|
 |
 |
 |
 |
 |
 |
The Art
of Lossless
Data Compression
vol. 26t
Here are the results of tests performed in December 2003 to compare
lossless compression of "plain" texts by all known good enough programs
developed for such purpose, including UHArc, PPMd, Bzip2, RAR, ACE and 7-zip.
See Archive Comparison Test by J.Gilchrist for more details:
http://compression.ca
If anybody wants to start or continue such tests,
or can suggest some other sets of texts, or other compression programs,
(not sources or algorithm descriptions, executable programs only)
or knows we have missed something important,
(some new fantastic technology, an algorithm or even a program capable
of lossless compression of up to 1000:1 etc.)
please let us know immediately: artest@inbox.ru Thank you!
[[1]] COMPRESSION QUALITY
=========================
(see also
[[2]] Speed
[[3]] Details
[[4]] Comments)
Last seventh line shows results for the sum of all 1231 texts in six sets.
Origin DURILCA Entropy Slim RKC EPM Compressia PAQ6 PPMonstr PPMN
555.90% 100.04 100.47 100.13 100% 101.76 100.92 101.06 102.24 106.01
567.66% 100% 104.53 108.58 104.13 110.79 108.38 110.82 113.42 115.01
455.43% 100% 104.56 108.09 106.31 109.96 108.47 110.10 112.89 111.94
513.18% 100% 104.14 109.85 107.26 112.58 110.55 113.25 115.19 114.70
799.24% 100.41 101.80 108.25 124.62 102.27 106.97 102.55 100% 115.93
432.59% 100% 122.52 122.68 127.29 123.95 130.67 124.45 125.58 123.82
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
506.08% 100% 108.04 111.99 112.20 113.73 114.12 114.19 115.87 116.11
ASH BEE PPMd RAR UHArc DC SBC BZip2 7-zip pkzip
101.08 107.67 108.14 109.58 105.80 109.17 109.49 124.64 152.86 159.97
113.49 117.44 118.78 120.19 117.81 119.46 120.79 136.85 178.63 186.09
112.30 116.43 117.05 117.39 115.89 116.26 117.70 130.16 163.57 170.65
114.86 119.36 120.27 120.44 120.33 120.57 121.66 137.33 174.83 181.91
110.63 111.40 109.06 110.11 117.68 121.44 118.06 149.67 197.06 205.34
126.87 128.81 129.39 130.16 138.55 136.80 135.14 143.16 175.08 181.85
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~..~~~~~~~~~~~~~~~~~~~~~~
116.43 120.13 120.75 121.21 123.03 123.18 123.45 137.84 173.93 181.03
Results of many other programs are in full version only, TEXTS.DAT file.
[[2]] Speed
===========
Canterbury Corpus Large Set http://corpus.canterbury.ac.nz/resources/large.zip
was used for this test, and a 970MHz PC with 256Mb RAM and Windows98.
Programs, Compression/ Overall Average Users' Compressed
options Extraction, Score Score Size
seconds seconds, % seconds, % bytes , %
no compression 0 0 4446 559 4446 577 16005619 600
7za a -t7z 88 1 1104 138 1024 133 3650998 138
7za a -t7z -mx 154 1 982 123 843 109 2975903 113
7za a -tzip 23 0 1244 156 1223 158 4393623 167
7za a -tzip -mx 44 0 1268 159 1227 159 4401160 167
ash04a /o6 /m230 110 114 1100 138 1001 130 3154310 119
ash04a /o9 /m230 164 170 1130 142 982 127 2863384 108
ash04a /o16 /m230 221 210 1524 191 1324 171 3932449 149
ash04a /o6 /m230 /s16 117 121 1113 140 1007 130 3146157 119
ash04a /o9 /m230 /s16 172 177 1142 143 987 128 2855460 108
ash04a /o16 /m230 /s16 235 224 1541 193 1329 172 3895407 148
bee a -m1 70 71 1049 132 986 128 3268345 124
bee a -m2 144 148 1167 146 1037 134 3147914 119
bee a -m3 201 204 1267 159 1085 140 3099133 117
durilca e -o8 -t2(31) 77 77 964 121 895 116 2915855 110
durilca e -o9 -t2(31) 80 80 969 121 897 116 2911029 110
durilca e -o10 -t2(31) 82 83 970 122 896 116 2899602 110
durilca e -o12 -t2(31) 83 85 970 122 895 116 2886375 109
durilca e -o16 -t2(31) 85 86 972 122 895 116 2880306 109
durilca e -o32 -t2(31) 87 88 975 122 896 116 2878335 109
durilca e -o64 -t2(31) 88 89 977 122 898 116 2878490 109
durilca e -o128 -t2(31) 88 90 978 123 898 116 2878614 109
epm9 c008 117 116 1034 130 929 120 2882720 109
epm9 c012 142 142 1119 140 991 128 3007492 114
epm9 c016 147 147 1138 143 1006 130 3040288 115
grzipii e 16 10 908 114 893 115 3170895 120
paq6v2a 388 387 1597 200 1247 161 2957673 112
paq6v2a -6 2551 2710 6002 754 3706 481 2664718 101
rar a -m1 13 1 1239 155 1227 159 4408290 167
rar a -m2 21 1 1169 147 1150 149 4131324 157
rar a -m3 36 1 1156 145 1123 145 4026937 153
rar a -m4 19 11 914 114 896 116 3178761 120
rar a -m5 25 16 920 115 898 116 3164814 120
rar a -m5 -s 25 16 923 116 900 116 3173148 120
rar a -mc16t -s 35 26 992 124 960 124 3347746 127
rar a -mc16t+ -s 35 26 992 124 960 124 3347746 127
rar a -mc16:128t -s 40 31 955 120 919 119 3180234 120
rar a -mc16:128t+ -s 40 31 955 120 919 119 3180234 120
rar32 a -mc16t -s 37 28 996 125 962 124 3347746 127
rkc -mf -M230M -o8 94 6 1083 136 998 129 3537812 134
rkc -mx -M230M -o8 120 130 1019 128 911 118 2766843 105
rkc -mxx -M230M -o8 276 276 1320 166 1071 139 2763319 105
rkc -mxx -M230M -o12 297 297 1334 167 1067 138 2663521 101
rkc -mxx -M230M -o16 318 312 1361 171 1074 139 2630041 100
rkc -mxx -M230M -ft 316 316 1367 172 1082 140 2645990 100
rkc -mf -M230M -td+ 94 6 1083 136 998 129 3537812 134
rkc -mx -M230M -td+ 144 156 1032 129 902 117 2633402 100
rkc -mxx -M230M -td+ 305 311 1346 169 1072 139 2630041 100
slim a -d32 -w21 581 650 2034 255 1511 196 2890141 109
slim a -d16 -w21 577 648 2029 255 1509 195 2890280 109
slim a -d8 -w21 566 626 1995 250 1486 192 2891166 109
slim a -d4 -w21 521 583 1908 239 1439 186 2892897 109
uhbc e 40 31 951 119 914 118 3164344 120
//previous
ace32 a -d4096 66 2 1124 141 1058 137 3801917 142
ace32 a -d4096 -m1 31 2 1134 143 1104 143 3965841 149
ace32 a -d4096 -m5 206 2 1249 157 1045 136 3746553 140
arh a 38 40 1091 137 1053 137 3647067 137
arh a -2 -1 68 40 1121 141 1054 137 3647067 137
ba -k -50 35 12 964 121 929 121 3298943 124
bix a -mdg -s 92 1 1069 134 978 127 3514944 132
boa -m1 86 88 1253 158 1168 152 3886863 146
boa -m15 139 141 1165 146 1027 133 3182739 119
boa -m15 -s 138 140 1148 144 1011 131 3132810 117
bzip2 -k 21 6 1032 130 1011 131 3616113 136
bzip2 -k -9 20 6 1031 130 1011 131 3616113 136
Entropy t o12 94 95 1003 126 910 118 2932445 110
Entropy t o16 98 99 1001 126 904 117 2892711 108
Entropy t o32 105 106 1009 127 905 118 2873677 108
Entropy t o64 112 111 1022 128 911 118 2873318 108
compcl c -b15 37 20 904 114 868 113 3049569 114
compcl c -b15 -s 38 29 808 102 770 100 2668128 100
dc e 13 7 903 114 890 116 3179173 119
dc e -b16300 -mt5 17 7 795 100 778 101 2773427 104
eri a 39 17 936 118 897 116 3168414 119
eri a -m3 59 21 996 125 937 122 3295385 124
eri a -m6 59 21 989 124 931 121 3272926 123
gcac a 26 12 980 123 954 124 3390603 127
gcac s 26 12 981 123 955 124 3395064 127
imp98 a -mm 31 1 1175 148 1143 148 4112387 154
imp98 a -mm -2 13 5 999 126 986 128 3533761 132
imp98 a -2 -s4 13 5 999 126 986 128 3533693 132
pkzip -es 1 1 1654 208 1652 215 5945622 223
pkzip -a 4 1 1308 164 1304 169 4691491 176
pkzip -exx 16 1 1296 163 1280 166 4605942 173
ppmdi e -o7 -m232 11 12 904 114 893 116 3169000 119
ppmdi e -o12 -m232 25 26 915 115 891 116 3113630 117
ppmdi e -o16 -m232 27 28 916 115 890 116 3100943 116
ppmn_km e -o6 -MT1 30 30 931 117 901 117 3132278 117
ppmn_km e -o8 -MT1 64 65 993 125 929 121 3107654 116
ppmn_km e -o9 62 63 990 125 929 121 3115560 117
ppmn_km e -o9 -M:50 49 50 949 119 900 117 3058436 115
ppmonstr e -o7 -m232 64 67 974 123 911 118 3035498 114
ppmonstr e -o8 -m232 71 74 980 123 910 118 3007964 113
ppmonstr e -o64 -m232 101 103 1020 128 920 119 2937387 110
qlfc a 22 11 973 122 952 124 3385084 127
rk -mf2 50 20 1108 139 1058 137 3735704 140
rk -mx1 144 143 1147 144 1004 130 3093640 116
rk -mx2 173 173 1203 151 1032 134 3086312 116
sbc c -b63 29 9 914 115 885 115 3151930 118
sbc c -os -b63 29 9 810 102 782 101 2779632 104
szip -o4 4 10 1027 129 1023 133 3647445 137
szip -o6 17 14 996 125 979 127 3475264 130
szip -o8 -b41 27 17 973 122 947 123 3348344 125
zzip a 21 11 977 123 956 124 3400243 127
zzip a -mx 22 12 973 122 952 124 3383060 127
zzip a -mx -30m 30 12 940 118 910 118 3233147 121
abc13 -c 20 9 950 119 931 120 3313820 124
abc24 -c 29 16 923 116 897 116 3159570 118
uharc a -m1 -md32768 63 5 1026 129 969 125 3446069 129
uharc a -m2 -md32768 100 5 980 123 890 115 3151572 118
uharc a -m3 -md32768 110 5 973 122 874 113 3087249 115
uharc a -mz -md32768 8 9 1084 136 1077 139 3842041 144
uharc a -mx -md32768 60 55 936 117 882 114 2953184 110
ybs -m1m 22 8 952 119 931 120 3316356 124
ybs -m2m 25 8 937 117 915 118 3255538 122
ybs -m4m 28 8 919 115 894 116 3178183 119
ybs -m8m 31 8 905 113 877 113 3116271 116
ybs -m16mu 33 9 835 105 805 104 2852642 106
ybs -m16mu -r 34 9 841 105 811 105 2874130 107
ybs_d -m16mu 34 9 836 105 805 104 2852642 106
Overall score is calculated by adding compression time, extraction time, and
time it would take to transfer the compressed file over a 28,800bps network:
(compressed_size)/3600
Average Users' score is calculated by adding (compress_time/10)+ extract_time +
time it would take to transfer the compressed file over a 28,800bps network.
Compression time is divided by 10 here, because more than 90% of people would
never compress anything during their life (with compression programs), but they
use compressed data almost _every_ time they use computers and/or Internet.
That's why compression time is not so actual for them.
[[3]] Details
=============
are no longer put to this main text
(thousands of lines reporting about 60,000 results on 1231 files in 6 sets),
but can be found in FULL version with TEXTS.DAT and *.BAT
at http://compression.ru/artest/artest26.zip
or http://artest1.tripod.com/artest26.zip
[[4]] Comments
==============
Links to download programs
and Homepages
are now in links.htm file
What's new:
~~~~~~~~~~~
12 new programs were tested:
ASH 04a
7-zip 3.13
RAR 3.30b5
UHBC 1.0
EPM 9
Slim 0.021a
BEE 0.7.7
Durilca 0.3a
PAQ 6v2
RKC 1.02
GRZipII 0.2.3
BWIC
Latest beta versions of DC, Entropy, UHArc were available
from authors by e-mail request:
Entropy: artest@inbox.ru
DC: EdgarBinder@t-online.de
UHArc: Uwe.Herklotz@gmx.de
Results of many other programs are in full version only, TEXTS.DAT file.
The set of Russian texts was at http://arte.nm.ru, but now only artest@inbox.ru
WARNINGS:
~~~~~~~~~
Beta versions of RKC, EPM and BWIC fail to compress and/or decompress many
files. Authors are notified.
ASH 04a can fail to decompress some large files if it lacks memory.
BA 1.00beta5 can't correctly decompress shaks12.txt and set used for speed
measurements.
DC 0.99.158b failed to decompress 1DFRE10.dc , ANDES10.dc , and BTI0110.dc,
saying "Corrupted block" (while t(est) command writes "Test successful").
Problems in all other compressors were not found.
ESP, Rkive and many other programs are not tested any more,
their results and links can be found in previous volumes of ARTest.
The LATEST RELEASE, and all previous volumes can be found
at http://compression.ru/artest/
Send your suggestions, comments to artest@inbox.ru
With best kind regards,
A.Ratushnyak
Back to main ARTest page
|
|
Последнее обновление:
22-January-2004
|
Сайт о сжатии >>
Новинки |
О сервере |
Статистика
Книга "Методы сжатия данных" >>
Универсальные |
Изображений |
Видео
Разделы >>
Download (статьи+исходники) |
Ссылки |
Ru.compress |
Arctest |
Видео |
Каталог ссылок |
Форум
Проекты >>
Д.Ватолина
|
А.Ратушняка |
М.Смирнова |
В.Юкина |
Е.Шелвина |
А.Филинского |
Д.Шкарина |
С.Оснача
|
 |
 |
 |
 |
 |
 |
|