eln1

7 ms·

31.
by eln1 10y ago
Lorentz invariance appears naturally in practically all theories with waves - its violation would be a huge surprise. To see it, understand STR, the perfect model is sine-Gordon: just many coupled pendula - we get particles ("kinks&quo
32.
by eln1 10y ago
As the main hypothesis in this paper is that the momentum is transferred by pilot waves, it is worth to emphasize the (cited there) most known recent approaches - Couder's experiments which use classical object with wave-particle duali
33.
GST: GPU-decodable Supercompressed Textures (~2x smaller than standard DXT1)(gamma.cs.unc.edu)
3 points by eln1 10y ago|0 comments
34.
by eln1 10y ago
Golomb-Rice with base M is prefix code optimal for approximately geometric probability distribution Pr(x) ~ sqrt(2)^(-Mx). Arithmetic coding or FSE/tANS would allow to use the actual probability distribution. The question is how large
35.
by eln1 10y ago
It depends if you have static or adaptive coder. Static is much cheaper, uses the same probabilities for the entire data block (e.g. 30 kB), probabilities are stored in the header - practically all Huffman and tANS compressors (however, the
36.
by eln1 10y ago
Before Yann, this ANS variant was implemented by Andrew Polar in 2008: http://www.ezcodesample.com/abs/abs_article.html Here is a list of implementations: http://encode.ru/threads/2078-List-of-Asym
37.
Wikipedia article about Asymmetric Numeral Systems(en.wikipedia.org)
1 points by eln1 10y ago|0 comments
38.
by eln1 10y ago
Code: https://aomedia.googlesource.com/aom/+/master
39.
AV1 royalty-free video codec from AOM (Google,Mozilla,Cisco,Intel,Microsoft,...)(en.wikipedia.org)
2 points by eln1 10y ago|1 comments
40.
Zstd has just landed in Blosc(blosc.org)
3 points by eln1 10y ago|0 comments
41.
by eln1 10y ago
WebP is switching to rANS: https://chromium-review.googlesource.com/#/c/338781/ Here is a superfast implementation of rANS: https://github.com/jkbonfield/rans_static
42.
by eln1 10y ago
Here are explanations of various people: http://www.ezcodesample.com/abs/abs_article.html http://cbloomrants.blogspot.com/2014/02/02-18-14-understandi... http://fastcompression.blo
43.
by eln1 10y ago
It points to even older paper for ANS: http://arxiv.org/abs/0710.3861 The more recent one is: http://arxiv.org/abs/1311.2540 The first one is about kind of multidimensional analogue of Fibonacci c
44.
by eln1 10y ago
The difference is size probably come from suboptimalities which were repaired after open-sourcing: http://encode.ru/threads/2221-LZFSE-New-Apple-Data-Compressi... The difference in speed is more surprising, it could co
45.
by eln1 10y ago
It is poorly optimized for 32bits - some benchmarks: http://encode.ru/threads/2221-LZFSE-New-Apple-Data-Compressi...
46.
by eln1 10y ago
Arithmetic coding is very expensive - zstd uses new entropy coding with compression ratio like arithmetic coding, but with Huffman-like speed: https://github.com/Cyan4973/zstd https://github.com/Cyan497
47.
by eln1 10y ago
There is an attempt to patent something ANS-related: http://cbloomrants.blogspot.com/2015/05/05-21-15-software-pa...
48.
by eln1 10y ago
Generally, a symbol of probability p contains lg(1/p) bits of information: 1 bit if p=1/2, 2 bits if p=1/4 etc. In Huffman you directly assign a concrete bit sequence to every symbol - it is perfect if their probabilities are
49.
by eln1 10y ago
Beside Huffman coding, there is also arithmetic/range coding and now is becoming widely used ANS coding - e.g. LZFSE (Apple), ZSTD (Facebook), VP10 (Google): http://encode.ru/threads/2078-List-of-Asymmetric-Numeral
50.
by eln1 10y ago
Apple tANS implementation was suboptimal - both from the quantization and symbol spread point of view: http://encode.ru/threads/2221-LZFSE-New-Apple-Data-Compressi... The author of this post (Eric Biggers) has made cor
51.
by eln1 10y ago
Google VP10 and WebP are also planned to be ANS-based: https://chromium.googlesource.com/webm/libvpx/+/fb9186d68d49... https://chromium-review.googlesource.com/#/c/338781/
52.
by eln1 10y ago
Also lzturbo is very interesting, closed source but you one can download exec: https://sites.google.com/site/powturbo/ https://github.com/powturbo/TurboBench
53.
by eln1 11y ago
Let's look at the most interesting there: fastest encoding: 210,194,996 bytes, 0.333 sec., 0.371 sec., lz4 -1 fastest decoding: 169,476,113 bytes, 2.072 sec., 0.352 sec., lz4 -9 149,642,742 bytes, 2.069 sec., 0.646 sec., zhuff_beta -c0
54.
by eln1 11y ago
Here are some benchmarks and discussion of data compression specialists: http://encode.ru/threads/2313-Brotli?p=44970&viewfull=1#post... "Input: 812,392,384 bytes, HTML top 10k Alexa crawled (8,998 with HTML r
55.
by eln1 11y ago
FSE is implementation of this new ANS entropy coding, also used e.g. in lzturbo 1.2. It's surprising that even still using huffman, brotli is slowly approaching its performance: http://encode.ru/threads/2313-Brotli