations that occur.

<Tb>
<Tb>

Processing <SEP> Test <SEP> Multiplication <SEP> Addition <SEP> Assignment
<tb> VLD <SEP> 2
<tb> Reconstruction <SEP> vector <SEP> 7 <SEP> 5 <SEP> 13 <SEP> 11
<tb> movement
<tb> Reconstruction <SEP> Macroblock <SEP> 1288 <SEP> 1400 <SEP> 2328 <SEP> 1540
<tb> difference
<tb><SEP> Construction of the <SEP> Macro Block <SEP> 512 <SEP> 256
<tb> predicts
<tb> Rebuild <SEP> of <SEP> Macroblock <SEP> 256 <SEP> 256
<tb> Total <SEP> 1295 <SEP> 1405 <SEP> 3109 <SEP> 2065
<Tb>

Table 2
Thus, for the Inter or Predictive mode, the total complexity of decoding a coded macro-block in Inter mode is 7874 elementary operations.

 Once the decoding complexity of a segment composed of Nlntra + Nlnter images has been determined, step E434 is followed by a step E435.

 During this step, a test is performed to determine whether, on the one hand, the total memory capacity ME obtained for decoding an image coded with the current coding parameters is less than or equal to a predetermined memory capacity. M and, on the other hand, the decoding complexity 0 obtained is less than or equal to a predetermined decoding complexity C.

<Desc / Clms Page number 46>

 The predetermined memory capacity and the predetermined decoding complexity respectively correspond, in the case of a client-server architecture, to the memory capacity available in the client machine 2 of FIG. 1 and to its computing capacity.

 By thus determining one or more coding parameters in a manner adapted to this memory capacity available in the client machine 2 and to its computing capacity, the device 4 of FIG. 1 will be able to transmit to the machine 2 the data required by the user. and coded with this or these parameters, that is to say in a manner adapted to the aforementioned characteristics of the machine 2.

 This adaptation of the content of the transmitted data takes more account of the specificities of the client machine 2 than the adaptation which is provided in FIG. 4a.

 With reference to FIG. 5, an example of a programmable apparatus embodying the invention is described. This apparatus is adapted to process digital multimedia data for transmission.

 According to the embodiment chosen and shown in FIG. 5, an apparatus embodying the invention is for example a microcomputer 500 or a workstation connected to different peripherals, for example a digital camera 501 (or a scanner, or any means of acquisition or image storage) connected to a graphics card and providing the device with data to be adapted and transmitted.

 The device 500 comprises a communication bus 502 to which are connected: a central processing unit 503 (microprocessor), which performs the function of the control unit 19 of FIG. 1, a read-only memory 504, which may comprise the program "Progr", - a RAM 506, having registers 507 adapted to record variables created and modified during the execution of the aforementioned program and in particular the variables r, fr, BR, BRo and PSNR mentioned with reference to the figures preceding,

<Desc / Clms Page number 47>

 a screen 508 making it possible to display the data to be processed and / or to serve as a graphical interface with the user who can interact with the program according to the invention, using a keyboard 510 or any other means such as that a pointing device not shown, such as for example a mouse or an optical pencil, a hard disk 512 which may comprise the "Progr" program mentioned above, a floppy disk drive 514 adapted to receive a floppy disk 516 and to read from it or to write data processed or to be processed according to the invention, - a communication interface 518 connected to a communication network 520, for example an Internet network, or Ethernet, or IEEE1394 or 802.1 wireless type, the interface being able to transmit and receive data.

 In the case of audio data, the apparatus further comprises an input / output card connected to a microphone which are both unrepresented.

 The communication bus allows communication and interoperability between the various elements included in the microcomputer 500 or connected to it. The representation of the bus is not limiting and, in particular, the central unit is able to communicate instructions to any element of the microcomputer 500 directly or through another element of the microcomputer 500.

 The program executable code denoted "Progr" allowing the programmable apparatus to implement the context recovery (FIG. 2), coding / decoding (FIGS. 3a and 3b) and adaptation (FIGS. 4a and 4b) recovery methods. data according to the invention can be stored for example in the hard disk 512 or in the read only memory 504 as shown in FIG. 5.

 Although only one program is identified, it is possible to have several programs or sub programs to implement the invention.

 According to one variant, the floppy disk 516 can contain compressed and stored data as well as the executable code of the program or programs

<Desc / Clms Page number 48>

 according to the invention which, once read by the apparatus 500, will be stored in the hard disk 512.

 In the second variant, the executable code of the program or programs may be received via the communication network 520, via the interface 518, to be stored identically to that described above.

 Floppies can be replaced by any information medium such as, for example, a compact disc (CD-ROM) or a memory card. In general, an information storage means, readable by a computer or by a microprocessor, integrated or not, possibly removable, is adapted to store a program whose execution allows the implementation of the method according to the invention.

 More generally, the program can be loaded into one of the storage means of the device 500 before being executed.

 The central unit 503 will control and direct the execution of the instructions or portions of software code of the program or programs according to the invention, which instructions are stored in the hard disk 512 or the read only memory 504 or else in the other storage elements. supra. When powering up, the program or programs that are stored in a non-volatile memory, for example the hard disk 512 or the ROM 504, are transferred into the RAM RAM 506 which will then contain the executable code of the program or programs. according to the invention, as well as registers for storing the variables necessary for the implementation of the invention.

 It should be noted that the data processing apparatus comprising a device according to the invention may also be a programmed apparatus.

 This device then contains the code of the computer program or programs for example frozen in a specific application integrated circuit (ASIC).

 It will be noted that the apparatus of FIG. 5 which contains the program "Progr" previously mentioned can be, in the context of a communication architecture of the client-server type, the device 4 of FIG.

Claims (38)

A method for determining at least one coding parameter of multimedia digital data organized in elementary data units intended to be coded according to at least one coding mode, characterized in that said method comprises the following steps: determination of an elementary memory capacity necessary for the decoding of each elementary unit of data encoded according to a different encoding mode, - determining a total memory capacity necessary for the decoding of a given number of elementary units of data coded according to a on the one hand, at least one elementary memory capacity determined for the decoding of an elementary unit of data coded according to an encoding mode and, on the other hand, the number of elementary data units coded according to each of the coding modes, - decision as to the determination of at least one coding parameter of the data according to the memory capacity tot ale previously determined and a predetermined memory capacity.  2. Method according to claim 1, characterized in that, several coding modes being used for the coding of the given number of elementary data units, the step of determining the total memory capacity is more particularly carried out as a function of on the one hand, the elementary memory capacity determined for the decoding of each elementary data unit coded according to a different coding mode and, on the other hand, the number of elementary data units coded according to each of the coding modes.  3. Method according to claim 1, characterized in that, several coding modes being used for the coding of the given number of elementary data units, said method comprises an additional step of selecting the highest elementary memory capacity among the different basic memory capacities determined for each of the coding modes.  4. Method according to claim 3, characterized in that the step of determining the total memory capacity is more particularly carried out in <Desc / Clms Page number 50>  function, on the one hand, the highest elementary memory capacity previously selected and, on the other hand, the given number of elementary units of coded data.  5. Method according to one of claims 1 to 4, characterized in that the steps of determining the elementary memory capacity for each mode of coding, determination of the total memory capacity and decision are performed in a first communication device connected to a second communication apparatus by a communication network.  6. Method according to claim 5, characterized in that, prior to the decision step, the method comprises a step of comparing the total memory capacity determined for decoding the given number of elementary units of coded data and the capacity. memory available in the second communication apparatus for decoding this data.  7. Method according to claim 6, characterized in that, prior to the comparison step, the method comprises a step of obtaining by the first communication device the available memory capacity in the second communication device.  8. Method according to one of claims 1 to 7, characterized in that it further comprises the following steps: - determination of the basic number of operations to be performed to decode each elementary unit of data encoded according to a different coding mode determining the total number of operations to be performed during a predetermined time interval for decoding a plurality of elementary data units coded as a function, on the one hand, of the previously determined elementary number and, on the other hand, of the number of elementary units coded according to each of the coding modes.  9. Method according to claim 8, characterized in that the decision as to the determination of said at least one coding parameter is also taken according to the total number of operations previously determined and a total number of predetermined operations. <Desc / Clms Page number 51>  The method according to claims 5 and 9, characterized in that the steps of determining the basic number of operations, determining the total number of operations during a predetermined time interval and decision are performed in the first communication apparatus. .  11. The method of claim 10, characterized in that, prior to the decision step, the method comprises a step of comparing the total number of operations determined during the time interval and the total number of operations that could be carried out during the same time interval in the second communication apparatus.  12. Method according to claim 11, characterized in that, prior to the comparison step, the method comprises a step of obtaining by the first communication device the total number of operations that can be performed by the second communication device. .  13. Method according to one of claims 5 to 7, characterized in that the decision as to the determination of said at least one coding parameter is also taken according to characteristics of the communication network.  14. Method according to one of claims 1 to 13, characterized in that it comprises a step of estimating the quality of at least one elementary unit of data by comparing said at least one elementary unit of data coded with the determined coding parameter and said at least one uncoded unit of data.  15. Method according to claim 14, characterized in that, when several coding parameters have been determined, the method comprises the following steps: estimation of the quality of at least one elementary unit of data for each combination of coding parameters determined, - selection of the best quality among the various qualities estimated for the different coding parameters.  16. The method of claim 14 or 15, characterized in that it comprises the following steps: coding of at least one elementary unit of data with the determined coding parameter, <Desc / Clms Page number 52>  decoding said at least one coded elementary data unit; estimating the quality of said at least one elementary data unit by comparing this at least one uncoded elementary data unit with that previously decoded unit.  17. Method according to one of claims 1 to 16, characterized in that the data constitutes a video and the given number of elementary data units corresponds to an image of the video.  18. The method as claimed in claim 17, characterized in that the elementary data unit is a macro-block of data.  19. A device for determining at least one coding parameter for multimedia digital data organized in elementary data units intended to be coded according to at least one coding mode, characterized in that said device comprises: means for determining an elementary memory capacity necessary for the decoding of each elementary unit of data coded according to a different encoding mode; means for determining a total memory capacity necessary for decoding a given number of elementary units of data coded according to, on the one hand, at least one elementary memory capacity determined for the decoding of an elementary unit of data coded according to an encoding mode and, on the other hand, the number of elementary data units coded according to each of the modes. coding means - decision means for determining at least one coding parameter of the data as a function of the total memory capacity previously determined and a predetermined memory capacity.  20. Device according to claim 19, characterized in that, several coding modes being used for the coding of the given number of elementary data units, said device comprises means for selecting the highest elementary memory capacity among the different elementary memory capacities determined for each of the coding modes.  21. Device according to claim 19 or 20, characterized in that the means for determining the elementary memory capacity for each mode. <Desc / Clms Page number 53>  encoding, determining the total memory capacity and decision are integrated in a first communication device connected to a second communication device by a communication network.  22. Device according to claim 21, characterized in that it comprises means for comparing the total memory capacity determined for decoding the given number of coded data elementary units and the available memory capacity in the second communication device for decoding of these data.  23. Device according to claim 22, characterized in that it comprises means for obtaining by the first communication device the memory capacity available in the second communication device.  24. Device according to one of claims 19 to 23, characterized in that it further comprises: - means for determining the basic number of operations to be performed to decode each elementary unit of data encoded according to a different coding mode means for determining the total number of operations to be performed during a predetermined time interval for decoding a plurality of elementary data units coded as a function, on the one hand, of the previously determined elementary number and, on the other hand, , the number of elementary units coded according to each of the coding modes.  25. Device according to claim 24, characterized in that the decision as to the determination of said at least one coding parameter is also taken according to the total number of operations previously determined and a total number of predetermined operations.  Device according to claims 21 and 25, characterized in that the means for determining the basic number of operations, determining the total number of operations during a predetermined time interval and decision are integrated in the first communication device. .  27. Device according to claim 26, characterized in that it comprises means for comparing the total number of operations determined during <Desc / Clms Page number 54>  the time interval and the total number of operations that could be performed during the same time interval in the second communication apparatus.  28. Device according to claim 27, characterized in that it comprises means for obtaining by the first communication device the total number of operations that can be performed by the second communication device.  29. Device according to one of claims 21 to 23, characterized in that the decision as to the determination of said at least one coding parameter is also taken according to characteristics of the communication network.  30. Device according to one of claims 19 to 29, characterized in that it comprises means for estimating the quality of at least one elementary unit of data which compares said at least one elementary unit of data coded with the determined coding parameter and said at least one uncoded unit of data.  31. Device according to claim 30, characterized in that the device comprises: means for estimating the quality of at least one elementary unit of data for each combination of coding parameters determined, means for selecting the better quality among the different qualities estimated for the different coding parameters.  32. Device according to claim 30 or 31, characterized in that it comprises: coding means of at least one elementary data unit with the determined coding parameter; means for decoding said at least one unit; coded elementary data element; means for estimating the quality of said at least one elementary unit of data which compares this at least one elementary unit of uncoded data with that previously decoded.  33. Device according to one of claims 19 to 32, characterized in that the data constitutes a video and the given number of elementary data units corresponds to an image of the video. <Desc / Clms Page number 55>  34. Device according to claim 33, characterized in that the elementary data unit is a macro-block of data.  35. Communication apparatus, characterized in that it comprises a device for determining at least one data coding parameter according to one of claims 19 to 34.  36. Computer-readable information storage medium or microprocessor comprising code instructions of a computer program for performing the steps of the method of determining at least one data encoding parameter according to the present invention. one of claims 1 to 18.  37. Removable, partially or completely readable information storage medium by a computer or a microprocessor comprising code instructions of a computer program for performing the steps of the method of determining at least one coding parameter data set according to one of claims 1 to 18.  38. Computer program loadable in a programmable device, characterized in that it comprises instruction sequences or portions of software code to implement the steps of the method for determining at least one data encoding parameter according to one of claims 1 to 18, when this computer program is loaded and executed by the programmable device.