WO2000065838A2 - Conversion d'un fichier media en format variable pour une transmission progressive - Google Patents

Conversion d'un fichier media en format variable pour une transmission progressive Download PDF

Info

Publication number
WO2000065838A2
WO2000065838A2 PCT/GB2000/001614 GB0001614W WO0065838A2 WO 2000065838 A2 WO2000065838 A2 WO 2000065838A2 GB 0001614 W GB0001614 W GB 0001614W WO 0065838 A2 WO0065838 A2 WO 0065838A2
Authority
WO
WIPO (PCT)
Prior art keywords
media
quality
file
slice
bitstream
Prior art date
Application number
PCT/GB2000/001614
Other languages
English (en)
Other versions
WO2000065838A3 (fr
Inventor
Tony Richard King
Timothy Holroyd Glauert
Ian David Wilson
Phil El Well
Original Assignee
Telemedia Systems Limited
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Telemedia Systems Limited filed Critical Telemedia Systems Limited
Priority to EP00927450A priority Critical patent/EP1145556A2/fr
Priority to JP2000614662A priority patent/JP2002543690A/ja
Publication of WO2000065838A2 publication Critical patent/WO2000065838A2/fr
Publication of WO2000065838A3 publication Critical patent/WO2000065838A3/fr

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/25Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
    • H04N21/266Channel or content management, e.g. generation and management of keys and entitlement messages in a conditional access system, merging a VOD unicast channel into a multicast channel
    • H04N21/2662Controlling the complexity of the video stream, e.g. by scaling the resolution or bitrate of the video stream based on the client capabilities
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/30Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using hierarchical techniques, e.g. scalability
    • H04N19/36Scalability techniques involving formatting the layers as a function of picture distortion after decoding, e.g. signal-to-noise [SNR] scalability
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/40Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using video transcoding, i.e. partial or full decoding of a coded input stream followed by re-encoding of the decoded output stream
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/62Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding by frequency transforming in three dimensions
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/60Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
    • H04N19/63Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using sub-band based transform, e.g. wavelets
    • H04N19/64Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using sub-band based transform, e.g. wavelets characterised by ordering of coefficients or of bits for transmission
    • H04N19/647Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding using sub-band based transform, e.g. wavelets characterised by ordering of coefficients or of bits for transmission using significance based coding, e.g. Embedded Zerotrees of Wavelets [EZW] or Set Partitioning in Hierarchical Trees [SPIHT]
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/234Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
    • H04N21/2343Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs involving reformatting operations of video signals for distribution or compliance with end-user requests or end-user device requirements
    • H04N21/234327Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs involving reformatting operations of video signals for distribution or compliance with end-user requests or end-user device requirements by decomposing into layers, e.g. base layer and one or more enhancement layers
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/25Management operations performed by the server for facilitating the content distribution or administrating data related to end-users or client devices, e.g. end-user or client device authentication, learning user preferences for recommending movies
    • H04N21/258Client or end-user data management, e.g. managing client capabilities, user preferences or demographics, processing of multiple end-users preferences to derive collaborative data
    • H04N21/25808Management of client data
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04NPICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/60Network structure or processes for video distribution between server and client or between remote clients; Control signalling between clients, server and network components; Transmission of management data between server and client, e.g. sending from server to client commands for recording incoming content stream; Communication details between server and client 
    • H04N21/61Network physical structure; Signal processing
    • H04N21/6106Network physical structure; Signal processing specially adapted to the downstream path of the transmission network
    • H04N21/6125Network physical structure; Signal processing specially adapted to the downstream path of the transmission network involving transmission via Internet

Definitions

  • This invention relates to methods for the delivery of media files, such as video and audio files, from a server to one or more devices using a network, in which the files stored on the server are encoded so that the generation and/or reception of their associated bitstreams can be controlled to yield a file at the device which meets the specific criteria (for example, image quality) of that device.
  • the network may include the Internet and the device may be a client or another server.
  • the media is encoded as an ordered set of discrete blocks.
  • Each block encodes the entire media in such a way that reception of the first block allows some aspect of the entire media object to be viewed, and each successive block received produces a discrete step in the quality of the representation at the client.
  • the client can exert some control over the use of network resource, and over the quality of the media delivered, by stopping the delivery when the representation has the desired quality.
  • the client does not have control over the content of the block and therefore over the characteristics of the progressive improvement. Such decisions are made by the content provider when the media file is compiled. Neither does the client have any control over the ordering of the blocks or over the ordering of data within the blocks.
  • the quality improvements available from the system are only available in fixed steps, the number of such steps also being predetermined by the content provider.
  • Sharpening up of the image occurs as a result of 2 factors: first, as all of the 4 sub-bands which form each decomposition level are received and decoded, the image quality level increases slightly. As successive decomposition levels are received and decoded, the image quality increases more significantly.
  • One characteristic of the system taught in the 5,880,856 patent is that the sub-bands and decomposition levels are all transmitted to clients in the same, fixed order, which is relatively inflexible. Whilst this scheme allows the client control over quality, it provides no other parameters that can be controlled.
  • JPEG 2000 image coding system draft specification (ISO/TEC CD 15444-1 : 1999). This describes a coding method for still pictures using a wavelet transform in conjunction with coefficient significance-ordering in a manner similar to the SPIHT system described above. Markers are inserted into the bitstream to allow the decoder to extract a reconstructed image of the desired quality.
  • New applications for media delivery over networks require more flexibility and greater interactivity than is possible with any of the prior art schemes described above.
  • Such an application is that of professional media browsing where the aim is rapidly to find, examine and annotate many different aspects or views of the media.
  • Such a system must be capable of the following:
  • a component can be a time-ordered sequence of media parts, a media part sampled at a particular time, a spatial area within a media part, a colour within a media part, or any other aspect which can be described.
  • the present invention is therefore based on the insight that it is possible to design a universal new layered file format into which files encoded with many different schemes (e.g. block based schemes such as MPEG, JPEG and DV; subband based schemes such as wavelets) can be converted.
  • Media files conforming to this new format can then be manipulated, e.g. at a client device, without any detailed knowledge of the original compression format.
  • Various quality axes define the layering in the new format: extensive manipulation functionality may then be supported by the new layered file format, allowing complex editing and viewing tasks to be performed.
  • An implementation of this new format is the IPV format from Telemedia Systems Limited of the United Kingdom.
  • the method comprises the step of structuring the media bitstream at an encoder using a protocol which converts the media bitstream into a file with a structure including (i) layers which sit in a dependency hierarchy and (ii) groups of those layers which are related by dependency.
  • a protocol which converts the media bitstream into a file with a structure including (i) layers which sit in a dependency hierarchy and (ii) groups of those layers which are related by dependency.
  • a codebook may be used to store information that maps the properties of a particular encoding into a form conformable to a layered representation to generate a consistent view of any media file so that the media file can be manipulated without detailed information relating to the original compression format used; the codebook preferably exploits the two-fold structure insight.
  • layer information is encapsulated and labelled to form a 'slice': a group of layers related by dependency is referred to as a 'context'.
  • the slice is a transport mechanism for which the layer information is in effect the payload.
  • the 2 fold structure is represented in MPEG, JPEG, 2-d wavelets and 3-d wavelets as follows in the IPV implementation:
  • a media encoding scheme can be described in terms of units of information which represent an input signal captured within a temporal window and digitally sampled, then transformed by a particular coding scheme into a set of conceptual layers
  • These layers initialise or refine particular quality axes within the encoded media file and, again conceptually, can be added or subtracted in order to improve or reduce the quality of the decoded media file.
  • these layers may operate spatially, temporally, or along any other quality axis.
  • the number of quality axes will vary between encoding schemes, as will the granularity with which quality is represented along each axis.
  • Dependencies will exist between the layers such that it is impossible to decode a layer without also being able to decode the layer on which it depends.
  • Such a protocol can be used to provide a consistent view of media files from the point of view of tools designed for manipulation of these layers. This allows operations to be defined that are valid for any files in this layered format, irrespective of their original encoding.
  • An Encoder encodes the media, adds the layering protocol, and transmits the media as a bitstream onto a network;
  • a Server stores the layered bitstream as a network resource;
  • a Filter selects data for a particular layer or set of layers and appends a Filter Mask to the bitstream to signal the bitstream content to downstream tools;
  • a Client decodes and views the media.
  • the term network should be expansively construed to cover any kind of data connection between 2 or more devices.
  • the network includes the Internet.
  • the device at which the media file is reconstructed may be a client or may be a server, such as an edge server.
  • a file is any consistent set of data, so that a media file is a consistent set of data representing one or more media samples, such as frames of video or audio levels.
  • Another aspect of the invention includes a media file which has been reconstructed from a bitstream using any of the above inventive methods.
  • a further aspect is a computer readable data signal in a transmission medium which can be reconstructed as a media file, in which the data signal is generated using any of the above inventive methods.
  • a further aspect is a computer program which when running on a client enables the client to receive and playback a media file reconstructed from a bitstream which has been converted using any of the above inventive methods.
  • a final aspect is a computer program which when running on a server or encoder enables the server or encoder to perform any of the above inventive methods.
  • Figure 1 is a schematic representation of a network used in performing the method of media file delivery according to the present invention
  • Figure 2 is a schematic representing the sub-bands which result from applying wavelet transforms to an image in a conventional multi-scale decomposition with examples of partial reconstruction according to the present invention
  • Figure 3 is a schematic representation of the format of the 'chunk' data structure according to the present invention.
  • Figure 4 is a schematic representing the typical path of a unit of media through the layering system according to the present invention.
  • Figure 5 is a schematic representation of the labelling mechanism as applied to a wavelet encoding according to the present invention.
  • Figure 6 is an example of a fragment of a Codebook for the labelling example of Figure 5, according to the present invention.
  • Figure 7 is a schematic representation of slice filtering as applied to the labelling example of Figure 5 according to the present invention.
  • Figure 8 is a schematic representation of slice merging as applied to the labelling example of Figure 5 according to the present invention
  • Figure 9 is a schematic representation of the labelling mechanism as applied to MPEG encoding according to the present invention
  • Figure 10 is an example of a fragment of a Codebook for the labelling example of Figure 9, according to the present invention.
  • Figure 11 is a schematic representation of slice filtering as applied to the labelling example of Figure 9 according to the present invention.
  • Figure 12 is a schematic representation of the labelling mechanism as applied to DV encoding according to the present invention.
  • Figure 13 is an example of a fragment of a Codebook for the labelling example of Figure 12, according to the present invention.
  • Figure 14 is a schematic representation of slice filtering as applied to the labelling example of Figure 12 according to the present invention.
  • Block-based, motion compensated encoding schemes Block-based, motion compensated encoding schemes.
  • MPEG Discrete Cosine Transform
  • P-frames a set of spatial frequencies obtained through the Discrete Cosine Transform (DCT), where the low spatial frequencies are generally represented with much greater accuracy than the high spatial frequencies.
  • DCT Discrete Cosine Transform
  • a particular Macroblock in a P-frame is generated at the encoder by searching a previous I-encoded frame (or P-frame) for a Macroblock which best matches that under consideration. When one is found the vector is calculated which represents the offset between the two. This motion vector, together with the error between the predicted block and the actual block, is all that need be sent in order to reconstruct the P-frame at the receiver.
  • the third kind is called a B (Bi-directional) frame.
  • B-frames are based on motion vectors obtained by past and future I and P- frames; they provide a smoothing effect, increase the compression factor and reduce noise.
  • MPEG exemplifies three properties of encoded media files that are factors in the design of a system such as is described here.
  • the first property is that the initial sample structure is preserved through the encoding, i.e., the identity of individual frames is not lost in the encoded bitstream. This means that temporal properties can be manipulated by adding or removing frames in the compressed domain.
  • a temporal window is defined (the MPEG Group of Pictures or GOP) within which temporal redundancy is exploited to achieve compression.
  • a complex set of dependencies is defined by the encoding; in MPEG, P-frames require I-frames, and B-frames require I and P-frames, for decoding.
  • Block-based encoding schemes that utilise motion detection and compensation including H.261 and H.263.
  • JPEG and DV have other schemes, notably JPEG and DV (as defined in SMPTE 314M-1999), that use block-based encoding without motion compensation.
  • the basic scheme is to transform blocks of pixels into frequency components using the DCT, quantise the components by multiplying by a set of weighting factors, then variable-length code the result to produce the encoded bitstream.
  • DV introduces the concept of feed-forward quantisation that optimises the compression process prior to compression being applied. To do this the DCT-transformed image is examined and classified into areas of low, medium and high spatial detail. Using this information, different tables of quantisation factors are selected and used according to area, with the object of matching the fidelity with which frequency coefficients are represented, to the frequency response of the human visual system.
  • a second feature of DV is it's use of block-based adaptive field/frame processing. What this means is that a 64-entry DCT block can represent either an 8-by-8 area of pixels in a de-interlaced frame (an 8-8 DCT), or two 4-by-8 areas in the first and second fields of a frame (a 2-4-8 DCT). The choice between the two is done by detecting motion. The former scheme is used if there is little motion occurring between fields, the latter if motion is detected; this choice being made on a per-block basis.
  • a transform such as DCT
  • An example of this is the Wavelet Transform.
  • the wavelet transform has only relatively recently matured as a tool for image analysis and compression.
  • the FWT generates a hierarchy of power-of-two images or subbands where at each step the spatial sampling frequency - the 'fineness' of detail which is represented - is reduced by a factor of two in x and y.
  • This procedure decorrelates the image samples with the result that most of the energy is compacted into a small number of high-magnitude coefficients within a subband, the rest being mainly zero or low-value, offering considerable opportunity for compression.
  • Each subband describes the image in terms of a particular combination of spatial/frequency components. This is illustrated in Figure 2 A.
  • a wavelet filter is applied twice to an image. After the first application 4 subbands at Level 0 result, with each subband a quarter the scale of the original; after the second application, four new one-eigth-scale subbands are created. To reconstruct an image fully, the procedure is followed in reverse order using an inverse wavelet filter.
  • the Figure also illustrates the scheme used to name subbands: 'L' and ⁇ ' refer to Low-pass and High-pass wavelet filters which are applied in x and y to generate a subband. Applying the H filter in both x and y gives the 'HH' subband, applying 'L' in x and 'H' in y results in a 'LH', and so forth.
  • Subbands need not always be fully reconstructed into the original image; they may be combined in different ways to produce a final image to satisfy some set of individual requirements.
  • Three examples are shown in Figure 2.
  • the LL sub-band of Level 1 is used as a one-sixteenth scale version of the original.
  • Figure 2C the four subbands of Level 0 are inversely transformed to reconstruct a one-quarter resolution version of the original.
  • Figure 2D the LL and LH sub-bands of Level 1 are used, together with the LH sub-band of Level 0 to reconstruct an original resolution image, but with horizontal features at all scales emphasised.
  • Tree-based data structures can be used by compression schemes to exploit spatial redundancy in images.
  • a recent development of such a scheme is the SPIHT algorithm; see Beong-Jo Kim, Zixiang Xiong and William Pearlman, "Very Low Bit- Rate Embedded Coding with 3D Set Partitioning in Hierarchical Trees ", submitted to IEEE Trans. Circuits & Systems for Video Technology, Special Issue on Image and Video Processing for Emerging Interactive Multimedia Services, Sept. 1998.
  • the SPIHT algorithm is an effective compression step for use in conjunction with a wavelet transform, since it can efficiently exploit the decorrelation of the input image provided by the transform to get high levels of data reduction.
  • a key feature of the SPIHT algorithm is its ability to analyse the transformed image in terms of Significance. It efficiently locates and partially orders all the samples with respect to bit-significance, i.e., the position at which the highest- order bit in a sample is set. Since this corresponds to the magnitude of the sample, the samples are effectively ordered with respect to their energies, or the contribution they make to the reconstructed object.
  • a significance layer is generated by choosing a bit- position and outputting the value of that bit-position (1 or 0) for all the samples for which that bit-position is defined.
  • a set of dependencies between related image components is defined by schemes such as Wavelet/SPIHT compression.
  • a significance layer can be decoded only if the next highest significance layer can also be decoded.
  • the parent subband LHl in Figure 1
  • Wavelet and SPIHT compression schemes can be extended to the third (time) dimension in a fairly straightforward manner.
  • a sequence of frames are captured and are treated as a 3-dimensional block of data.
  • the wavelet filter is applied along the time dimension as many times as it is along the vertical and horizontal dimensions, resulting in a transformed block comprising a set of spatio-temporal cubes, each of which is a subband.
  • the 3D-SPIHT algorithm represent these subbands in terms of octtrees (as opposed to quadtrees in the 2D case) and generates the bitstream accordingly.
  • the system capable of performing the method of the present invention comprises an Encoder, a Network, a Server and N clients, as shown in Figure 1.
  • a Encoder compresses the incoming media, structures it into layers and adds layer labelling information as described in the next section, prior to transmitting it as a bitstream over a digital communications network, to be stored on the Server.
  • FIG. 1 there are N clients, each of which engages in a session with the media Server during which media content and control information are transmitted. For each client a control channel to the Server is maintained which allows that client to request that media be transmitted to it.
  • the types of requests that can be made include (but are not limited to) the following :-
  • the application may then determine that this particular section of the media is to be rendered at greater quality and issue an appropriate request to the Server.
  • the Server calculates which extra layers should be delivered to improve the media information in the way required and transmits only those layers. Note that no extra processing of the media file is required during this procedure other than to locate and send the appropriate sets of data. Note also that the quality improvement may occur in a number of ways, including :
  • a Layered bitstream is built out of a media encoding by splitting it up into packets (called chunks here) in a manner determined by the encoding type, and appending Labels to the chunks that uniquely describe their contribution to the reconstructed file.
  • static signalling information (constant throughout a stream's lifetime) is made available, either as a reference to a globally available file, or as data sent once in the bitstream. For configuration information that must vary within the stream, provision is made for dynamic signalling information to be carried in the bitstream.
  • the first requirement is addressed through the concept of a Filter that specifies which parts of the file are to be selected for transmission.
  • the second requirement needs in- band signalling information to describe and delineate the layered structure within the bitstream.
  • the third requirement is addressed through an indirection mechanism that maps the subtleties of a particular encoding to the uniform layered view. This indirection mechanism is termed the Codebook.
  • a Codebook is defined for every new encoding scheme, or variations of existing schemes with new parameters.
  • Layer Labelling Format This section defines the format of the layer labels used in the present invention to convert a media file to a layered stream format. Streams are stored and transmitted as contiguous groups of chunks, each with a chunk header defining the length and type of the chunk. Some chunk types are reserved for carrying data (termed slices), whereas others are used for signalling.
  • Extra configuration information for these streams can be associated with the data stream.
  • the static portion of this information can be stored externally, or in-band using signalling chunks; dynamic information is always stored in-band.
  • the 2-bit type field is used to distinguish between 4 independent sets of chunk types. (The issue of independence is important, since there is no way to specify a dependency between two labels belonging to different types.) The following type values are currently defined:
  • the 22-bit length field gives the length of the chunk in bytes, including the header.
  • a chunk containing no payload information i.e. just a header
  • the 8-bit label field is the layer label for the chunk.
  • the interpretation of the label is dependent on the type field (see above), so the header format allows for 4 sets of 256 labels.
  • Data chunks are used to encapsulate a stream of multimedia data.
  • the labels attached to the slices are used to indicate the nature of the slice contents and their dependency relationships with other slices in the stream.
  • Signalling chunks can be used within the stream to pass extra information about the data, such as details about how it has been encoded or filtered.
  • the slice label is taken as an indication of the data contents of the slice, and is used by the system to separate and merge the slices to obtain data streams of differing quality.
  • the slice label In order to allow the stream decoders to be stateless, the slice label must identify uniquely the actual data contents. As a result, slices which are too long to be encapsulated as a single chunk must be split in such a way as to enable the decoder to determine which is the base data and which is the continuation information.
  • the slice labelling must also reflect the dependency hierarchy, with all slices having numerically lower label values than those which depend on them. This means that when the slices of a stream are encoded in dependency order, the label values increase
  • a group of slices which are related by dependency is called a "context".
  • the label values for the slices within the context must be arranged so that they are in numerically increasing order. This usually allows the boundaries between contexts to be detected implicitly, without the need for an explicit signalling chunk. Based on the label allocation scheme described above, slices within the context will also, therefore, be arranged in dependency order such that slices always precede those which are dependent on them.
  • the maximum number of slices which comprise a context is 256.
  • the dependency hierarchy for the slice labels is defined in a code book (see below) and is fixed for the duration of the data stream. Note that it is not mandatory for slice labels to be allocated consecutively (i.e. gaps in the allocation scheme are allowed); nor is it essential for a specific slice label to be present in all contexts.
  • Contexts are assumed to be separated from their neighbours in the temporal domain, and be independent of them; in other words, when contexts are extracted from the data stream, each should be capable of being decoded as a stand alone entity.
  • encoding schemes such as MPEG which have temporal dependencies between adjacent Groups Of Pictures. This is handled by associating an "overlap" with each context allowing temporal units (see sample below) to be marked as dependent on the immediately preceding context.
  • an overlap value of n implies that the first n samples of a context are dependent on the previous context. Since the normal situation is for contexts to be temporally self-contained, the default overlap value is zero. In the case where it is not zero, the slice dependency hierarchy must reflect the inter-context peculiarities of the encoding scheme, with "phantom" dependencies being added if necessary.
  • Multimedia data streams are usually represented as a sequence of discrete units, each with a well defined temporal position in the stream. This is particularly true with video data, which is usually modelled as a set of separate frames (even though the frames may have been reordered and grouped in the encoding process, such as with MPEG).
  • the data streams preserve the natural temporal unit of the encoding scheme, with each discrete unit being termed a sample. Whether a context contains a single sample or a group of samples is dependent on the encoding technique used, but specific techniques usually follow a very rigid repeating pattern. By default, therefore, each context is assumed to contain a fixed number of samples, with the context spacing (samples per context repeat count) defined in the system header (see below).
  • contexts and samples are important when contemplating the temporal dependencies of the multimedia data (such as a 3-D Wavelet/SPIHT scheme) and the ability to perform temporal decimation (such as playing every fourth frame of a video sequence).
  • a context contains the smallest number of samples which make up a temporal encoding unit (GOP for MPEG, GOF for Wavelet/SPIHT), with the spatial and temporal dependencies being handled in exactly the same manner.
  • each slice is given a temporal priority (see below) which allows for intelligent decimation of the temporal sequence, but does not in itself imply any kind of temporal dependency.
  • the system header is used to define the stream attributes which are relevant at the system layer only. Basically, this means parameters which are needed by the Server to provide the level of service required of them. For example, the default context spacing is needed in order to construct the mappings between contexts and samples.
  • the media header is used to define the stream attributes which are relevant at the media layer only. This means parameters which are needed by the stream decoder in order to make sense of the data stream, but which are not required by the Server. Examples which fall into this category are: the horizontal and vertical resolutions of a video encoding; the inter-context dependency overlap; and a reference to the code book which has been used for the encoding.
  • the code book is used to define the dependency and quality relationships of the slices within a context, and is necessary in order to construct a filter for a data stream. Irrespective of whether the dependencies are spatial or temporal, the relationships between the individual slices can be represented as a simple hierarchy: each slice type (i.e. slice with a specific label) has a predetermined set of other slice types on which it is dependent. The nature of the hierarchy is, obviously, dependent on the labelling scheme used; but the code book technique allows a great deal of flexibility providing the basic rules of slice labelling are not violated (see above).
  • 3-D Wavelet/SPIHT encoded video can be thought of as having four quality axes: scale (image size), fidelity (image quality), colour and temporal blur.
  • the bulk of the code book comprises a series of tables (one for each slice label) containing the dependency data, the temporal priority and a set of quality parameters, one for each quality axis.
  • the number and type of quality axes defined is dependent on the nature of the multimedia data and the encoding scheme used. To simplify matters, three important restrictions are applied to quality axes: Quality axes are assigned names (or quality tags) which have global significance. Code books which employ the same quality tag must use the same allocation scheme when it comes to quality parameters for the relevant axis.
  • Quality parameters are always specified with regard to the highest quality (unfiltered) version of the data stream. In the case of video, for example, this allows a common code book to be used irrespective of the image size.
  • the scale parameters for video data might represent the width or height of a scaled image compared to the original, assuming an unchanged aspect ratio.
  • the code book header contains the list of quality tags for which quality parameters will be found in the tables which follow. This, combined with the restrictions outlined above, allows for filter creation to be done in a manner which is entirely independent of the multimedia data type and encoding scheme. A filter created using a quality axis which is either absent or unspecified always results in slices being unfiltered with respect to that axis.
  • inter-slice dependencies of the relevant samples may prohibit simple decimation. This is true, for example, with MPEG video, where there is a complex web of temporal dependencies between the I, P and B frames.
  • each slice in a context is assigned a temporal priority which is used to control decimation.
  • a temporal priority which is used to control decimation.
  • slices belonging to the I, P and B frames would be allocated temporal priorities of 0, 1 and 2 respectively. It is these temporal priorities which are used when performing temporal filtering below the level of the context.
  • the signalling chunks are used to identify and annotate the data chunks in a data stream. Whether they are stored together by Servers or generated on the fly at the time of delivery is an implementation detail which is outside the scope of this document.
  • the signalling chunks fall into three distinct categories: static, dynamic and padding.
  • the static chunks define parameters which do not change during the lifetime of the data stream. This includes the definitions of the basic default values which apply to the whole stream, such as the identity of the code book which has been employed to generate the data slices. Since these chunks are static, there is no necessity for them to be included as part of the data stream: it is just as valid for the information to be sent out-of-band or, as in the case of re-usable code book tables, by reference. If they are transmitted or stored in-band, however, they must be unique, precede the first data slice and be stored in numerically increasing opcode order.
  • the dynamic chunks define those parameters which vary depending on their position within the data stream. This includes filtration information, since the filters used to generate the stream can vary as the slices are delivered. It also includes all the variable context and sample information, such as the indication of "seek" operations and the handing of contexts with a context spacing which is not constant. Dynamic chunks, by their very nature, carry positional information within the data stream, and hence must always be sent in-band. Where present, dynamic chunks are only valid at context boundaries.
  • the padding chunks have no semantic influence on the data stream, and hence can be positioned at any chunk boundary.
  • Contents Information which is specific to the system layer only (Media Server), and is not required to filter or decode the data stream.
  • the parameters are: context spacing
  • Position Out-of-band, or in-band before first data slice.
  • Contents Information which is specific to the media layer only (Decoder), and is not required for the operation of the Media Server.
  • the parameters are: original media parameters context dependency overlap code book reference
  • Position Out-of-band, or in-band before first data slice.
  • Position Out-of-band, by reference, or in-band before first data slice.
  • Contents Information which is pertinent to the data stream, but which is not required for its storage, filtration or transmission. Examples are: date, time & location edit history copyright notice
  • Opcode 0x05, 0x06, 0x07, 0x08 Contents: Private information, added into the data stream by the generator of the encoding, but which is not interpreted by the system in any way.
  • Contents A context spacing for the immediately following context. Format: 32-bit little-endian binary. Status: Optional. Only necessary if the context spacing is different to the default. Position: In-band at context boundaries.
  • Position In-band at context boundaries.
  • Opcode 0x83 Contents: Used to encode stream filtration information for the data slices which follow. It takes the form of a pair of bit sequences (possibly compressed) indicating which data has been filtered out of the subsequent stream.
  • the slice mask indicates which of the 256 possible slice types have been filtered out of the data stream. It does not imply that the remaining slice types will be found in the stream; nor does it guarantee that the slices which are present conform to the dependency criteria specified in the code book.
  • the context mask is used to indicate which whole contexts have been filtered out of the data stream.
  • Each context has a unique position in the sequence, and mask refers to the position value modulo 240 (a number chosen because of its numerical properties, being the common multiple of 2, 3, 4, 5, 6, 8, 10, 12, 15, 16, 20, 24 and 30).
  • the context mask can be used to determine which contexts are present and which are absent. In the case of inter-context dependencies, however, there is no guarantee that the relevant information is actually present.
  • Padding chunks may be necessary in the case where fixed bit-rate streams are encoded or specific byte alignments are required. They can also be used to re-label a slice in place for data filtration experiments.
  • Figure 4 gives an overview of some typical transformations undergone by a media file as it is layered, filtered and decoded.
  • the Figure shows the two styles of filtering that are possible: filtering at the Slice level (within a dependency unit), and filtering at the Context level (across entire dependency units).
  • Slice-level filtering can be used to change the colour depth, spatial resolution, fidelity or temporal resolution of the media.
  • Context-level filtering has a coarser 'grain' and is designed mainly to support temporal decimation of the media.
  • Figure 5 shows a more detailed example of labelling Wavelet/SPIHT-encoded data.
  • a depth 2 wavelet decomposition with four significance levels results in a Context of 28 slices.
  • a fragment of one possible Codebook for this encoding is shown in Figure 6 and an example of how the Codebook is used to generate a Filter is illustrated in Figure 7.
  • the header of the Codebook includes a CodebooklD field for self-identification, a QualityTags field that specifies the quality axes available for this encoding, a QualityDivisions field that specifies, using plain-text names, how those axes are labelled, and an OriginalSize field that specifies the spatial resolution of the original media, to provide a baseline for scale calculations.
  • Each encoding scheme may have its own individual Codebook defined, but any Codebooks that share common QualityTags entries must use exactly the same QualityDivisions scheme to label the quality axes.
  • Textual names are used to map properties of the different media encodings, which may be similar but not identical, into a single quality scheme. This results in an important capability of the system: that an application can be written to manipulate media using these names in a manner that is completely independent of encoding format. In the Figure it is assumed that a file manipulation tool wishes to extract a medium fidelity, monochrome, half-scale version of a 352 by 288 video encoding.
  • Filtered slices are preceded by a Filter Definition signalling chunk which holds the slice mask used to perform the filtering. This value can be used by downstream tools that require knowledge of the stream characteristics, for example, to initialise a decoder, or set up memory areas for buffering.
  • the Slice Mask is also used when filtered bitstreams are merged to determine the slice content of the resulting stream, as illustrated in Figure 8.
  • a refinement stream containing slices 16, 17, 20, 21, 24 and 25 is merged with the half-scale, low-resolution filtered stream of Figure 7 to obtain a full-scale, low-resolution encoding.
  • the slice mask for the combined stream is obtained using a simple bitwise OR of the two input Slice Masks.
  • Figure 9 shows an example of labelling MPEG-encoded data
  • Figure 10 shows a fragment of the corresponding Codebook.
  • a file manipulation tool wants to generate a Filter that operates on slices in such a way as to generate a fast playback stream.
  • a Slice represents an MPEG frame, so the effect of the Filter is to decimate the stream temporally.
  • the result of applying the Filter is shown in Figure 11.
  • Figure 12 shows an example of labelling DV-encoded data
  • Figure 13 shows a fragment of the corresponding Codebook.
  • the DV-encoded data is labelled according to the following scheme: data that represents both 2-by-4-by-8 and 8-by-8 DCT blocks from field 1 are assigned slice label 0; data that represents 8-by-8 DCT blocks only from field 2, are assigned slice label 1.
  • Figure 14 illustrates the result of applying a Filter that selects label-0 Slices only, so producing a half vertcal resolution video stream.
  • a function can be analysed into a set of components with different time/space/frequency content. These components are called scales, and the process of analysis into scales is called multi-scale decomposition. The analysis is performed by a waveform of limited duration and zero integral, called a wavelet. Components that are highly localised in space or time but have an ill-defined frequency spectrum are small-scale and capture fine detail. Components that are spread through space but have a precisely-defined frequency spectrum are large-scale and capture general trends.
  • a slice is a labelled encapsulation of a layer.
  • a Layer is a conceptual part of a media file that initialises or refines a single homogeneous component of that file, where that component is data representing spatial resolution, temporal resolution, sample fidelity, colour or any other quality axis.
  • a layer where all the refinement information refers to a particular bit-position for all the coefficients undergoing refinement.
  • a layer for which all the refinement information is for a particular scale is for a particular scale.
  • a layer for which all the refinement information is for a particular connected region in a space or time-varying function is for a particular connected region in a space or time-varying function.
  • PSNR Peak Signal-to-Noise Ratio
  • MSE Mean Squared Error
  • QualityTags The set of axes that are available to represent different aspects of perceived quality, for example, spatial resolution, temporal resolution, fidelity of a reconstructed pixel with respect to the original, etc.
  • the marking scheme for the QualityTags axes for example, the spatial resolution axis may be marked with low, medium, high.
  • a set of classifiers that can be used to describe the contribution of an item of encoded media to the perceived quality of the reconstruction.
  • the Codebook achieves format-independence through use of the QualityParams classification system.
  • An information structure that defines a partitioning of a layered media file into two parts; one part representing the media file at a reduced quality, and the other representing the enhancement information needed to reconstruct the original file.
  • the simplest implementation is a bitmask where 'zero' and 'one' at bit positions n specifies whether a data item with a particular label ( ⁇ ) is or is not required in the lower-quality output file.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Databases & Information Systems (AREA)
  • Computer Graphics (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)
  • Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
  • Communication Control (AREA)

Abstract

La présente invention permet de distribuer sur un réseau un seul fichier vidéo ou audio codé qui peut être lu par différents clients avec des débits binaires, des résolutions et des niveaux de qualité différents, déterminés individuellement par chaque client. Un codeur insère dans les trains de bits du fichier média des informations de « signalisation de couche » qui délimitent et identifient plusieurs couches différentes (par exemple, différentes « couches de signification/échelle » et différentes « couches de région »). Un serveur de média stocke les fichiers média et peut distribuer à différents clients sur le réseau des trains de bits aux propriétés différentes en fonction des couches demandées par chaque client ou convenant à chaque client.
PCT/GB2000/001614 1999-04-26 2000-04-26 Conversion d'un fichier media en format variable pour une transmission progressive WO2000065838A2 (fr)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP00927450A EP1145556A2 (fr) 1999-04-26 2000-04-26 Conversion d'un fichier media en format variable pour une transmission progressive
JP2000614662A JP2002543690A (ja) 1999-04-26 2000-04-26 プログレッシブ送信のためのメディアファイルのスケーラブルフォーマットへの変換

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
GBGB9909605.9A GB9909605D0 (en) 1999-04-26 1999-04-26 Networked delivery of media files to clients
GB9909605.9 1999-04-26

Publications (2)

Publication Number Publication Date
WO2000065838A2 true WO2000065838A2 (fr) 2000-11-02
WO2000065838A3 WO2000065838A3 (fr) 2001-11-01

Family

ID=10852302

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/GB2000/001614 WO2000065838A2 (fr) 1999-04-26 2000-04-26 Conversion d'un fichier media en format variable pour une transmission progressive

Country Status (4)

Country Link
EP (1) EP1145556A2 (fr)
JP (1) JP2002543690A (fr)
GB (1) GB9909605D0 (fr)
WO (1) WO2000065838A2 (fr)

Cited By (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2816793A1 (fr) * 2000-11-10 2002-05-17 Satake Eng Co Ltd Dispositif de traitement d'information multimedia
WO2002071757A2 (fr) * 2001-03-07 2002-09-12 Internet Pro Video Limited Procede de traitement de donnees video dans un train de bits code
FR2826823A1 (fr) * 2001-06-27 2003-01-03 Canon Kk Procede et dispositif de traitement d'un signal numerique code
FR2831728A1 (fr) * 2001-10-25 2003-05-02 Canon Kk Procede et dispositif de formation d'un signal numerique derive a partir d'un signal numerique compresse
JP2003216520A (ja) * 2002-01-11 2003-07-31 Xerox Corp ドキュメントビューイング方法
FR2842057A1 (fr) * 2002-07-05 2004-01-09 Canon Kk Procede et dispositif de traitement de donnees dans un reseau de communication
FR2842983A1 (fr) * 2002-07-24 2004-01-30 Canon Kk Transcodage de donnees
GB2403620B (en) * 2002-04-29 2005-08-24 Sony Electronics Inc Generic adaptation layer for jvt video
EP1519587A3 (fr) * 2003-09-17 2006-02-08 Lg Electronics Inc. Appareil et méthode pour mise à disposition d'un service de téléchargement à haut de débit de données multimédia
US7116833B2 (en) 2002-12-23 2006-10-03 Eastman Kodak Company Method of transmitting selected regions of interest of digital video data at selected resolutions
GB2426401A (en) * 2005-05-20 2006-11-22 Bosch Gmbh Robert Handling scalable media data
US7895625B1 (en) 2003-12-24 2011-02-22 Time Warner, Inc. System and method for recommending programming to television viewing communities
US8196168B1 (en) 2003-12-10 2012-06-05 Time Warner, Inc. Method and apparatus for exchanging preferences for replaying a program on a personal video recorder
GB2552944A (en) * 2016-08-09 2018-02-21 V-Nova Ltd Adaptive content delivery network
US9959383B1 (en) 2004-04-30 2018-05-01 Time Warner, Inc. Apparatus, method and system for brokering and provision of intelligent advertisement
WO2023137284A3 (fr) * 2022-01-11 2023-09-07 Bytedance Inc. Procédé, appareil et support de traitement vidéo

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2879387B1 (fr) * 2004-12-15 2007-04-27 Tdf Sa Procede de transmission a debit binaire variable a travers un canal de transmission.

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1996029818A1 (fr) * 1995-03-17 1996-09-26 Imperial College Of Science, Technology & Medicine Transmission progressive d'images
WO1998037699A1 (fr) * 1997-02-25 1998-08-27 Intervu, Inc. Systeme et procede permettant d'envoyer et de recevoir une video comme montage de diapositives sur un reseau d'ordinateurs
WO1998037698A1 (fr) * 1996-12-17 1998-08-27 Adaptive Media Technologies Procede et dispositif constituant un systeme de remise de supports conformables

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR100540275B1 (ko) * 1997-07-11 2006-01-10 미디어텍 인크. 멀티스케일 제로트리 엔트로피 인코딩용 장치 및 방법

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1996029818A1 (fr) * 1995-03-17 1996-09-26 Imperial College Of Science, Technology & Medicine Transmission progressive d'images
WO1998037698A1 (fr) * 1996-12-17 1998-08-27 Adaptive Media Technologies Procede et dispositif constituant un systeme de remise de supports conformables
WO1998037699A1 (fr) * 1997-02-25 1998-08-27 Intervu, Inc. Systeme et procede permettant d'envoyer et de recevoir une video comme montage de diapositives sur un reseau d'ordinateurs

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
BEONG-JO KIM ET AL: "An embedded wavelet video coder using three-dimensional set partitioning in hierarchical trees (SPIHT)" PROCEEDINGS DCC '97. DATA COMPRESSION CONFERENCE (CAT. NO.97TB100108), PROCEEDINGS DCC '97. DATA COMPRESSION CONFERENCE, SNOWBIRD, UT, USA, 25-27 MARCH 1997, pages 251-260, XP002143648 1997, Los Alamitos, CA, USA, IEEE Comput. Soc. Press, USA ISBN: 0-8186-7761-9 *
See also references of EP1145556A2 *

Cited By (31)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
FR2816793A1 (fr) * 2000-11-10 2002-05-17 Satake Eng Co Ltd Dispositif de traitement d'information multimedia
WO2002071757A2 (fr) * 2001-03-07 2002-09-12 Internet Pro Video Limited Procede de traitement de donnees video dans un train de bits code
WO2002071757A3 (fr) * 2001-03-07 2003-01-03 Internet Pro Video Ltd Procede de traitement de donnees video dans un train de bits code
FR2826823A1 (fr) * 2001-06-27 2003-01-03 Canon Kk Procede et dispositif de traitement d'un signal numerique code
US7215819B2 (en) 2001-06-27 2007-05-08 Canon Kabushiki Kaisha Method and device for processing an encoded digital signal
FR2831728A1 (fr) * 2001-10-25 2003-05-02 Canon Kk Procede et dispositif de formation d'un signal numerique derive a partir d'un signal numerique compresse
US7113643B2 (en) 2001-10-25 2006-09-26 Canon Kabushiki Kaisha Method and device for forming a derived digital signal from a compressed digital signal
US7797455B2 (en) 2002-01-11 2010-09-14 Xerox Corporation Method for document viewing
US8019897B2 (en) 2002-01-11 2011-09-13 Xerox Corporation Method for viewing, on a client-side device, documents requested from a server-side device
US7975221B2 (en) 2002-01-11 2011-07-05 Xerox Corporation Method for document viewing
US7529755B2 (en) 2002-01-11 2009-05-05 Xerox Corporation Method for document viewing
JP2003216520A (ja) * 2002-01-11 2003-07-31 Xerox Corp ドキュメントビューイング方法
US7765473B2 (en) 2002-01-11 2010-07-27 Xerox Corporation Method for document viewing
GB2403620B (en) * 2002-04-29 2005-08-24 Sony Electronics Inc Generic adaptation layer for jvt video
CN100399824C (zh) * 2002-04-29 2008-07-02 索尼电子有限公司 用于jvt视频的通用适配层
FR2842057A1 (fr) * 2002-07-05 2004-01-09 Canon Kk Procede et dispositif de traitement de donnees dans un reseau de communication
US7467184B2 (en) 2002-07-05 2008-12-16 Canon Kabushiki Kaisha Method and device for data processing in a communication network
FR2842983A1 (fr) * 2002-07-24 2004-01-30 Canon Kk Transcodage de donnees
US7260264B2 (en) 2002-07-24 2007-08-21 Canon Kabushiki Kaisha Transcoding of data
US7512283B2 (en) 2002-12-23 2009-03-31 Eastman Kodak Company Method of transmitting selected regions of interest of digital video data at selected resolutions
US7116833B2 (en) 2002-12-23 2006-10-03 Eastman Kodak Company Method of transmitting selected regions of interest of digital video data at selected resolutions
EP1519587A3 (fr) * 2003-09-17 2006-02-08 Lg Electronics Inc. Appareil et méthode pour mise à disposition d'un service de téléchargement à haut de débit de données multimédia
US8196168B1 (en) 2003-12-10 2012-06-05 Time Warner, Inc. Method and apparatus for exchanging preferences for replaying a program on a personal video recorder
US7895625B1 (en) 2003-12-24 2011-02-22 Time Warner, Inc. System and method for recommending programming to television viewing communities
US9959383B1 (en) 2004-04-30 2018-05-01 Time Warner, Inc. Apparatus, method and system for brokering and provision of intelligent advertisement
GB2426401A (en) * 2005-05-20 2006-11-22 Bosch Gmbh Robert Handling scalable media data
GB2426401B (en) * 2005-05-20 2009-11-04 Bosch Gmbh Robert Method of handling scalable medial data
GB2552944A (en) * 2016-08-09 2018-02-21 V-Nova Ltd Adaptive content delivery network
US10951930B2 (en) 2016-08-09 2021-03-16 V-Nova International Limited Adaptive content delivery network
GB2552944B (en) * 2016-08-09 2022-07-27 V Nova Int Ltd Adaptive content delivery network
WO2023137284A3 (fr) * 2022-01-11 2023-09-07 Bytedance Inc. Procédé, appareil et support de traitement vidéo

Also Published As

Publication number Publication date
EP1145556A2 (fr) 2001-10-17
GB9909605D0 (en) 1999-06-23
WO2000065838A3 (fr) 2001-11-01
JP2002543690A (ja) 2002-12-17

Similar Documents

Publication Publication Date Title
EP1175785B1 (fr) Acheminement en reseau de fichiers supports profiles vers des clients
Marcellin et al. An overview of JPEG-2000
WO2000065838A2 (fr) Conversion d'un fichier media en format variable pour une transmission progressive
US20050166245A1 (en) Method and device for transmitting scalable video bitstream
KR100664928B1 (ko) 비디오 코딩 방법 및 장치
JP4318918B2 (ja) ビデオのスケーラブル圧縮方法および装置
US20090316778A1 (en) Method And System For Optimal Video Transcoding Based On Utility Function Descriptors
Lee JPEG 2000: Retrospective and new developments
US8014451B2 (en) Video encoder/decoder with macroblock arrangement of significant item
US6828967B1 (en) Method of and apparatus for digital data storage
US20090003270A1 (en) Server-Driven Progressive Image Transmission
Liang Highly scalable image coding for multimedia applications
CN101888553B (zh) 用于可伸缩视频编码的方法和装置
WO2000065834A1 (fr) Procede de copie de fichier media
Auli-Llinas et al. Enhanced JPEG2000 quality scalability through block-wise layer truncation
Sparenberg et al. A concept for file-based content exchange using scalable media
Pandiammal et al. An Enhanced Approach for Video Compression
Zhang et al. High performance full scalable video compression with embedded multiresolution MC-3DSPIHT
Akujuobi Application of Wavelets to Video Compression
Barina et al. JPEG 2000: guide for digital libraries
Bojkovic et al. Advanced image compression techniques: Comparative functionalities and performances
Danyali et al. Fully scalable texture coding of arbitrarily shaped video objects
Zhao et al. Highly scalable differential JPEG 2000 wavelet video codec for Internet video streaming
MILOVANOVIC et al. DEVELOPMENT OF WAVELET TECHNIQUES IN JPEG2000 IMAGE CODING STANDARD
Cheng et al. Audio/video compression applications using wavelets

Legal Events

Date Code Title Description
AK Designated states

Kind code of ref document: A2

Designated state(s): GB JP US

AL Designated countries for regional patents

Kind code of ref document: A2

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE

WWE Wipo information: entry into national phase

Ref document number: 2000927450

Country of ref document: EP

WWE Wipo information: entry into national phase

Ref document number: 09720588

Country of ref document: US

121 Ep: the epo has been informed by wipo that ep was designated in this application
WWP Wipo information: published in national office

Ref document number: 2000927450

Country of ref document: EP

AK Designated states

Kind code of ref document: A3

Designated state(s): GB JP US

AL Designated countries for regional patents

Kind code of ref document: A3

Designated state(s): AT BE CH CY DE DK ES FI FR GB GR IE IT LU MC NL PT SE

WWW Wipo information: withdrawn in national office

Ref document number: 2000927450

Country of ref document: EP