Table of Contents
QccENTExponentialGolombEncode, QccENTExponentialGolombDecode, QccENTExponentialGolombEncodeSymbol,
QccENTExponentialGolombDecodeSymbol - exponential-Golomb encoding and decoding
#include "libQccPack.h"
int QccENTExponentialGolombEncode(QccBitBuffer
*output_buffer, const int *symbols, int num_symbols, int signed_symbols);
int QccENTExponentialGolombDecode(QccBitBuffer *input_buffer, int *symbols,
int num_symbols, int signed_symbols);
int QccENTExponentialGolombEncodeSymbol(QccBitBuffer
*output_buffer, int symbol, int signed_symbol);
int QccENTExponentialGolombDecodeSymbol(QccBitBuffer
*input_buffer, int *symbol, int signed_symbol);
QccENTExponentialGolombEncode()
performs exponential-Golomb coding of the symbol stream in symbols, outputting
the binary codewords to output_buffer. num_symbols indicates how many symbols
are in the symbol stream (which can be as short as one symbol). signed_symbols
indicates whether the symbols in the symbol stream are signed integers
(signed_symbols = 1) or unsigned integers (signed_symbols = 0). output_buffer
must be a QccBitBuffer(3)
that has been opened for writing prior to calling
QccENTExponentialGolombEncode(). After calling QccENTExponentialGolombEncode(),
QccBitBufferFlush(3)
must be called to ensure that the last few bits of
the bitstream are actually written to the output file.
QccENTExponentialGolombDecode()
performs exponential-Golomb decoding of the bits in the bitstream input_buffer,
producing an output stream of symbols that are stored in symbols. The
symbols array must be allocated with space sufficient for holding num_symbols
integers; this allocation must be done prior to calling QccENTExponentialGolombDecode().
Additionally, input_buffer must be opened for reading prior to calling
QccENTExponentialGolombDecode().
QccENTExponentialGolombEncodeSymbol()
and QccENTExponentialGolombDecodeSymbol() encode and decode a single symbol.
QccENTExponentialGolombEncode() and QccENTExponentialGolombDecode() are
implemented simply as repeated calls to QccENTExponentialGolombEncodeSymbol()
and QccENTExponentialGolombDecodeSymbol(), respectively.
Golomb
coding was originally proposed by Golomb in 1966, and a special case of
Golomb coding with the Golomb-coding parameter constrained to be a power
of 2 was proposed independently by Rice in 1979, with these latter codes
often called Rice codes or Golomb-Rice codes. Exponential-Golomb codes were
proposed by Teuhola as another variant of Golomb codes similar to Rice
codes with the property that the number of codes with a given length grows
exponentially with the length. The implementation of exponential-Golomb
coding follows that described in the H.264 standard (Sec. 9.1). Although the
codeword lengths produced by this implementation are exactly the same
as in H.264, the actual codewords are somewhat different.
These routines return 0 on success, and 1 on failure.
QccBitBuffer(3)
,
QccPackENT(3)
, QccPack(3)
S. W. Golomb, "Run-Length Encodings," IEEE Transactions
on Information Theory, vol. 12, pp. 399-401, July 1966.
R. F. Rice, "Some Practical
Universal Noiseless Coding Techniques," Tech. Rep. JPL-79-22, Jet Propulsion
Laboratory, Pasadena, CA, March 1979.
J. Teuhola, "A Compression Method
for Clustered Bit-Vectors," Information Processing Letters, vol. 7, pp. 308-311,
October 1978.
ITU-T, Advanced Video Coding for Generic Audiovisual Services,
May 2003, ITU-T Recommendation H.264.
Copyright (C) 1997-2021 James
E. Fowler
Table of Contents
