Arithmetic coding approaches source entropy better than table-based Huffman methods

Evidence

In order to have longer symbol lengths, which allow greater coding efficiencies, other approaches are used. Golomb-Rice codes use a predetermined statistical distribution to remove the need for tables, thus allowing longer code words; they are used in some lossless encoding schemes. Another approach is arithmetic coding, which does not need predetermined distributions, but, by using the statistics of the symbols occurrence, can encode very long symbols and so approach the source entropy much more closely. — David M. Howard, Jamie Angus, Acoustics and Psychoacoustics, 2644

In order to have longer symbol lengths, which allow greater coding effi­ ciencies, other approaches are used. Golomb Rice Codes use a predeter­ mined statistical distribution to remove the need for tables, thus allowing longer code words; they are used in some lossless encoding schemes. Another approach is Arithmetic Coding, which does not need predeter­ mined distributions but, by using the statistics of the symbols occurrence, can encode very long symbols and so approach the source entropy much more closely. — David Howard, Jamie Angus, coustics and Psychoacoustics, 2641

Categorization

• XREF: Golomb-Rice and arithmetic coding recur in lossless audio formats like FLAC and in MPEG codecs. — connects to: A data source’s entropy gives the minimum bits needed to encode it — David Howard, Jamie Angus, coustics and Psychoacoustics, 2625; Entropy coding trade-offs limit how close code rates approach information-theoretic limits; Quantizer levels need not be powers of two when entropy coding is used — David M. Howard, Jamie Angus, Acoustics and Psychoacoustics, 2656 (+5 more)
• QUESTION: How close to entropy do practical lossless audio codecs actually get, and where is the gap felt?
• corroborated: 3 independent passages support this claim — David M. Howard, Jamie Angus, Acoustics and Psychoacoustics, 2644; David Howard, Jamie Angus, coustics and Psychoacoustics, 2641

Broader Connections

• Quantizer levels need not be powers of two when entropy coding is used #expands (0.80) — David M. Howard, Jamie Angus, Acoustics and Psychoacoustics, 2656
• Information compression works by removing only the predictable parts of a signal #corroborates (0.75)
• Audio compression maximizes information by encoding only unpredictable signal parts #corroborates (0.70)
• Perceptual coding works by maximizing the surprise value of coded signals #corroborates (0.70)
• Lossless audio coding works by predicting and stripping signal correlations #corroborates (0.70)