dirac::ArithCodec< T > Class Template Reference

Abstract binary arithmetic coding class. More...

#include <arith_codec.h>

List of all members.

Public Member Functions
	ArithCodec (BasicOutputManager *bits_out, size_t number_of_contexts)
	Constructor for encoding.
	ArithCodec (BitInputManager *bits_in, size_t number_of_contexts)
	Constructor for decoding.
virtual	~ArithCodec ()
	Destructor.
int	Compress (T &in_data)
	Compresses the input and returns the number of bits written.
void	Decompress (T &out_data, int num_bits)
	Decompresses the bitstream and writes into the output.
Protected Types
typedef uint16_t	code_t
typedef uint32_t	calc_t
Protected Member Functions
virtual void	InitContexts ()=0
	The method by which the contexts are initialised.
virtual void	Update (const bool symbol, const int context_num)=0
	The method by which the counts are updated.
virtual void	Resize (const int context_num)=0
	The method by which the counts are resized.
virtual void	ResetAll ()=0
	The method by which _all_ the counts are resized.
virtual int	ChooseContext (const T &data, const int bin_number) const =0
	Choose the context based on previous data.
virtual int	ChooseContext (const T &data) const =0
	Choose the context based on previous data.
virtual void	DoWorkCode (T &in_data)=0
	Does the work of actually coding the data.
void	InitEncoder ()
	Initialises the Encoder.
void	EncodeTriple (const Triple &c)
	encodes a triple and writes to output
void	EncodeSymbol (const bool symbol, const int context_num)
	encodes a symbol and writes to output
void	FlushEncoder ()
	flushes the output of the encoder.
void	FlushDecoder ()
	flushes the input of the encoder by reading all the remaining bits
virtual void	DoWorkDecode (T &out_data, int num_bits)=0
	virtual decode-only functions
void	InitDecoder ()
	Initialise the Decoder.
void	RemFromStream (const Triple &c)
	Remove the symbol from the coded input stream.
void	SetCurrentCount (const int context_num)
	Rescales the interval midpoint according to the context.
void	DecodeSymbol (bool &symbol, int context_num)
	Decodes a symbol given a context number.
Protected Attributes
std::vector< Context >	m_context_list
	List of contexts.
Static Protected Attributes
static const code_t	CODE_MAX = 0xffff
static const code_t	CODE_MSB = ((0xffff + 1) >> 1)
static const code_t	CODE_2ND_MSB = ((0xffff + 1) >> 2)
Classes
class	Context
	A class for binary contexts. More...
class	Triple
	A class for encapsulating interval fractions for use in arithmetic coding. More...

---

Detailed Description

template<class T>
class dirac::ArithCodec< T >

Abstract binary arithmetic coding class.

This is an abtract binary arithmetic encoding class, used as the base for concrete classes that encode motion vectors and subband residues.

Parameters:

T

a container (most probably, or array) type

Definition at line 73 of file arith_codec.h.

---

Constructor & Destructor Documentation

template<class T> dirac::ArithCodec< T >::ArithCodec (  BasicOutputManager *  bits_out, size_t  number_of_contexts )

Constructor for encoding. Creates an ArithCodec object to decode input based on a set of parameters. Parameters: bits_out  output for encoded bits number_of_contexts  the number of contexts used Definition at line 457 of file arith_codec.h. Referenced by dirac::ArithCodec< T >::Context::GetSymbol().

template<class T> dirac::ArithCodec< T >::ArithCodec (  BitInputManager *  bits_in, size_t  number_of_contexts )

Constructor for decoding. Creates an ArithCodec object to decode input based on a set of parameters. Parameters: bits_in  source of bits to be decoded number_of_contexts  the number of contexts used Definition at line 447 of file arith_codec.h.

template<class T> dirac::ArithCodec< T >::~ArithCodec (   )  [virtual]

Destructor. Destructor is virtual as this class is abstract. Definition at line 466 of file arith_codec.h.

---

Member Function Documentation

template<class T> int dirac::ArithCodec< T >::Compress (  T &  in_data  )

Compresses the input and returns the number of bits written. Compress takes a type T object (a container or array) and compresses it using the abstract function DoWorkCode() which is overridden in subclasses. It returns the number of bits written. Parameters: in_data  the input to be compressed. Non-const, since the compression may be lossy. Definition at line 472 of file arith_codec.h. References dirac::ArithCodec< T >::DoWorkCode(), dirac::ArithCodec< T >::FlushEncoder(), and dirac::ArithCodec< T >::InitEncoder(). Referenced by dirac::FrameCompressor::Compress(), and dirac::CompCompressor::Compress().

template<class T> void dirac::ArithCodec< T >::Decompress (  T &  out_data, int  num_bits )

Decompresses the bitstream and writes into the output. Decompresses the bitstream, up to the number of bits specified and writes into the output subclasses. Parameters: out_data  the output into which the decompressed data is written. num_bits  the number of bits to be read from the bitstream. [May be eliminated in future revisions. TJD 13 April 2004.] Definition at line 481 of file arith_codec.h. References dirac::ArithCodec< T >::DoWorkDecode(), dirac::ArithCodec< T >::FlushDecoder(), dirac::BitInputManager::FlushInput(), and dirac::ArithCodec< T >::InitDecoder(). Referenced by dirac::CompDecompressor::Decompress().

---

The documentation for this class was generated from the following file:

• C:/guliverkli/src/filters/parser/diracsplitter/libdirac/libdirac_common/arith_codec.h

---