Classifications

• • H—ELECTRICITY
  • H03—ELECTRONIC CIRCUITRY
  • H03M—CODING; DECODING; CODE CONVERSION IN GENERAL
  • H03M7/00—Conversion of a code where information is represented by a given sequence or number of digits to a code where the same, similar or subset of information is represented by a different sequence or number of digits
  • H03M7/30—Compression; Expansion; Suppression of unnecessary data, e.g. redundancy reduction
  • H03M7/40—Conversion to or from variable length codes, e.g. Shannon-Fano code, Huffman code, Morse code
  • H03M7/42—Conversion to or from variable length codes, e.g. Shannon-Fano code, Huffman code, Morse code using table look-up for the coding or decoding process, e.g. using read-only memory

Definitions

• the invention generally relates to data compression, and more specifically relates to a form of entropy coding.
• data is represented as a long sequence of symbols which is input to an encoder.
• the input data is encoded by an encoder, transmitted over a communication channel (or simply stored), and decoded by a decoder.
• the input is pre-processed, sampled, converted, compressed or otherwise manipulated into a form for transmission or storage. After transmission or storage, the decoder attempts to reconstruct the original input.
• Audio coding techniques can be generally categorized into two classes, namely the time-domain techniques and frequency-domain ones.
• Time-domain techniques e.g., ADPCM, LPC
• the frequency-domain techniques transform the audio signals into the frequency domain where the actual compression happens.
• the frequency-domain codecs can be further separated into either subband or transform coders although the distinction between the two is not always clear.
• Processing an audio signal in the frequency domain is motivated by both classical signal processing theories and human perception models (e.g., psychoaoustics).
• the inner ear specifically the basilar membrane, behaves like a spectral analyzer and transforms the audio signal into spectral data before further neural processing proceeds.
• the frequency-domain audio codecs often take advantage of many kinds of auditory masking that are going on with the human hear system to modify the original signal and eliminate a great many details/redundancies. Since the human ears are not capable of perceiving these modifications, efficient compression is achieved.
• Masking is usually conducted in conjunction with quantization so that quantization noise can be conveniently "masked.”
• the quantized spectral data are usually further compressed by applying entropy coding, e.g., Huffman coding. Compression is required because a fundamental limitation of the communication model is that transmission channels usually have limited capacity or bandwidth. Consequently, it is frequently necessary to reduce the information content of input data in order to allow it to be reliably transmitted, if at all, over the communication channel.
• Huffman encoding which is a particular form of entropy encoding.
• Entropy coding assigns code words to different input sequences, and stores all input sequences in a code book.
• the complexity of entropy encoding depends on the number m of possible values an input sequence X may take. For small m, there are few possible input combinations, and therefore the code book for the messages can be very small (e.g., only a few bits are needed to unambiguously represent all possible input sequences).
• the code alphabet is most likely a series of binary digits ⁇ 0, 1 ⁇ , and code word lengths are measured in bits.
• variable length sequences are assigned fixed length codes (or conversely, variable length sequences are assigned fixed length codes).
• variable-to-variable entropy encoder codes variable length input sequences with variable length codes.
• entropy-type codes may be assigned to only probable inputs, and alternate codes used to identify less probable sequences.
• each section may further be sorted according to the probability associated with the entry. For example, each section may be sorted with highest probable inputs located first in the section, thus increasing the likelihood that a match will be found quickly.
• Matching code cook entries depends on the internal representation of the book. For example, in a tree structure, nodes may represent each character of the input such that reaching a leaf signifies the end and identification of a particular grouping of input symbols.
• a pattern matching algorithm can be applied to each table entry within the appropriate section.
• searching may be facilitated by recognition that only as many input symbols as the longest grouping in the code book section need to be considered.
• the corresponding entropy-type code can be output and the search repeated with the next symbol following the matched input.
• the input stream is expected to be any data stream, such as numbers, characters, or a binary data which encodes audio, video or other types of data.
• the input stream is referenced herein as a series of symbols, where each "symbol” refers to the appropriate measurement unit for the particular input.
• the input stream may originate from local storage, or from intranets, the Internet, or streaming data (e.g., Microsoft's "NETSHOW”TM client/server streaming architecture).
• FIG. 1 is a block diagram of a computer system that may be used to implement variable to variable entropy encoding.
• FIG. 2 shows a basic communication model for transmitting streaming and non-streaming data.
• FIG. 3 is a flowchart showing creation of a code book having variable length entries for variable length symbol groupings.
• FIGS. 4-10 illustrate creation of a code book pursuant to FIG. 3 for an alphabet ⁇ A, B, C ⁇ .
• FIG. 1 1 shows encoding of audio data.
• FIG. 12 illustrates an entropy encoder.
• the invention has been implemented in an audio/visual codec (compressor/de-compressor). This is only one example of how the invention may be implemented.
• the invention is designed to be utilized wherever entropy-type coding may be utilized, and is applicable to compression of any type of data. Briefly described, optimal entropy encoding requires excessive resources, and the illustrated embodiments provide a nearly optimal encoding solution requiring far fewer resources.
• FIG. 1 and the following discussion are intended to provide a brief, general description of a suitable computing environment in which the invention may be implemented. While the invention will be described in the general context of computer- executable instructions of a computer program that runs on a personal computer, those skilled in the art will recognize that the invention also may be implemented in combination with other program modules. Generally, program modules include routines, programs, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Moreover, those skilled in the art will appreciate that the invention may be practiced with other computer system configurations, including hand-held devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, minicomputers, mainframe computers, and the like.
• the illustrated embodiment of the invention also is practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. But, some embodiments of the invention can be practiced on stand alone computers. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.
• an exemplary system for implementing the invention includes a computer 20, including a processing unit 21 , a system memory 22, and a system bus 23 that couples various system components including the system memory to the processing unit 21 .
• the processing unit may be any of various commercially available processors, including Intel x86, Pentium and compatible microprocessors from Intel and others, the Alpha processor by Digital, and the PowerPC from IBM and Motorola. Dual microprocessors and other multi-processor architectures also can be used as the processing unit 21.
• the system bus may be any of several types of bus structure including a memory bus or memory controller, a peripheral bus, and a local bus using any of a variety of conventional bus architectures such as PCI, AGP, VESA, MicroChannel, ISA and EISA, to name a few.
• the system memory includes read only memory (ROM) 24 and random access memory (RAM) 25.
• ROM read only memory
• RAM random access memory
• BIOS basic input/output system
• BIOS basic routines that help to transfer information between elements within the computer 20, such as during start-up, is stored in ROM 24.
• the computer 20 further includes a hard disk drive 27, a magnetic disk drive 28, e.g., to read from or write to a removable disk 29, and an optical disk drive 30, e.g., for reading a CD-ROM disk 31 or to read from or write to other optical media.
• the hard disk drive 27, magnetic disk drive 28, and optical disk drive 30 are connected to the system bus 23 by a hard disk drive interface 32, a magnetic disk drive interface 33, and an optical drive interface 34, respectively.
• the drives and their associated computer- readable media provide nonvolatile storage of data, data structures, computer-executable instructions, etc. for the computer 20.
• a number of program modules may be stored in the drives and RAM 25, including an operating system 35, one or more application programs (e.g., Internet browser software) 36, other program modules 37, and program data 38.
• application programs e.g., Internet browser software
• a user may enter commands and information into the computer 20 through a keyboard 40 and pointing device, such as a mouse 42.
• Other input devices may include a microphone, joystick, game pad, satellite dish, scanner, or the like.
• These and other input devices are often connected to the processing unit 21 through a serial port interface 46 that is coupled to the system bus, but may be connected by other interfaces, such as a parallel port, game port or a universal serial bus (USB).
• a monitor 47 or other type of display device is also connected to the system bus 23 via an interface, such as a video adapter 48.
• personal computers typically include other peripheral output devices (not shown), such as speakers and printers.
• the computer 20 is expected to operate in a networked environment using logical connections to one or more remote computers, such as a remote computer 49.
• the remote computer 49 may be a web server, a router, a peer device or other common network node, and typically includes many or all of the elements described relative to the computer 20, although only a memory storage device 50 has been illustrated in FIG. 1 .
• the computer 20 can contact the remote computer 49 over an Internet connection established through a Gateway 55 (e.g., a router, dedicated-line, or other network link), a modem 54 link, or by an intra-office local area network (LAN) 51 or wide area network (WAN) 52.
• LAN local area network
• WAN wide area network
• the present invention is described below with reference to acts and symbolic representations of operations that are performed by the computer 20, unless indicated otherwise. Such acts and operations are sometimes referred to as being computer-executed. It will be appreciated that the acts and symbolically represented operations include the manipulation by the processing unit 21 of electrical signals representing data bits which causes a resulting transformation or reduction of the electrical signal representation, and the maintenance of data bits at memory locations in the memory system (including the system memory 22, hard drive 27, floppy disks 29, and CD-ROM 31 ) to thereby reconfigure or otherwise alter the computer system's operation, as well as other processing of signals.
• the memory locations where data bits are maintained are physical locations that have particular electrical, magnetic, or optical properties corresponding to the data bits.
• FIG. 2 shows a basic communication model.
• a data source or sender 200 there is a data source or sender 200, a communication channel 204, and a data receiver 208.
• the source may be someone speaking on a telephone, over telephone wires, to another person.
• the source may be a television or radio broadcast transmitted by wireless methods to television or radio receivers.
• the source may be a digital encoding of some data, whether audio, visual, or other, transmitted over a wired or wireless communication link (e.g., a LAN or the Internet) to a corresponding decoder for the information.
• a wired or wireless communication link e.g., a LAN or the Internet
• an encoder 202 is used to prepare the data source for transmission over the communication channel 204.
• the encoder is responsible for converting the source data into a format appropriate for the channel 204.
• a format appropriate for the channel 204 For example, in the context of a common telephone call, one's voice is typically converted by the phone's handset from voice sounds to analog impulses that are sent as analog data to local telephone receiving equipment. This analog signal is then converted into digital form, multiplexed with numerous other conversations similarly encoded, and transmitted over a common line towards the receiver.
• the channel 204 corresponds in large part to a common pathway shared by multiple senders and receivers.
• the channel 204 is commonly an intranet or the Internet.
• a decoder 206 is required to reverse the encoding process so as to present sensible data to the receiver.
• This simple model does not take into account, however, the real-world demands of application programs.
• a client e.g., an application program
• a streaming delivery system is required, i.e., an adaptive data transmission system that allows application-level bandwidth reservation for a data stream.
• Streaming environments contrast traditional networking programs, such as certain versions of Internet browsers that download web page content on a non-prioritized basis, and allow data content delivery over the network link 240 to be orchestrated (and optimized) for particular retrieval needs (such as a slow dial-up link).
• An exemplary streaming format is the Microsoft Active Streaming Format.
• a SF defines the structure of complex, synchronized object data streams. Any object can be placed into a SF data stream, including audio and video data objects, scripts, ActiveX controls, and HTML documents.
• SF data can be stored in a file or created in real-time using audio and video capture hardware.
• An Application Programming Interface corresponding to an implementation of the SF can provide an application with support for delivering and receiving streaming content.
• One such API is the Microsoft Audio Compression Manager (ACM), which provides functions for processing (e.g., compressing and delivering) audio data.
• ACM Microsoft Audio Compression Manager
• Win32 Internet Extensions (Winlnet), WinSock, and TCP/IP APIs.
• Winlnet Microsoft Win32 Internet Extensions
• WinSock WinSock
• TCP/IP APIs TCP/IP
• networks such as the internet convert the source data into packet form suitable for the network.
• Packets generally include routing information as well as the actual data.
• SF data streams are preferably made up of data packets that can be transmitted over any digital network by inserting them one at a time into the data field of network packets.
• Each SF packet may contain a prioritized mix of data from different objects within the stream, so that the bandwidth can be concentrated on higher priority objects (or organized to optimize throughput).
• This data can be captured in real time, stored to nonvolatile storage, converted from existing audio or video formats, created by combining audio with pictures and scripts, or delivered over the network to a client program or viewer.
• the client receiver 208 of the streaming data can be a traditional "helper" application (for compatibility with the old Web publishing approach), or a more modern web page control (e.g., an ActiveX object) embedded in a web page.
• SF data streams are distinguished over traditional network content as being viewed progressively in real time as a client receives it.
• playback of streamed content becomes susceptible to transmission delays. If data does not arrive reliably, or if transmission speed falls below an acceptable minimum, playback of the content cannot continue at an acceptable rate.
• Smooth-streaming playback at a client requires that the transmission require a bandwidth less than the client's available bandwidth (e.g. the speed of the link 204 less networking overhead).
• a dial-up connection to the Internet provides a bandwidth of 28-34 Kbps. Consequently, audiovisual source data (which is bandwidth intensive) must be significantly compressed to allow its transmission over low bit-rate connections. The degree of compression necessarily impacts the quality of the reproduced signal.
• a server provides multiple sources optimized for different networking speeds, or utilizes an adaptive feedback system to perform real-time analysis of the client's actual throughput.
• SF data packets are encoded 202 and placed inside network packets and sent over the network 204, the routing technology of the network takes care of delivering the network packets to the receiver 208.
• a variety of network and application protocols such as UDP, TCP, RTP, IP Multicast, IPX, and HTTP, are supported by the broadcast sender 200.
• Entropy coding capitalizes on data coherency, and is effective when symbols have non-uniform probability distribution.
• FIG. 3 is a flowchart showing a preferred method for generating an entropy encoder's code book.
• FIG. 3 illustrates how to create a code book having variable length code assignments for variable length symbol groupings.
• prior art techniques either require fixed-length codes or fixed blocks of input.
• Preferred implementations overcome the resource requirements of large dimension vector encoding, and the inapplicability of coding into words of equal lengths, by providing an entropy based variable-to-variable code, where variable length code words are used to encode variable length X sequences.
• Equation L 1 N *i .
• the leaf-node having highest probability is selected for extension (see also FIG. 3 discussion regarding processing sequence).
• group A 400 is tentatively expanded by each character in the alphabet (or one may limit the expansion to some subset thereof as described for creating the initial grouping).
• Probabilities are then recomputed with respect to the input stream "A A A B B A A C A B A B B A B” to determine values for the tentative extensions A 406, B 408, and C 410.
• the result is nine parsing groups, where "A A” appears 2/9, "A B” appears 4/9, and "A C" appears 0/9.
• FIG. 6 shows the collapse into X1 for FIG. 5. Processing repeats with identification of the node having highest probability, e.g., node B 408 at probability 4/9.
• this node 408 is tentatively extended with symbols A 414, B 416, C 418, and as discussed above, the tentative grouping with highest probability is retained. After recalculating probabilities, the result is eight parsing groups in which the symbol sequence "A B A” 414 appears once, "A B B” 416 appears once, and
• the next step is to expand the node currently having highest probability with respect to the input stream.
• the highest node in the tree (XO 402) is extended. (Although it is only necessary to be consistent, it is preferable to expand higher level nodes since this may increase coding efficiency by increasing the number of long code words.)
• FIG. 1 1 shows a transmission model for transmitting audio data over a channel 460.
• channel 460 is of limited bandwidth, and therefore some compression of source data 450 is required before the data can be reliably sent.
• this discussion focuses on transmission of audio data, the invention applies to transfer of other data, such as audio visual information having embedded audio data (e.g., multiplexed within an MPEG data stream), or other data sources having compressible data patterns (e.g., coherent data).
• source data 450 is input to a time / frequency transform encoder 352 such as a filter bank or discrete-cosine type transform.
• Transform encoder 452 is designed so as to convert a continuous or sampled time-domain input, such as an audio data source, into multiple frequency bands of predetermined (although perhaps differing) bandwidth. These bands can then be analyzed with respect to a human auditory perception model 454 (for example, a psychoacoustic model) in order to determine components of the signal that may be safely reduced without audible impact. For example, it is well known that certain frequencies are inaudible when certain other sounds or frequencies are present in the input signal (simultaneous masking). Consequently, such inaudible signals can be safely removed from the input signal.
• Use of human auditory models is well known, e.g., the MPEG 1 , 2 and 4 standards. (Note that such models may be combined into a quantization 456 operation.)
• frequency coefficients within each range are quantized 456 to convert each coefficient (amplitude levels) to a value taken from a finite set of possible values, where each value has a size based on the bits allocated to representing the frequency range.
• the quantizer may be a conventional uniform or non-uniform quantizer, such as a midriser or midtreader quantizer with (or without) memory.
• the general quantization goal is identifying an optimum bit allocation for representing the input signal data, i.e., to distribute usage of available encoding bits to ensure encoding the (acoustically) significant portions of the source data.
• Various quantization methods such as quantization step size prediction to meet a desired bit rate (assuming constant bit rate) can be used.
• FIG. 1 2 shows one method for implementing the entropy encoder 458 of FIG. 1 1 through application of the code book of FIG. 3 to the quantized data.
• the code book for variable-to-variable encoding can be used to encode other types of data.
• the quantized data is received 480 as input to the entropy encoder 458 of FIG. 1 1 . It is understood that the input is in some form of discrete signals or data packets, and that for simplicity of discussion, all input is simply assumed to be a long series of discrete symbols.
• the received input 480 is scanned 482 in order to locate 484 a corresponding code book key in the code book of FIG. 3. Such scanning corresponds to a data look-up, and depending on how the data structure used to implement the code book, the exact method of look-up will vary.
• One structure for a code book is traversal and storage of a N-ary (e.g., binary, tertiary, etc.) tree, where symbol groupings guide a traversal of the tree structure.
• the path to a leaf node of the tree represents the end of a recognized symbol sequence, where a entropy-type code is associated with the sequence.
• Nodes can be coded in software as a structure, class definition, or other structure allowing storage of a symbol or symbols associated with the node, and association of a corresponding entropy-type code 486.
• the code book may be structured as a table having each string of input symbol sorted by probability of occurrence, with highly probable input at the top of the table.
• the table can be sorted according to the first symbol, i.e., all symbol series beginning with "A” are grouped together, followed by series starting with "B", etc. With this arrangement, all entries within the grouping are sorted according to their probabilities of occurrence. The position of the beginning of each section is marked/tracked so that a hash-type function (e.g., a look-up based on the first symbol) can be used to locate the correct portion of the code book table.
• a hash-type function e.g., a look-up based on the first symbol
• the corresponding table section is exhaustively searched until a matching entry is located.
• the code 484 associated with the matching entry is then output 486 as the encoded substitute.
• this output is transmitted over the communication channel 460.
• the receiving end 470 implements a reverse-encoding process, i.e., a series of steps to undo the encoding of the source data 450. That is, the encoded data 486 is received as input to an entropy decoder 462 which performs a reverse code book look-up to convert the encoded output 486 back into the original input symbol series 480 (FIG. 12).
• the recovered input data 480 is then processed by a de-quantizer 464 and time/frequency transform decoder 466 to reverse the original coding operations, resulting in a reconstructed data 468 that is similar to the original source data 450. It should be noted that the reconstructed data 468 only approximates the original source data 450 when, as it presumed herein, a lossy system is employed.

Landscapes

• Engineering & Computer Science (AREA)
• Theoretical Computer Science (AREA)
• Compression, Expansion, Code Conversion, And Decoders (AREA)
• Vehicle Body Suspensions (AREA)

Priority Applications (4)

Applications Claiming Priority (2)

Publications (1)

Family

ID=22786310

Family Applications (1)

Country Status (6)

Cited By (5)

Families Citing this family (44)

Citations (7)

Family Cites Families (36)

• 1998
  • 1998-12-14 US US09/211,294 patent/US6377930B1/en not_active Expired - Lifetime
• 1999
  • 1999-12-07 WO PCT/US1999/029108 patent/WO2000036752A1/en not_active Ceased
  • 1999-12-07 JP JP2000588900A patent/JP4559632B2/ja not_active Expired - Lifetime
  • 1999-12-07 EP EP99966062A patent/EP1142129B1/en not_active Expired - Lifetime
  • 1999-12-07 DE DE69918172T patent/DE69918172T2/de not_active Expired - Lifetime
  • 1999-12-07 AT AT99966062T patent/ATE269605T1/de not_active IP Right Cessation

Patent Citations (8)

Non-Patent Citations (2)

Also Published As

Similar Documents

Legal Events