EP1905038A1

EP1905038A1 - Method for storing individual data elements of a scalable data flow in a file and corresponding device

Info

Publication number: EP1905038A1
Application number: EP06763760A
Authority: EP
Inventors: Peter Amon; Andreas Hutter; Thomas Rathgen
Original assignee: Siemens AG
Current assignee: Siemens AG
Priority date: 2005-07-20
Filing date: 2006-06-16
Publication date: 2008-04-02
Also published as: CN101228586A; US20090119314A1; WO2007009849A1; CN101228586B; DE102005033981A1; JP2009502066A

Abstract

The invention relates to a method for storing individual data elements of a scalable data flow in at least one file. Quality levels of the scalable data flow are described by at least one scaling feature in respective scaling levels and at least one data element is associated with every scaling level of the scaling features. A processing index is allocated to the data element in such a manner that only one or more data elements having a low value of the respective processing index have to be taken into consideration for processing. At least one descriptive list is produced in such a manner that it comprises descriptive elements associated with the data element, namely the scaling level of the respective scaling features, the time index and/or the processing index. At least one of the descriptive lists and the data elements associated therewith can be stored in one of the files in an organized manner. The invention also relates to a device for implementing and carrying out the method.

Description

description

Method for storing individual data elements of a scalable data stream in a file and associated device

The invention relates to a method according to the preamble of claim 1 and a device according to the preamble of claim 10.

In many applications, media data streams, for example, video data streams or audio data streams, in different qualities are needed. For example, a mobile phone is only able to reproduce the video data stream with a small image resolution, for example 176x144 pixels. On the other hand, portable computers, such as a tablet PC, can display on their display the video data stream with up to 1280 x 768 pixels.

In order to offer a media data stream for different terminals with different terminal properties, the media data stream can be created in several data streams of different quality. This procedure is disadvantageous since a large storage capacity must be made available for storing these large numbers of data streams for a media data stream.

In another variant, the media data stream is encoded in a base data stream and a plurality of sub-data streams, wherein an improvement of the quality, for example image quality, compared to the base data stream is achieved by adding one or more partial data streams to the base data stream. By means of this scalable coding, by adding one or more partial data streams to the basic data stream, a terminal can obtain a data stream to be decoded with a selectable quality such that it is suitable for the device properties of the specific terminal. The scalable coding ensures that on the one hand on the other hand that an adaptation of the device properties for the terminals is made possible by the addition of one or more partial data streams.

Encoded media data streams, such as a video stream, are organized into files. From [1], for example, a file format is known in which coded media data streams are stored. According to Chapter 7 of this document [1], this file format supports "layers" and "subsequences". It is explicitly pointed out that the "layers" and the "subsequences" are limited to reading the data format, whereas this information is not intended to describe properties of a codec, that is, of scalable data streams.

Reference is further made to document [2], which defines a specific hierarchical format with a fixed structure for the order of scalability directions, that is to say for a specific preferred application. The format proposed in [2] has the disadvantage that it does not support flexible definition of scaling directions.

The invention has for its object to provide a method and an apparatus which in the creation of at least one file for storing scalable data streams both a flexible definition of scaling directions and an adaptation of the scalable data streams to device properties of one or more different Endge- rates in simple and efficient way.

This object is achieved on the basis of the method according to the O-berbegriff of claim 1 by its characterizing features. Furthermore, this object is achieved according to the preamble of claim 10 by its characterizing features.

Other developments of the invention are given in the dependent claims. In the method for storing individual data elements of a scalable data stream in at least one file, wherein quality levels of the scalable data stream are described by at least one scaling feature in each of several scaling levels, each scaling level of the scaling features is assigned at least one data element, the scalable data stream in a high quality level at least one of those data elements is represented, whose scaling levels of the respective scaling features are equal to or smaller than the scaling levels of the respective scaling features associated with the high quality level, the data element is assigned a time index at which the respective data element is to be displayed relative to other data elements , the data element is assigned a processing index such that only one or more data elements with a lower value are used to process the data element At least one write list is generated in such a way that it comprises for the data element associated description elements consisting of the scaling level of the respective scaling features, the time index and / or the processing index, at least one of the description lists and the associated data elements are stored in one of the files organized.

The inventive method ensures that data elements and their description, which are presented in the form of at least one description list, can be stored in a very flexible manner.

If the description elements in the description list are organized according to at least one sorting criterion, in particular sorted from the lowest scaling level to the highest scaling level, a rapid retrieval of data elements required for processing can be achieved. Preferably, the description list is generated in such a way that the description list can be assigned to a terminal functionality, in particular to a computing power or reproduction unit of the terminal. As a result, a terminal with a special terminal functionality can find the data elements required for itself quickly in the file in a simple manner. Furthermore, by specifying several description lists, terminals with a wide variety of terminal functionalities can be served.

If references to finding the associated data elements are preferably added to the description list and the data elements are organized in a data area in one of the files, then the description list and the data area can be managed separately from one another and separated, for example, when the file or files are transmitted be transmitted to each other. In this transmission, a special error protection for the description list and the data area can be used in each case, whereby bandwidth can be achieved compared to a single error protection.

If, in addition, the value of the processing index of the data element is determined from the reference associated with the data element, a data volume of the write list to be stored can be reduced.

Preferably, in addition, the data elements in the data area can be stored in an organized manner depending on the reference associated with the respective data element. Thus, the associated processing index can be determined by a position of the respectively stored data element. As a result, on the one hand, the data volume can be reduced by omitting the concrete processing indices. For others, the data area for reading the data elements only needs to be processed in one direction, thereby avoiding time-consuming jumps. In an alternative extension of the method according to the invention, in addition to a description list, a data list is stored in accordance with an MPEG-4 AVC description format for MPEG-4 AVC-compatible data elements for describing one of the lowest quality levels in one of the files. The MEPG-4 AVC description format is known from the standard ISO / IEC MPEG-4 AVC. This ensures that the file can also be read and evaluated by terminals that only know the MEPG-4 AVC description format.

Preferably, quality groups are furthermore formed from the quality levels in such a way that the quality group is assigned to one of the quality levels and all data elements are assigned to process this quality group. This can support a scalable coding method which supports layered multiple scaling levels.

If in the description list only those data elements or references of the quality group are assigned, which are added or required in addition to the data elements or references to the quality group which are next to the quality group, then a compact and memory-efficient representation of the quality group can be provided Description list are guaranteed.

Furthermore, the invention relates to a device for storing individual data elements of a scalable data stream in at least one file, wherein quality levels of the scalable data stream are described by at least one scaling feature in each of several scaling levels, each scaling level of the scaling features is assigned at least one data element, the scalable data stream in one high quality level is represented by at least one of those data elements whose scaling levels of the respective scaling characteristics are equal to or smaller than the ska associated respectively to the high quality level. lierungsstufen the respective scaling features, the data element is assigned a time index at which the respective data element is to be displayed relative to other data elements having a generator module, which is adapted to the data element, a processing index are assigned such that for processing the data element only one or more data elements with a lower value of the respective processing index are to be taken into account, at least one description list is generated such that it comprises descriptive elements associated with the data element, consisting of the scaling level of the respective scaling characteristics, the time index and / or the processing index At least one of the description lists and the associated data items should be stored in one of the files organized.

With this device, the inventive method can be implemented and executed. The device can be realized in hardware, in software running on a processor or in a combination of hardware and software.

The invention and its developments are explained in more detail with reference to the figures 1 to 15.

Show it:

Figure 1 quality levels of a scalable data stream;

Figure 2 shows procedures for achieving data elements representing different scaling features (refresh rate and image resolution);

Figure 3 description elements and data element for each

Quality level;

FIG. 4 shows a circuit diagram for generating data elements from a plurality of images and for producing reconstructed images; FIG. 5A A procedure for creating a reconstructed image for the quality level with high image resolution;

FIG. 5B: Procedure for creating a reconstructed image for the quality level with high refresh rate;

FIG. 6 A description list for the quality levels or data element according to FIG. 2 or 3;

Figure 7 A structure of a file consisting of description list and data area;

FIG. 8A shows a variant of a description list;

FIG. 8B shows a binary representation of the description list according to FIG. 8A;

FIG. 9 A device for storing individual data elements of a scalable data stream in a file;

FIG. 10 shows an application example for using the device within a video server in a network;

Figure 11 A description list considering quality groups; ;

FIG. 12 shows another variant of a description list taking into account device properties;

Figure 13 dependencies of a number of images in one

Description List; FIG. 14 two files, wherein in the first file several description lists and a data list and in the second file the data elements are stored in an organized manner.

Elements with the same function and mode of operation are provided in FIGS. 1 to 14 with the same reference numerals.

The inventive method for storing individual data elements of a scalable data stream in a file is explained in more detail using a video data stream. The video data stream represents one possible type of scalable data stream. Other types of scalable data streams are, for example, a speech signal, a piece of music or a data record that can be displayed in several quality levels.

For example, scalable video data streams have the property that portions of these data streams are of reduced quality, e.g. less sharpness, less local or temporal resolution and / or omission of certain objects, can be decoded. In certain coding methods it is possible to extract any coherent subsets from the overall data stream, which leads to a reduced local and / or temporal and / or to a worsened sharpness. Further, portions of the scalable data stream SD may be tagged depending on a purpose of use, e.g. that certain scenes of the scalable data stream are only accessible to certain age groups.

FIG. 1 shows the concept of "fill scalable" video coding. In this case, three scaling features, image repetition rate T, spatial resolution S and image sharpness B, are plotted on the axes. Each of the scaling features T, S, B can each be resolved in several scaling stages. TO = 7.5 fps (fps - frames per second = frames per second), Tl = 15 fps and T2 = 30 fps. Furthermore, the scaling levels of the spatial resolution S in FIG. 1 are SO = QCIF (QCIF = Quarter Common Intermediate Format = 176x144 image quality). points), Sl = CIF (CIF - Coramon Intermediate Format = 352x288 pixels) and S2 = 4CIF (4CIF - 4 times Coramon Interchange Format = 704x576 pixels). Finally, the scaling feature image sharpness B comprises two scaling stages BO, Bl, where BO corresponds to a coarse quantization and Bl to a fine quantization.

In Figure 1, each box represents one of the possible quality levels Q0, ..., Q4, each quality level being represented by an individual triple at scaling levels of the scaling features T, S, B. For example, the quality level Q3 is characterized by the scaling levels Tl, Sl, BO of the scaling features T, S, B. In FIG. 1, each quality level Q0,... Q4 represents a specific quality of the scalable data stream SD. The quality of the quality level QO is low, since both the image refresh rate, the spatial resolution and the image sharpness are low or coarse. To improve the quality of the scalable data stream SD, for example, the second quality level Q2 is selected. The spatial resolution S is increased from QCIF to CIF. Thus, in a "fill scalable" encoding method, as illustrated in Figure 1, an improvement in image quality can be achieved by traveling along the axes representing the scaling levels of the individual scaling features.

The production or coding and processing or decoding of the scalable data stream will be described in more detail below with the aid of FIGS. 2 to 5B. Methods for encoding and decoding a scalable data stream are known to the person skilled in the art, such as, for example, the ISO / IEC MPEG standardization activities for the SVC (Scalable Video Coding). For this reason, only the coding and decoding of a scalable data stream will be explained in general terms with the aid of FIGS. 2 to 5B. A scalable data stream or data elements of a scalable data stream are to be generated, which allow a refresh rate of T0 = 15 fps and Tl = 30 fps, as well as a spatial resolution SO = QCIF and Sl = CIF. As in To see gur 3, four different quality levels QO, ... Q3 can be realized. For each quality level, at least one data element D0,..., D5 is generated, which contains the relevant data of the respective quality levels Q0,..., Q3.

Based on FIG. 2, the generation of the data elements DO,..., D5 will be described in greater detail. To generate these data elements, three images PO, P1, P2 are considered, e.g. have been recorded at the times tθ, tl, t2. These

Pictures PO, Pl, P2 have a spatial resolution of CIF. First, in a first step X10, the image PO is converted from CIF to QCIF, e.g. by means of a two-dimensional subsampling filter, and optionally compressed, so that the result is the data element DO. This step is analogous in the

Step X12 performed for the image P2, which generates the data element Dl. In order to generate the data element D2, in step XI, for example, first the image P1 is converted from CIF to QCIF, and then a difference image in the form of a B-image is generated by means of the decompressed data elements DO, D1 (steps X20, X21). This difference image is compressed and thus yields the data element D2.

In a next step X30, the decompressed data element DO is enlarged from QCIF to CIF, e.g. by means of a two-dimensional filter. Then, in step X40, a difference image from the image PO and the enlarged and decompressed data element DO is detected and compressed. This compressed differential image is referred to as data element D3. To create the data element D4, the procedure is analogous using steps X32 and X42.

To generate the data element D5, the data element D2 is first decompressed in step X31 and enlarged from QCIF to CIF. Subsequently, a difference image is calculated taking into account the image Pl, the decompressed and enlarged image D2 and the decompressed and the reconstructed images on the basis of the data elements DO, D3 and the data element. elements Dl and D4. In step X50, one of the reconstructed images is determined from the decompressed data element DO enlarged by QCIF to CIF together with the decompressed data element D3. This is done analogously in step X51 for the data elements Dl and D4. In step

X41 is also the generation of the difference image and its compression, so that there creates the data element D5.

In FIG. 3, the scaling stages of the scaling features spatial resolution S and repetition frequency T of the data elements generated in FIG. 2 are plotted along the scaling stages. Also indicated is a respective time index ZIO, ..., ZI5 at which the associated data item is output, i. For example, to be displayed on a screen to be.

In addition, in FIG. 3, for each data element DO,..., D5 an associated processing index VO,..., V5 is listed. The processing index indicates in which order the individual data elements are processed, e.g. decompressed and processed. For example, if the data item D4, i. the quality level Q2, are processed, at least the data element Dl must be processed, e.g. decompressed and enlarged. In general, the processing index indicates that prior to processing the associated data element D4, one or more data elements Dl having a smaller value of the processing index V1 must already have been processed. Values for the respective processing index VO,..., V5 of the associated data element DO,..., D5 are plotted in the lower section of FIG. This list is exemplary and can also be represented by one or more alternative orders.

In Figures 2 and 3, only a few images, for example, PO, Pl and P2, have been considered. For a scalable data stream DS, a multiplicity of images is considered, for which data elements are generated in each case. It should also be noted that compression or decompression tion is optional. This compression or decompression can be enabled by the standard ISO / IEC MPEG-4 AVC.

FIG. 4 shows a circuit diagram for generating data elements from a plurality of images and for producing reconstructed images. As is known, for example, from FIG. 2, the data elements DO,..., D5 are generated with the aid of an encoding module COD from the images PO,..., P2. By using a decoding module DEC, one or more reconstructed images R 1, R 1 ', R 2, R 2' can be generated from one or more data elements. In this case, the reconstructed images Rl, R2 have a low spatial resolution SO and the images Rl ', R2' have a high spatial resolution Sl. If only the low repetition rate pictures TO, i. the reconstructed images Rl and Rl 'are displayed, the repetition rate TO = 7.5 fps. If additionally the reconstructed images R2 or R2 'are displayed, a reconstructed data stream of Tl = 15 fps results.

The procedure for generating reconstructed images is explained in greater detail in FIGS. 5A and 5B. If, for example, the reconstructed image R1 corresponds to the quality level Q1 shown in FIG. 3, the data elements DO, D1 and D2 are required for its creation. This dependence can be seen from FIG. If the quality level Q2 has been selected, then, as can be seen in FIG. 5B, the reconstructed image R2 'is created by means of the data elements D1 and D2.

Because of this modular approach in the generation of reconstructed images, a terminal, depending on its terminal functionalities, only the quality level and thus only consider those data elements that it can, for example, process or represent.

It is further noted that a high quality, scalable data stream takes into account both the data element of this high quality level and one or more lower quality data elements. That means, the scalable data stream of a high quality level is reproduced by one or more of those data elements, the scaling levels of the respective scaling characteristics being equal to or smaller than the scaling levels of the respective scaling characteristics respectively associated with that high quality level. This type of scaling is also known as "fill scalability." If, for example, the high quality level is Q3, the data packet D5 as well as, for example, all data packets of the lower quality levels must be taken into account in order to generate the scalable data stream DS, namely DO,..., D4.

In a subsequent step of the method according to the invention, a description list L1 is obtained by sorting the writing elements, e.g. the scaling levels of the scaling features, and / or the time index ZI9, ..., ZI5 and / or the processing index VO, ..., V5 include. The result of this sorting is shown, for example, in FIG. 6. The following three sort criteria were executed in the following sort order:

1. Time index ZIO, ..., ZI5 or tθ, ..., t2

2. Refresh rate TO, Tl

3. Spatial resolution SO, Sl.

The description list L1 is arranged in the file F such that, after specifying the sorting parameters, the associated data element is specified together with its processing index, such as "t0, TO, SO, V0, DO".

FIG. 7 shows an alternative representation of the description list from FIG. 6. In this case, instead of the processing index VO,..., V5 and the associated data element DO,..., D5, a respective reference VDO,..., VD5 is inserted into the description list L1. This reference

VDO, ..., VD5 has two functions. On the one hand, it points to an entry within a data section DAT of the file F, in which the individual data elements DO,..., D5 are organized are stored. Thus, by means of the reference VDO, ..., VD5 the data element DO, ..., D5 can be found. This is indicated symbolically in FIG. 7 by dashed arrows with the reference symbols VDO,..., DV5. On the other hand, from the reference VDO,..., VD5 of the data element DO,

..., D5 associated processing index VO, ..., V5 are obtained. Thus, the processing index VO can be determined from the position of the data element DO within the data section DAT. If the data element DO is in the first position, then the associated processing index VO = 0. If this data element DO is at the fourth position, then the processing index VO = 3. This applies analogously to the further data elements D1,..., D5.

In the above exemplary embodiments according to FIGS. 6 and 7, the data elements and the description list are stored sorted in a single file F. Generally, these can be stored in more than a single file, e.g. in the file F the description list Ll and in another file Fl the data elements DO, ..., D5.

FIG. 8A shows a further alternative representation of the description list L2. In this case, those entries within the description list L2, for which no data element exists, have been assigned a marking word "FREI" instead of the reference.

FIG. 8B shows a binary representation of the entries of FIG. 8A. This is, for example, a description list L2 '. The description elements assume the following binary values:

00: tθ, 01: tl, 10: t2 0: TO, 1: Tl 0: SO, 1: Sl

VDO: 000, ..., VD5: 101, FREE: 111 For example, the fourth line has the following bit pattern: "00, 1, 1, 111". If the individual binary symbols are replaced by the previously mentioned reference symbols, this fourth row reads: "tθ, Tl, Sl, FREI". The meaning of the reference symbols as binary symbols t.sup.th, t.sub.l, T0, Tl, S0, S, D0, Dl, D2, D3, FREI are known to a decoder or, for example, a priori or are communicated to it separately.

In FIGS. 6 to 8B, a uniform sorting sequence has been selected. This is one of the possible embodiments of the method according to the invention. In general, any order in sorting the descriptors, i. e.g. the time index and scaling levels of the scaling features. In particular, when creating the description lists

Terminal functionalities, such as a computing power or a playback unit of a terminal are observed. For example, a scalable data stream SD with a local resolution SO = QCIF can be processed and displayed on a terminal. FIG. 12 shows by way of example a description list L4 according to the following sorting order:

1. Spatial resolution S 2. Time index or time 3. Refresh rate T.

When evaluating the description list L4, the terminal only searches in that section of the file F which has the spatial resolution SO as the first sorting parameter. In the section of the description list L4, which begins with the spatial resolution Sl, the terminal does not have to search for possible data elements, since the terminal can not process and display a spatial resolution Sl.

In addition, individual data elements and / or scaling levels of the scaling features and / or time indices can be marked with an access index. The access index allows Licht an additional sorting criterion in the selection of the data elements to be processed. Thus, temporal sections of the scalable data stream SD can only be approved for adults, while remaining time sections are suitable for adolescents of 12 to 18 years. By marking the time indexes ZIO,..., ZI5 by means of an access index, processing of undesired data elements can be prevented or permitted.

Based on the embodiment of Figure 2, three images PO, ..., P2 have been processed. In this case, an influence length of the processing (= coding) is GOP = 2, since, apart from an auxiliary image PO required for processing, the influence length comprises the images P1 and P2. In general, any arbitrary length of influence GOP with the aid of the invention

Process are presented. If a larger influence length GOP, e.g. GOP = I6, selected so a number of scaling levels of the scaling feature refresh rate T can be increased. FIG. 13 shows by way of example that the influence lengths GOP, GOP1, GOP2 can change over time t, e.g. GOP1 = 2 pictures, GOP2 = 3 pictures.

In an extension of the method according to the invention, more than one description list L1, L2, 13 can be contained in the file F. In this case, a terminal can individually select one of the description lists for processing the data elements. Thereby, e.g. by specifying a number of a description list Ll, the terminal a certain procedure in the processing of the data elements are given. For example. a terminal should only be able to scale into quality groups. For this purpose, the terminal is instructed to process only the description list L3. This description list L3 will be explained later.

In an alternative extension of the method according to the invention, a data list DL can be added to the file F, whereby this data list DL is allocated according to an MPEG-4 AVC Description format for MPEG-4 AVC-compliant data elements describing one of the lowest quality levels QO, Ql in one of the files F, Fl organized. This can be used in the file to achieve backward compatibility with already existing terminals that can not evaluate the description lists L1. In Fig. 14, the optional data list DL has been inserted into the file F.

Additionally or alternatively, the description list Ll and the data area DAT may be stored in a file F or in a plurality of files F, Fl. As can be seen from FIG. 14, the description lists L1,..., L3 are stored in the file F and the data area DAT is organized in the further file F1.

FIG. 9 shows an exemplary embodiment of a device for carrying out the method according to the invention. In this case, uncoded images Pl,..., P3 are recorded by a camera K and transferred to the coding module COD, which generates the data elements DO,..., D5. These are transferred to a generator module GM which, after sorting the description elements, generates at least one of the description lists L 1 and stores them in at least one of the files F. Furthermore, the generator module GM creates the data area DAT within one of the files F, Fl, which comprises the data elements DO,..., D5. In this case, the file F may for example be stored on a memory module SM, in particular a hard disk.

FIG. 10 shows a possible application example of the device according to the invention. Within a network NET is a video server VX, which comprises the device according to the invention. The memory module SM can be coupled with the file F to the video server VX. Further, a camera K may be connected to the video server VX. In addition, the video server VX is able to send the created file or files with the description list L 1 and the data area DAT to a mobile device MG, for example a GSM Device (GSM - Global System Mobile Communications) or to a computer CG.

In a variant of the method according to the invention, the description list L3 can be generated in such a way that the data elements belonging to a respective time index are combined into quality groups G1,..., G3. Reference is first made to Figure 3, in which the quality groups Gl, ..., G3 are located. In this case, in the exemplary embodiment according to FIG. 3, not all quality levels can be selectively selected by the quality groups G1,..., G3. In Figure 3, three quality groups Gl, ..., G3 can be found. The first quality group Gl comprises only the data elements DO, Dl of the quality level QO. The second quality group G2 contains the data elements DO, D1, D3, D4 of the quality levels Q0 and Q1. The third quality group G3 comprises the data elements DO,..., D5 of the quality levels Q0,..., Q3. The grouping into quality groups according to FIG. 3 is known to the person skilled in the art as "layered coding" in the context of scalable coding. The layers are characterized by the fact that it is no longer possible to achieve any combination of scaling levels of the scaling features.

FIG. 11 shows a further variant of a description list L3 which describes the quality groups G1,..., G3 and has been created according to the following sorting order:

1. Time index ZIO, ..., ZI5, 2. Quality group Gl, ..., G3

In this case, the respective quality group G1,..., G3 is shown as representative of the scaling features T, S. Instead of specifying all associated data elements or references for each quality group, only those data elements or references per quality group G3 are listed, which in addition to the next smaller quality group G2, are additionally listed during processing or decoding. must be considered. If, for example, a terminal selects the quality group G3, all data elements or references of the lower quality groups G1 and G2 are selected in addition to the data elements or references listed there. With this approach, a compact and memory-efficient representation of the description list can be achieved when using quality groups.

In the preceding exemplary embodiments, only six data elements with the scaling features image repetition rate and (bi-id) spatial resolution were displayed. According to the present invention, more description elements, ie, for example also a larger number of scaling level and / or scaling features, may be present. For example, in FIG. 1, in addition to the image repetition rate T, the spatial resolution F, the image sharpness B is drawn in a third dimension.

References cited in this document:

[1] ISO / IEC, "Coding of Moving Pictures and Audio, Information Technology - Coding of Audio-Visual Objects, Part 15: AVC File Format", ISO, JTC1 / SC29 / WG11, MPEG03 / N5652, March 21, 2003.

[2] M. Z. Wisharam et al. "Extensions to ISO / AVC File Format to Support the Storage of Scalable Video Coding (SVC) Bitstreams", ISO / IEC JTC1 / SC29 / WG11, MPEG2005 / M12062, Buthan, Korea, April 2005.

Claims

claims

1. A method for storing individual data elements (DO, ..., D5) of a scalable data stream (SD) in at least one file (F, Fl), wherein a) quality levels (QO, ..., Q3) of the scalable data stream (SD) by at least one scaling feature (T, S, B) in each case a plurality of scaling stages (TO, Tl, SO, Sl, BO, Bl) are described, b) each scaling level of the scaling features (T, S, B) each at least one Data element (DO, ..., D3) is assigned, c) the scalable data stream (SD) in a high quality level (Q3) by at least one of those data elements (DO, ..., D5) is represented, their scaling levels of the respective scaling features (T, S, B) are equal to or smaller than the respectively to the high quality level (Q3) associated scaling stages (TO, Tl, SO, Sl, BO) of the respective scaling features (T, S, B) are, d) the data element ( DO, Dl) is assigned a time index (ZIO, ZIl) at which the respective data element (DO) relative to other Da elements (Dl) is to be represented, characterized in that e) the data element (D3) a processing index (V3) is assigned such that for processing the data element (D3) only one or more data element (DO, ..., D2) f) at least one description list (L 1, L 2) is generated in such a way that it is generated for the data element (D 0,..., D 5) with a lower value of the respective processing index (V 0,. associated description elements consisting of the scaling level of the respective scaling features (T, S, B), the time index (ZIO, ..., ZI5) and / or the processing index (VO, ..., V5), g) at least one of the description lists (Ll) and the associated data elements (DO, ..., D5) are organized in one of the files (F, Fl).

2. The method according to claim 1, characterized in that the description elements in the description list (Ll) according to at least one sorting criterion, in particular sorted by the lowest scaling level (SO) to the highest scaling level (Sl), are arranged organized.

3. The method according to any one of the preceding claims, characterized in that the description list (Ll) is generated such that the description list (Ll) of a terminal functionality, in particular a computing power or playback unit of the terminal, assignable.

4. The method according to any one of the preceding claims, characterized in that the description list (Ll) references (VDO, ..., VD5) to find the associated data elements (DO, ..., D5) are added, the data elements ( DO, ..., D5) are stored in a data area (DAT) organized in one of the files (F, Fl).

5. Method according to the preceding claim, characterized in that the value of the processing index (V2) of the data element (D2) is determined from the reference to the data element (D2) (VD2).

6. The method according to any one of claims 4 or 5, characterized in that the data elements (DO, ..., D5) in the data area (DAT) in dependence of the respective data element (DO, ..., D5) associated reference ( VDO, ..., VD5) can be stored in a organized manner.

7. The method according to any one of the preceding claims, characterized in that in addition to a description list (Ll), a data list (DL) according to an MPEG-4 AVC description format for MPEG-4 AVC-compatible data elements for the description of one of the lowest quality levels (QO, Ql) is organized in one of the files (F, Fl) ,

8. The method according to any one of the preceding claims, characterized in that from the quality levels (QO, ..., Q3) quality groups (G1, G2, G3) are formed such that the quality group (G2) assigned to one of the quality levels (Q2) and all data elements (D3, DO or Dl and D4) for processing this quality group (G2).

9. The method according to claim 8, characterized in that in the description list (L3) only those data elements (D2, D5) or references (VD2, VD5) of the quality group (G3) are assigned, which in addition to the quality group (G3 ) next lower quality group (G2) (D3) or reference (VD3) to form the quality group (G3).

10. Device for storing individual data elements (DO, ..., D5) of a scalable data stream (SD) in at least one file (F, Fl), in particular according to one of the preceding claims, wherein a) quality levels (QO, ..., Q3) of the scalable data stream (SD) can be described by at least one scaling feature (T, S, B) in a plurality of scaling stages (TO, T1, SO, S1, BO, B1), b) each scaling level of the scaling characteristics (T, S, B) in each case at least one data element (DO, ..., D3) is assigned, c) the scalable data stream (SD) is represented in a high quality level (Q3) by at least one of those data elements (DO, ..., D5), their scaling levels of the respective scaling features (T, S, B) the same as or smaller than the respectively to the high quality level (Q3) associated scaling levels (TO, Tl, SO, Sl, BO) of the respective scaling features (T, S, B) are, d) the data element (DO, Dl) a time index (ZIO, ZIl) is assigned, at which the respective data element (DO) to be displayed relative to other data elements (Dl), characterized by a generator module (GM), which is designed such that e) the data element (D3) a processing index ( V3) is assigned in such a way that only one or more data elements (DO,..., D2) with a lower value of the respective processing index (VO,..., V3) are to be considered for processing the data element (D3), f) at least one description list (L1, L2) is generated in such a way that it contains for the data element (DO, ..., D5) associated description elements, consisting of the scaling level of the respective scaling features (T, S, B), the time index (ZIO , ..., ZI5) and / or the processing index (VO, ..., V5), g) at least one of the description lists (Ll) and the associated data elements (DO, ..., D5) are stored in an organized manner in one of the files (F, Fl).