296:
2783:) is the differential entropy of a Gaussian random variable with variance D. This lower bound is extensible to sources with memory and other distortion measures. One important feature of the SLB is that it is asymptotically tight in the low distortion regime for a wide class of sources and in some occasions, it actually coincides with the rate–distortion function. Shannon Lower Bounds can generally be found if the distortion between any two numbers can be expressed as a function of the difference between the value of these two numbers.
5340:
5330:
3758:
31:
172:
2539:
3455:
3444:
510:
Rate–distortion theory gives an analytical expression for how much compression can be achieved using lossy compression methods. Many of the existing audio, speech, image, and video compression techniques have transforms, quantization, and bit-rate allocation procedures that capitalize on the general
3460:
Rate–distortion theory tell us that 'no compression system exists that performs outside the gray area'. The closer a practical compression system is to the red (lower) bound, the better it performs. As a general rule, this bound can only be attained by increasing the coding block length parameter.
1648:
3688:
3468:
This rate–distortion function holds only for
Gaussian memoryless sources. It is known that the Gaussian source is the most "difficult" source to encode: for a given mean square error, it requires the greatest number of bits. The performance of a practical compression system working
2283:
783:
3272:
2686:
is often difficult to obtain except in some instances for which we next offer two of the best known examples. The rate–distortion function of any source is known to obey several fundamental properties, the most important ones being that it is a
3000:
995:
3161:
1786:
533:
is a subject of on-going discussion. In the most simple case (which is actually used in most cases), the distortion is defined as the expected value of the square of the difference between input and output signal (i.e., the
1451:
1459:
2797:
When working with stationary sources with memory, it is necessary to modify the definition of the rate distortion function and it must be understood in the sense of a limit taken over sequences of increasing lengths.
2794:, is an elegant iterative technique for numerically obtaining rate–distortion functions of arbitrary finite input/output alphabet sources and much work has been done to extend it to more general problem instances.
3518:
873:
2534:{\displaystyle D_{Q}=\int _{-\infty }^{\infty }\int _{-\infty }^{\infty }P_{X,Y}(x,y)(x-y)^{2}\,dx\,dy=\int _{-\infty }^{\infty }\int _{-\infty }^{\infty }Q_{Y\mid X}(y\mid x)P_{X}(x)(x-y)^{2}\,dx\,dy.}
2866:
666:
283:, that should be communicated over a channel, so that the source (input signal) can be approximately reconstructed at the receiver (output signal) without exceeding an expected distortion
582:, but are often not easy to include in rate–distortion theory. In image and video compression, the human perception models are less well developed and inclusion is mostly limited to the
1256:
2770:
2646:
2263:
1046:
3439:{\displaystyle R(D)={\begin{cases}{\frac {1}{2}}\log _{2}(\sigma _{x}^{2}/D),&{\text{if }}0\leq D\leq \sigma _{x}^{2}\\0,&{\text{if }}D>\sigma _{x}^{2}.\end{cases}}}
1874:
2118:
3498:
1931:
504:
441:
574:. In audio compression, perceptual models (and therefore perceptual distortion measures) are relatively well developed and routinely used in compression techniques such as
2053:
3226:
1835:
1323:
1135:
653:
2877:
2598:
1972:
3902:
3837:
3751:
3718:
2673:
2175:
2148:
468:
405:
378:
351:
324:
3931:
3799:
1288:
4124:
PyRated is a very simple Python package to do the most basic calculation in rate-distortion theory: the determination of the "codebook" and the transmission rate
3863:
889:
3246:
3192:
2562:
2215:
2195:
2012:
1992:
1181:
1161:
1093:
1073:
624:
3865:
bits/symbol will be lost when transmitting this information over the given channel. For the user to have any hope of reconstructing with a maximum distortion
2706:
3011:
1663:
2702:
and thus the shape for the function in the examples is typical (even measured rate–distortion functions in real life tend to have very similar forms).
1339:
1643:{\displaystyle H(Y\mid X)=-\int _{-\infty }^{\infty }\int _{-\infty }^{\infty }Q_{Y\mid X}(y\mid x)P_{X}(x)\log _{2}(Q_{Y\mid X}(y\mid x))\,dx\,dy.}
3801:
bits/symbol of information from the source must reach the user. We also know from
Shannon's channel coding theorem that if the source entropy is
4164:
236:
3683:{\displaystyle R(D)=\left\{{\begin{matrix}H_{b}(p)-H_{b}(D),&0\leq D\leq \min {(p,1-p)}\\0,&D>\min {(p,1-p)}\end{matrix}}\right.}
208:
153:
4832:
4643:
189:
2709:(SLB), which in the case of squared error and memoryless sources, states that for arbitrary sources with finite differential entropy,
215:
4532:
5038:
4861:
4655:
2705:
Although analytical solutions to this problem are scarce, there are upper and lower bounds to these functions including the famous
796:
4346:
222:
5043:
4620:
2804:
4082:
4053:
778:{\displaystyle d(x,{\hat {x}})={\begin{cases}0&{\text{if }}x={\hat {x}}\\1&{\text{if }}x\neq {\hat {x}}\end{cases}}}
4773:
204:
1794:
The mutual information can be understood as a measure for 'prior' uncertainty the receiver has about the sender's signal (
3869:, we must impose the requirement that the information lost in transmission does not exceed the maximum tolerable loss of
85:
2675:. These definitions can be formulated measure-theoretically to account for discrete and mixed random variables as well.
5150:
4888:
4827:
4638:
4588:
4411:
112:
4271:
4256:
4157:
2544:
As the above equations show, calculating a rate–distortion function requires the stochastic description of the input
255:
96:
5263:
3960:
3462:
883:
The functions that relate the rate and distortion are found as the solution of the following minimization problem:
595:
591:
146:
4137:
5273:
5111:
4962:
4881:
4675:
5246:
4866:
4660:
4448:
193:
70:
1189:
5343:
4379:
5008:
229:
5333:
5236:
4778:
4336:
4150:
3942:
3166:
where superscripts denote a complete sequence up to that time and the subscript 0 indicates initial state.
2787:
1052:
122:
40:
2715:
2603:
2220:
1003:
5374:
4326:
4321:
279:; it addresses the problem of determining the minimal number of bits per symbol, as measured by the rate
139:
127:
5195:
5033:
5013:
4957:
4615:
4406:
4209:
3249:
1840:
2058:
1802:)), diminished by the uncertainty that is left after receiving information about the sender's signal (
5369:
5278:
5219:
5145:
4993:
4583:
4578:
4433:
4276:
3509:
3472:
1886:
3296:
1837:). Of course the decrease in uncertainty is due to the communicated amount of information, which is
705:
473:
410:
5283:
4856:
4650:
4351:
3768:
Suppose we want to transmit information about a source to the user with a distortion not exceeding
2692:
2995:{\displaystyle R_{n}(D)={\frac {1}{n}}\inf _{Q_{Y^{n}\mid X^{n}}\in {\mathcal {Q}}}I(Y^{n},X^{n})}
2017:
16:
Branch of information theory which provides the theoretical foundations for lossy data compression
5224:
4595:
4482:
4438:
4251:
4234:
4224:
3721:
3204:
1049:
182:
1805:
1293:
1098:
629:
4849:
4600:
4384:
4229:
2699:
276:
5121:
4041:
2567:
1936:
5253:
3263:
2679:
3872:
3816:
655:. Typical distortion functions are the Hamming distortion and the Squared-error distortion.
546:, watching pictures and video) the distortion measure should preferably be modeled on human
4937:
4399:
4361:
4182:
3730:
3696:
3465:
that operate at distances from the rate–distortion function that are practically relevant.
2683:
2651:
2153:
2126:
1654:
990:{\displaystyle \inf _{Q_{Y\mid X}(y\mid x)}I_{Q}(Y;X){\text{ subject to }}D_{Q}\leq D^{*}.}
559:
446:
383:
356:
329:
302:
65:
45:
3907:
3775:
1264:
8:
5168:
5059:
5018:
5003:
4972:
4967:
4876:
4783:
4716:
4685:
4670:
4453:
3842:
3195:
3169:
2688:
1330:
50:
4071:
5241:
5211:
5190:
5096:
5028:
4922:
4610:
4426:
4416:
4311:
4291:
4286:
4008:
3997:
3231:
3177:
2547:
2266:
2200:
2180:
1997:
1977:
1166:
1146:
1139:
1078:
1058:
609:
535:
295:
272:
60:
22:
4822:
5185:
5173:
5155:
5023:
4907:
4844:
4690:
4605:
4561:
4522:
4204:
4096:
4088:
4078:
4049:
3757:
3156:{\displaystyle {\mathcal {Q}}=\{Q_{Y^{n}\mid X^{n}}(Y^{n}\mid X^{n},X_{0}):E\leq D\}}
2274:
567:
539:
3503:
1781:{\displaystyle \inf _{Q_{Y\mid X}(y\mid x)}E]{\text{ subject to }}I_{Q}(Y;X)\leq R.}
1653:
The problem can also be formulated as a distortion–rate function, where we find the
5160:
5116:
5089:
5084:
4942:
4927:
4837:
4746:
4741:
4570:
4303:
4281:
4173:
3948:
3806:
1657:
over achievable distortions for given rate constraint. The relevant expression is:
542:
techniques operate on data that will be perceived by human consumers (listening to
117:
75:
5079:
4893:
4817:
4798:
4768:
4736:
4702:
4261:
4199:
2695:
1974:. Alternatively, if the communication channel is perfect and the received signal
571:
4871:
4665:
4394:
4389:
4246:
4219:
4191:
3969:
3904:
bits/symbol. This means that the channel capacity must be at least as large as
3254:
2791:
1446:{\displaystyle H(Y)=-\int _{-\infty }^{\infty }P_{Y}(y)\log _{2}(P_{Y}(y))\,dy}
515:
4119:
5363:
5178:
5126:
4793:
4788:
4763:
4695:
4316:
4214:
3954:
563:
55:
5299:
4266:
4241:
4142:
3454:
80:
4100:
5258:
5136:
4932:
4808:
4758:
4128:, given a utility function (distortion matrix) and a Lagrange multiplier
3975:
3461:
Nevertheless, even at unit blocklengths one can often find good (scalar)
1791:
The two formulations lead to functions which are inverses of each other.
555:
3763:
5315:
5106:
5101:
4988:
4947:
4753:
3998:"Rethinking Lossy Compression: The Rate-Distortion-Perception Tradeoff"
551:
547:
30:
3170:
Memoryless (independent) Gaussian source with squared-error distortion
2270:
2265:
and the prescribed maximum distortion, respectively. When we use the
171:
5229:
5074:
4731:
4013:
3199:
3957: – Process of reducing correlation within one or more signals
443:. We try to minimize the distortion between the original sequence
4998:
4472:
4421:
4092:
4073:
Rate
Distortion Theory: A Mathematical Basis for Data Compression
3504:
Memoryless (independent) Bernoulli source with
Hamming distortion
4512:
4005:
606:
Distortion functions measure the cost of representing a symbol
579:
1055:(PDF) of the communication channel output (compressed signal)
5347:
4952:
4545:
4492:
868:{\displaystyle d(x,{\hat {x}})=\left(x-{\hat {x}}\right)^{2}}
543:
4502:
4356:
4341:
4331:
3677:
3432:
1333:
of the output signal given the input signal, respectively:
771:
587:
583:
529:
per data sample to be stored or transmitted. The notion of
3450:
The following figure shows what this function looks like:
4477:
4443:
3228:, and if we assume that successive samples of the signal
2861:{\displaystyle R(D)=\lim _{n\rightarrow \infty }R_{n}(D)}
575:
526:
562:, distortion measures can ultimately be identified with
4120:"PyRated: a python package for rate distortion theory"
3542:
3910:
3875:
3845:
3819:
3778:
3764:
Connecting rate-distortion theory to channel capacity
3733:
3699:
3521:
3475:
3275:
3234:
3207:
3180:
3014:
2880:
2807:
2718:
2654:
2606:
2570:
2550:
2286:
2223:
2203:
2183:
2156:
2129:
2061:
2020:
2000:
1980:
1939:
1889:
1843:
1808:
1666:
1462:
1342:
1296:
1267:
1192:
1169:
1149:
1101:
1081:
1061:
1006:
892:
799:
669:
632:
612:
476:
449:
413:
386:
359:
332:
305:
3965:
Pages displaying wikidata descriptions as a fallback
3963: – decision algorithm used in video compression
4062:
2123:In the definition of the rate–distortion function,
196:. Unsourced material may be challenged and removed.
4070:
3925:
3896:
3857:
3831:
3793:
3745:
3712:
3682:
3492:
3438:
3240:
3220:
3186:
3155:
2994:
2860:
2764:
2667:
2640:
2592:
2556:
2533:
2257:
2209:
2189:
2169:
2142:
2112:
2047:
2006:
1986:
1966:
1925:
1868:
1829:
1780:
1642:
1445:
1317:
1282:
1250:
1175:
1155:
1129:
1087:
1067:
1040:
989:
867:
777:
647:
618:
498:
462:
435:
399:
372:
345:
318:
3945: – Class of algorithms in information theory
3469:on—say—images, may well be below the
5361:
4138:VcDemo Image and Video Compression Learning Tool
3772:. Rate–distortion theory tells us that at least
3647:
3603:
2914:
2824:
1668:
894:
518:in his foundational work on information theory.
299:Rate distortion encoder and decoder. An encoder
2600:, and then aims at finding the conditional PDF
275:which provides the theoretical foundations for
4158:
147:
4172:
3995:
3150:
3025:
878:
4039:
4027:
3978: – Type of signal in signal processing
788:
4165:
4151:
4040:Cover, Thomas M.; Thomas, Joy A. (2012) .
2648:that minimize rate for a given distortion
1048:, sometimes called a test channel, is the
154:
140:
4012:
3727:Plot of the rate-distortion function for
2761:
2521:
2514:
2396:
2389:
1630:
1623:
1436:
1247:
256:Learn how and when to remove this message
3951: – Compact encoding of digital data
1251:{\displaystyle I(Y;X)=H(Y)-H(Y\mid X)\,}
294:
4117:
601:
525:is usually understood as the number of
5362:
4068:
514:Rate–distortion theory was created by
4146:
3512:with Hamming distortion is given by:
1325:are the entropy of the output signal
658:
2765:{\displaystyle R(D)\geq h(X)-h(D)\,}
2641:{\displaystyle Q_{Y\mid X}(y\mid x)}
2269:as distortion measure, we have (for
2258:{\displaystyle Q_{Y\mid X}(y\mid x)}
1075:for a given input (original signal)
1041:{\displaystyle Q_{Y\mid X}(y\mid x)}
538:). However, since we know that most
511:shape of rate–distortion functions.
194:adding citations to reliable sources
165:
86:Limiting density of discrete points
13:
3508:The rate-distortion function of a
3266:for the rate–distortion function:
3017:
2953:
2834:
2437:
2432:
2419:
2414:
2331:
2326:
2313:
2308:
1518:
1513:
1500:
1495:
1374:
1369:
14:
5386:
4111:
1869:{\displaystyle I\left(Y;X\right)}
97:Asymptotic equipartition property
5339:
5338:
5329:
5328:
3756:
3453:
3252:(or equivalently, the source is
2113:{\displaystyle I(Y;X)=H(X)=H(Y)}
1879:As an example, in case there is
170:
29:
3996:Blau, Y.; Michaeli, T. (2019).
3493:{\displaystyle R\left(D\right)}
1926:{\displaystyle H(Y\mid X)=H(Y)}
521:In rate–distortion theory, the
470:and the reconstructed sequence
290:
181:needs additional citations for
113:Shannon's source coding theorem
4046:Elements of Information Theory
4033:
4021:
3989:
3920:
3914:
3891:
3885:
3788:
3782:
3669:
3651:
3625:
3607:
3583:
3577:
3561:
3555:
3531:
3525:
3348:
3322:
3285:
3279:
3141:
3138:
3112:
3106:
3097:
3058:
2989:
2963:
2897:
2891:
2855:
2849:
2831:
2817:
2811:
2758:
2752:
2743:
2737:
2728:
2722:
2635:
2623:
2587:
2581:
2505:
2492:
2489:
2483:
2470:
2458:
2380:
2367:
2364:
2352:
2252:
2240:
2107:
2101:
2092:
2086:
2077:
2065:
2036:
2024:
1955:
1943:
1920:
1914:
1905:
1893:
1824:
1812:
1766:
1754:
1736:
1733:
1721:
1708:
1700:
1688:
1620:
1617:
1605:
1586:
1570:
1564:
1551:
1539:
1478:
1466:
1433:
1430:
1424:
1411:
1395:
1389:
1352:
1346:
1312:
1300:
1277:
1271:
1244:
1232:
1223:
1217:
1208:
1196:
1124:
1112:
1035:
1023:
953:
941:
926:
914:
848:
824:
818:
803:
762:
730:
694:
688:
673:
639:
499:{\displaystyle {\hat {X}}^{n}}
484:
436:{\displaystyle {\hat {X}}^{n}}
421:
71:Conditional mutual information
1:
4118:Marzen, Sarah; DeDeo, Simon.
3982:
3972: – Geometrical structure
4042:"10. Rate Distortion Theory"
3961:Rate–distortion optimization
2048:{\displaystyle H(Y\mid X)=0}
1053:probability density function
123:Noisy-channel coding theorem
7:
3936:
3221:{\displaystyle \sigma ^{2}}
2177:are the distortion between
1994:is identical to the signal
1883:communication at all, then
10:
5391:
5220:Compressed data structures
4542:RLE + BWT + MTF + Huffman
4210:Asymmetric numeral systems
4007:. PMLR. pp. 675–685.
3250:stochastically independent
1830:{\displaystyle H(Y\mid X)}
1318:{\displaystyle H(Y\mid X)}
1130:{\displaystyle I_{Q}(Y;X)}
648:{\displaystyle {\hat {x}}}
626:by an approximated symbol
5324:
5308:
5292:
5210:
5135:
5067:
5058:
4981:
4915:
4906:
4807:
4724:
4715:
4631:
4579:Discrete cosine transform
4569:
4560:
4509:LZ77 + Huffman + context
4462:
4372:
4302:
4190:
4181:
3510:Bernoulli random variable
3262:), we find the following
879:Rate–distortion functions
407:which outputs a sequence
380:is then fed to a decoder
5284:Smallest grammar problem
3943:Blahut–Arimoto algorithm
2788:Blahut–Arimoto algorithm
2693:monotonically decreasing
2593:{\displaystyle P_{X}(x)}
1967:{\displaystyle I(Y;X)=0}
789:Squared-error distortion
554:: much like the use of
205:"Rate–distortion theory"
5225:Compressed suffix array
4774:Nyquist–Shannon theorem
4048:(2nd ed.). Wiley.
4028:Cover & Thomas 2012
3722:binary entropy function
353:. The encoded sequence
128:Shannon–Hartley theorem
3927:
3898:
3897:{\displaystyle H-R(D)}
3859:
3833:
3832:{\displaystyle C<H}
3795:
3747:
3714:
3684:
3494:
3440:
3242:
3222:
3188:
3157:
2996:
2862:
2766:
2669:
2642:
2594:
2558:
2535:
2259:
2211:
2191:
2171:
2144:
2114:
2049:
2008:
1988:
1968:
1927:
1870:
1831:
1782:
1741: subject to
1644:
1447:
1319:
1284:
1252:
1177:
1157:
1131:
1089:
1069:
1042:
991:
958: subject to
869:
779:
649:
620:
507:
500:
464:
437:
401:
374:
347:
320:
277:lossy data compression
269:Rate–distortion theory
102:Rate–distortion theory
5254:Kolmogorov complexity
5122:Video characteristics
4499:LZ77 + Huffman + ANS
4069:Berger, Toby (1971).
3928:
3899:
3860:
3834:
3805:bits/symbol, and the
3796:
3748:
3746:{\displaystyle p=0.5}
3715:
3713:{\displaystyle H_{b}}
3685:
3495:
3441:
3264:analytical expression
3243:
3223:
3198:random variable with
3189:
3158:
2997:
2863:
2767:
2670:
2668:{\displaystyle D^{*}}
2643:
2595:
2559:
2536:
2260:
2212:
2192:
2172:
2170:{\displaystyle D^{*}}
2145:
2143:{\displaystyle D_{Q}}
2115:
2050:
2009:
1989:
1969:
1928:
1871:
1832:
1783:
1645:
1448:
1320:
1285:
1253:
1178:
1158:
1132:
1090:
1070:
1043:
992:
870:
780:
650:
621:
501:
465:
463:{\displaystyle X^{n}}
438:
402:
400:{\displaystyle g_{n}}
375:
373:{\displaystyle Y^{n}}
348:
346:{\displaystyle X^{n}}
321:
319:{\displaystyle f_{n}}
298:
271:is a major branch of
5344:Compression software
4938:Compression artifact
4894:Psychoacoustic model
3926:{\displaystyle R(D)}
3908:
3873:
3843:
3817:
3794:{\displaystyle R(D)}
3776:
3731:
3697:
3519:
3473:
3273:
3232:
3205:
3178:
3012:
2878:
2805:
2716:
2684:minimization problem
2652:
2604:
2568:
2564:in terms of the PDF
2548:
2284:
2221:
2201:
2181:
2154:
2127:
2059:
2018:
2014:at the sender, then
1998:
1978:
1937:
1887:
1841:
1806:
1664:
1460:
1340:
1294:
1283:{\displaystyle H(Y)}
1265:
1190:
1167:
1147:
1099:
1079:
1059:
1004:
890:
797:
667:
630:
610:
602:Distortion functions
566:as used in Bayesian
560:lossless compression
474:
447:
411:
384:
357:
330:
303:
190:improve this article
66:Directed information
46:Differential entropy
5334:Compression formats
4973:Texture compression
4968:Standard test image
4784:Silence compression
3858:{\displaystyle H-C}
3500:lower bound shown.
3425:
3387:
3339:
3258:, or the signal is
2707:Shannon lower bound
2441:
2423:
2335:
2317:
1522:
1504:
1378:
1331:conditional entropy
326:encodes a sequence
51:Conditional entropy
5375:Information theory
5242:Information theory
5097:Display resolution
4923:Chroma subsampling
4312:Byte pair encoding
4257:Shannon–Fano–Elias
3923:
3894:
3855:
3829:
3791:
3743:
3710:
3680:
3675:
3490:
3446:
3436:
3431:
3411:
3373:
3325:
3238:
3218:
3184:
3174:If we assume that
3153:
2992:
2959:
2858:
2838:
2762:
2665:
2638:
2590:
2554:
2531:
2424:
2406:
2318:
2300:
2275:continuous signals
2267:mean squared error
2255:
2207:
2187:
2167:
2140:
2110:
2045:
2004:
1984:
1964:
1923:
1866:
1827:
1778:
1704:
1640:
1505:
1487:
1443:
1361:
1315:
1280:
1248:
1173:
1153:
1140:mutual information
1127:
1085:
1065:
1038:
987:
930:
865:
775:
770:
659:Hamming distortion
645:
616:
536:mean squared error
508:
496:
460:
433:
397:
370:
343:
316:
273:information theory
61:Mutual information
23:Information theory
5357:
5356:
5206:
5205:
5156:Deblocking filter
5054:
5053:
4902:
4901:
4711:
4710:
4556:
4555:
4084:978-0-13-753103-5
4077:. Prentice Hall.
4055:978-1-118-58577-1
3403:
3359:
3307:
3241:{\displaystyle X}
3187:{\displaystyle X}
2913:
2911:
2823:
2790:, co-invented by
2682:solution to this
2557:{\displaystyle X}
2210:{\displaystyle Y}
2190:{\displaystyle X}
2007:{\displaystyle X}
1987:{\displaystyle Y}
1742:
1667:
1176:{\displaystyle X}
1156:{\displaystyle Y}
1088:{\displaystyle X}
1068:{\displaystyle Y}
959:
893:
851:
821:
765:
748:
733:
716:
691:
642:
619:{\displaystyle x}
540:lossy compression
487:
424:
266:
265:
258:
240:
164:
163:
5382:
5370:Data compression
5342:
5341:
5332:
5331:
5161:Lapped transform
5065:
5064:
4943:Image resolution
4928:Coding tree unit
4913:
4912:
4722:
4721:
4567:
4566:
4188:
4187:
4174:Data compression
4167:
4160:
4153:
4144:
4143:
4134:
4105:
4104:
4076:
4066:
4060:
4059:
4037:
4031:
4025:
4019:
4018:
4016:
4002:
3993:
3966:
3949:Data compression
3932:
3930:
3929:
3924:
3903:
3901:
3900:
3895:
3864:
3862:
3861:
3856:
3838:
3836:
3835:
3830:
3807:channel capacity
3800:
3798:
3797:
3792:
3760:
3752:
3750:
3749:
3744:
3719:
3717:
3716:
3711:
3709:
3708:
3689:
3687:
3686:
3681:
3679:
3676:
3672:
3628:
3576:
3575:
3554:
3553:
3499:
3497:
3496:
3491:
3489:
3457:
3445:
3443:
3442:
3437:
3435:
3434:
3424:
3419:
3404:
3401:
3386:
3381:
3360:
3357:
3344:
3338:
3333:
3318:
3317:
3308:
3300:
3247:
3245:
3244:
3239:
3227:
3225:
3224:
3219:
3217:
3216:
3193:
3191:
3190:
3185:
3162:
3160:
3159:
3154:
3137:
3136:
3124:
3123:
3096:
3095:
3083:
3082:
3070:
3069:
3057:
3056:
3055:
3054:
3042:
3041:
3021:
3020:
3001:
2999:
2998:
2993:
2988:
2987:
2975:
2974:
2958:
2957:
2956:
2947:
2946:
2945:
2944:
2932:
2931:
2912:
2904:
2890:
2889:
2867:
2865:
2864:
2859:
2848:
2847:
2837:
2771:
2769:
2768:
2763:
2674:
2672:
2671:
2666:
2664:
2663:
2647:
2645:
2644:
2639:
2622:
2621:
2599:
2597:
2596:
2591:
2580:
2579:
2563:
2561:
2560:
2555:
2540:
2538:
2537:
2532:
2513:
2512:
2482:
2481:
2457:
2456:
2440:
2435:
2422:
2417:
2388:
2387:
2351:
2350:
2334:
2329:
2316:
2311:
2296:
2295:
2264:
2262:
2261:
2256:
2239:
2238:
2216:
2214:
2213:
2208:
2196:
2194:
2193:
2188:
2176:
2174:
2173:
2168:
2166:
2165:
2149:
2147:
2146:
2141:
2139:
2138:
2119:
2117:
2116:
2111:
2054:
2052:
2051:
2046:
2013:
2011:
2010:
2005:
1993:
1991:
1990:
1985:
1973:
1971:
1970:
1965:
1932:
1930:
1929:
1924:
1875:
1873:
1872:
1867:
1865:
1861:
1836:
1834:
1833:
1828:
1787:
1785:
1784:
1779:
1753:
1752:
1743:
1740:
1720:
1719:
1703:
1687:
1686:
1649:
1647:
1646:
1641:
1604:
1603:
1582:
1581:
1563:
1562:
1538:
1537:
1521:
1516:
1503:
1498:
1452:
1450:
1449:
1444:
1423:
1422:
1407:
1406:
1388:
1387:
1377:
1372:
1324:
1322:
1321:
1316:
1289:
1287:
1286:
1281:
1257:
1255:
1254:
1249:
1182:
1180:
1179:
1174:
1162:
1160:
1159:
1154:
1136:
1134:
1133:
1128:
1111:
1110:
1094:
1092:
1091:
1086:
1074:
1072:
1071:
1066:
1047:
1045:
1044:
1039:
1022:
1021:
996:
994:
993:
988:
983:
982:
970:
969:
960:
957:
940:
939:
929:
913:
912:
874:
872:
871:
866:
864:
863:
858:
854:
853:
852:
844:
823:
822:
814:
784:
782:
781:
776:
774:
773:
767:
766:
758:
749:
746:
735:
734:
726:
717:
714:
693:
692:
684:
654:
652:
651:
646:
644:
643:
635:
625:
623:
622:
617:
505:
503:
502:
497:
495:
494:
489:
488:
480:
469:
467:
466:
461:
459:
458:
442:
440:
439:
434:
432:
431:
426:
425:
417:
406:
404:
403:
398:
396:
395:
379:
377:
376:
371:
369:
368:
352:
350:
349:
344:
342:
341:
325:
323:
322:
317:
315:
314:
261:
254:
250:
247:
241:
239:
198:
174:
166:
156:
149:
142:
118:Channel capacity
76:Relative entropy
33:
19:
18:
5390:
5389:
5385:
5384:
5383:
5381:
5380:
5379:
5360:
5359:
5358:
5353:
5320:
5304:
5288:
5269:Rate–distortion
5202:
5131:
5050:
4977:
4898:
4803:
4799:Sub-band coding
4707:
4632:Predictive type
4627:
4552:
4519:LZSS + Huffman
4469:LZ77 + Huffman
4458:
4368:
4304:Dictionary type
4298:
4200:Adaptive coding
4177:
4171:
4114:
4109:
4108:
4085:
4067:
4063:
4056:
4038:
4034:
4026:
4022:
4000:
3994:
3990:
3985:
3964:
3939:
3909:
3906:
3905:
3874:
3871:
3870:
3844:
3841:
3840:
3818:
3815:
3814:
3777:
3774:
3773:
3766:
3732:
3729:
3728:
3704:
3700:
3698:
3695:
3694:
3674:
3673:
3650:
3639:
3630:
3629:
3606:
3589:
3571:
3567:
3549:
3545:
3541:
3537:
3520:
3517:
3516:
3506:
3479:
3474:
3471:
3470:
3430:
3429:
3420:
3415:
3400:
3398:
3389:
3388:
3382:
3377:
3356:
3354:
3340:
3334:
3329:
3313:
3309:
3299:
3292:
3291:
3274:
3271:
3270:
3233:
3230:
3229:
3212:
3208:
3206:
3203:
3202:
3179:
3176:
3175:
3172:
3132:
3128:
3119:
3115:
3091:
3087:
3078:
3074:
3065:
3061:
3050:
3046:
3037:
3033:
3032:
3028:
3016:
3015:
3013:
3010:
3009:
2983:
2979:
2970:
2966:
2952:
2951:
2940:
2936:
2927:
2923:
2922:
2918:
2917:
2903:
2885:
2881:
2879:
2876:
2875:
2843:
2839:
2827:
2806:
2803:
2802:
2717:
2714:
2713:
2659:
2655:
2653:
2650:
2649:
2611:
2607:
2605:
2602:
2601:
2575:
2571:
2569:
2566:
2565:
2549:
2546:
2545:
2508:
2504:
2477:
2473:
2446:
2442:
2436:
2428:
2418:
2410:
2383:
2379:
2340:
2336:
2330:
2322:
2312:
2304:
2291:
2287:
2285:
2282:
2281:
2228:
2224:
2222:
2219:
2218:
2202:
2199:
2198:
2182:
2179:
2178:
2161:
2157:
2155:
2152:
2151:
2134:
2130:
2128:
2125:
2124:
2060:
2057:
2056:
2019:
2016:
2015:
1999:
1996:
1995:
1979:
1976:
1975:
1938:
1935:
1934:
1888:
1885:
1884:
1851:
1847:
1842:
1839:
1838:
1807:
1804:
1803:
1748:
1744:
1739:
1715:
1711:
1676:
1672:
1671:
1665:
1662:
1661:
1593:
1589:
1577:
1573:
1558:
1554:
1527:
1523:
1517:
1509:
1499:
1491:
1461:
1458:
1457:
1418:
1414:
1402:
1398:
1383:
1379:
1373:
1365:
1341:
1338:
1337:
1295:
1292:
1291:
1266:
1263:
1262:
1191:
1188:
1187:
1168:
1165:
1164:
1148:
1145:
1144:
1106:
1102:
1100:
1097:
1096:
1080:
1077:
1076:
1060:
1057:
1056:
1011:
1007:
1005:
1002:
1001:
978:
974:
965:
961:
956:
935:
931:
902:
898:
897:
891:
888:
887:
881:
859:
843:
842:
835:
831:
830:
813:
812:
798:
795:
794:
791:
769:
768:
757:
756:
745:
743:
737:
736:
725:
724:
713:
711:
701:
700:
683:
682:
668:
665:
664:
661:
634:
633:
631:
628:
627:
611:
608:
607:
604:
572:decision theory
490:
479:
478:
477:
475:
472:
471:
454:
450:
448:
445:
444:
427:
416:
415:
414:
412:
409:
408:
391:
387:
385:
382:
381:
364:
360:
358:
355:
354:
337:
333:
331:
328:
327:
310:
306:
304:
301:
300:
293:
262:
251:
245:
242:
199:
197:
187:
175:
160:
17:
12:
11:
5:
5388:
5378:
5377:
5372:
5355:
5354:
5352:
5351:
5336:
5325:
5322:
5321:
5319:
5318:
5312:
5310:
5306:
5305:
5303:
5302:
5296:
5294:
5290:
5289:
5287:
5286:
5281:
5276:
5271:
5266:
5261:
5256:
5251:
5250:
5249:
5239:
5234:
5233:
5232:
5227:
5216:
5214:
5208:
5207:
5204:
5203:
5201:
5200:
5199:
5198:
5193:
5183:
5182:
5181:
5176:
5171:
5163:
5158:
5153:
5148:
5142:
5140:
5133:
5132:
5130:
5129:
5124:
5119:
5114:
5109:
5104:
5099:
5094:
5093:
5092:
5087:
5082:
5071:
5069:
5062:
5056:
5055:
5052:
5051:
5049:
5048:
5047:
5046:
5041:
5036:
5031:
5021:
5016:
5011:
5006:
5001:
4996:
4991:
4985:
4983:
4979:
4978:
4976:
4975:
4970:
4965:
4960:
4955:
4950:
4945:
4940:
4935:
4930:
4925:
4919:
4917:
4910:
4904:
4903:
4900:
4899:
4897:
4896:
4891:
4886:
4885:
4884:
4879:
4874:
4869:
4864:
4854:
4853:
4852:
4842:
4841:
4840:
4835:
4825:
4820:
4814:
4812:
4805:
4804:
4802:
4801:
4796:
4791:
4786:
4781:
4776:
4771:
4766:
4761:
4756:
4751:
4750:
4749:
4744:
4739:
4728:
4726:
4719:
4713:
4712:
4709:
4708:
4706:
4705:
4703:Psychoacoustic
4700:
4699:
4698:
4693:
4688:
4680:
4679:
4678:
4673:
4668:
4663:
4658:
4648:
4647:
4646:
4635:
4633:
4629:
4628:
4626:
4625:
4624:
4623:
4618:
4613:
4603:
4598:
4593:
4592:
4591:
4586:
4575:
4573:
4571:Transform type
4564:
4558:
4557:
4554:
4553:
4551:
4550:
4549:
4548:
4540:
4539:
4538:
4535:
4527:
4526:
4525:
4517:
4516:
4515:
4507:
4506:
4505:
4497:
4496:
4495:
4487:
4486:
4485:
4480:
4475:
4466:
4464:
4460:
4459:
4457:
4456:
4451:
4446:
4441:
4436:
4431:
4430:
4429:
4424:
4414:
4409:
4404:
4403:
4402:
4392:
4387:
4382:
4376:
4374:
4370:
4369:
4367:
4366:
4365:
4364:
4359:
4354:
4349:
4344:
4339:
4334:
4329:
4324:
4314:
4308:
4306:
4300:
4299:
4297:
4296:
4295:
4294:
4289:
4284:
4279:
4269:
4264:
4259:
4254:
4249:
4244:
4239:
4238:
4237:
4232:
4227:
4217:
4212:
4207:
4202:
4196:
4194:
4185:
4179:
4178:
4170:
4169:
4162:
4155:
4147:
4141:
4140:
4135:
4113:
4112:External links
4110:
4107:
4106:
4083:
4061:
4054:
4032:
4020:
3987:
3986:
3984:
3981:
3980:
3979:
3973:
3970:Sphere packing
3967:
3958:
3952:
3946:
3938:
3935:
3922:
3919:
3916:
3913:
3893:
3890:
3887:
3884:
3881:
3878:
3854:
3851:
3848:
3828:
3825:
3822:
3790:
3787:
3784:
3781:
3765:
3762:
3742:
3739:
3736:
3707:
3703:
3691:
3690:
3678:
3671:
3668:
3665:
3662:
3659:
3656:
3653:
3649:
3646:
3643:
3640:
3638:
3635:
3632:
3631:
3627:
3624:
3621:
3618:
3615:
3612:
3609:
3605:
3602:
3599:
3596:
3593:
3590:
3588:
3585:
3582:
3579:
3574:
3570:
3566:
3563:
3560:
3557:
3552:
3548:
3544:
3543:
3540:
3536:
3533:
3530:
3527:
3524:
3505:
3502:
3488:
3485:
3482:
3478:
3448:
3447:
3433:
3428:
3423:
3418:
3414:
3410:
3407:
3399:
3397:
3394:
3391:
3390:
3385:
3380:
3376:
3372:
3369:
3366:
3363:
3355:
3353:
3350:
3347:
3343:
3337:
3332:
3328:
3324:
3321:
3316:
3312:
3306:
3303:
3298:
3297:
3295:
3290:
3287:
3284:
3281:
3278:
3237:
3215:
3211:
3183:
3171:
3168:
3164:
3163:
3152:
3149:
3146:
3143:
3140:
3135:
3131:
3127:
3122:
3118:
3114:
3111:
3108:
3105:
3102:
3099:
3094:
3090:
3086:
3081:
3077:
3073:
3068:
3064:
3060:
3053:
3049:
3045:
3040:
3036:
3031:
3027:
3024:
3019:
3003:
3002:
2991:
2986:
2982:
2978:
2973:
2969:
2965:
2962:
2955:
2950:
2943:
2939:
2935:
2930:
2926:
2921:
2916:
2910:
2907:
2902:
2899:
2896:
2893:
2888:
2884:
2869:
2868:
2857:
2854:
2851:
2846:
2842:
2836:
2833:
2830:
2826:
2822:
2819:
2816:
2813:
2810:
2792:Richard Blahut
2773:
2772:
2760:
2757:
2754:
2751:
2748:
2745:
2742:
2739:
2736:
2733:
2730:
2727:
2724:
2721:
2662:
2658:
2637:
2634:
2631:
2628:
2625:
2620:
2617:
2614:
2610:
2589:
2586:
2583:
2578:
2574:
2553:
2542:
2541:
2530:
2527:
2524:
2520:
2517:
2511:
2507:
2503:
2500:
2497:
2494:
2491:
2488:
2485:
2480:
2476:
2472:
2469:
2466:
2463:
2460:
2455:
2452:
2449:
2445:
2439:
2434:
2431:
2427:
2421:
2416:
2413:
2409:
2405:
2402:
2399:
2395:
2392:
2386:
2382:
2378:
2375:
2372:
2369:
2366:
2363:
2360:
2357:
2354:
2349:
2346:
2343:
2339:
2333:
2328:
2325:
2321:
2315:
2310:
2307:
2303:
2299:
2294:
2290:
2254:
2251:
2248:
2245:
2242:
2237:
2234:
2231:
2227:
2206:
2186:
2164:
2160:
2137:
2133:
2109:
2106:
2103:
2100:
2097:
2094:
2091:
2088:
2085:
2082:
2079:
2076:
2073:
2070:
2067:
2064:
2044:
2041:
2038:
2035:
2032:
2029:
2026:
2023:
2003:
1983:
1963:
1960:
1957:
1954:
1951:
1948:
1945:
1942:
1922:
1919:
1916:
1913:
1910:
1907:
1904:
1901:
1898:
1895:
1892:
1864:
1860:
1857:
1854:
1850:
1846:
1826:
1823:
1820:
1817:
1814:
1811:
1789:
1788:
1777:
1774:
1771:
1768:
1765:
1762:
1759:
1756:
1751:
1747:
1738:
1735:
1732:
1729:
1726:
1723:
1718:
1714:
1710:
1707:
1702:
1699:
1696:
1693:
1690:
1685:
1682:
1679:
1675:
1670:
1651:
1650:
1639:
1636:
1633:
1629:
1626:
1622:
1619:
1616:
1613:
1610:
1607:
1602:
1599:
1596:
1592:
1588:
1585:
1580:
1576:
1572:
1569:
1566:
1561:
1557:
1553:
1550:
1547:
1544:
1541:
1536:
1533:
1530:
1526:
1520:
1515:
1512:
1508:
1502:
1497:
1494:
1490:
1486:
1483:
1480:
1477:
1474:
1471:
1468:
1465:
1454:
1453:
1442:
1439:
1435:
1432:
1429:
1426:
1421:
1417:
1413:
1410:
1405:
1401:
1397:
1394:
1391:
1386:
1382:
1376:
1371:
1368:
1364:
1360:
1357:
1354:
1351:
1348:
1345:
1314:
1311:
1308:
1305:
1302:
1299:
1279:
1276:
1273:
1270:
1259:
1258:
1246:
1243:
1240:
1237:
1234:
1231:
1228:
1225:
1222:
1219:
1216:
1213:
1210:
1207:
1204:
1201:
1198:
1195:
1172:
1152:
1126:
1123:
1120:
1117:
1114:
1109:
1105:
1084:
1064:
1037:
1034:
1031:
1028:
1025:
1020:
1017:
1014:
1010:
998:
997:
986:
981:
977:
973:
968:
964:
955:
952:
949:
946:
943:
938:
934:
928:
925:
922:
919:
916:
911:
908:
905:
901:
896:
880:
877:
876:
875:
862:
857:
850:
847:
841:
838:
834:
829:
826:
820:
817:
811:
808:
805:
802:
790:
787:
786:
785:
772:
764:
761:
755:
752:
744:
742:
739:
738:
732:
729:
723:
720:
712:
710:
707:
706:
704:
699:
696:
690:
687:
681:
678:
675:
672:
660:
657:
641:
638:
615:
603:
600:
564:loss functions
516:Claude Shannon
493:
486:
483:
457:
453:
430:
423:
420:
394:
390:
367:
363:
340:
336:
313:
309:
292:
289:
264:
263:
178:
176:
169:
162:
161:
159:
158:
151:
144:
136:
133:
132:
131:
130:
125:
120:
115:
107:
106:
105:
104:
99:
91:
90:
89:
88:
83:
78:
73:
68:
63:
58:
53:
48:
43:
35:
34:
26:
25:
15:
9:
6:
4:
3:
2:
5387:
5376:
5373:
5371:
5368:
5367:
5365:
5349:
5345:
5337:
5335:
5327:
5326:
5323:
5317:
5314:
5313:
5311:
5307:
5301:
5298:
5297:
5295:
5291:
5285:
5282:
5280:
5277:
5275:
5272:
5270:
5267:
5265:
5262:
5260:
5257:
5255:
5252:
5248:
5245:
5244:
5243:
5240:
5238:
5235:
5231:
5228:
5226:
5223:
5222:
5221:
5218:
5217:
5215:
5213:
5209:
5197:
5194:
5192:
5189:
5188:
5187:
5184:
5180:
5177:
5175:
5172:
5170:
5167:
5166:
5164:
5162:
5159:
5157:
5154:
5152:
5149:
5147:
5144:
5143:
5141:
5138:
5134:
5128:
5127:Video quality
5125:
5123:
5120:
5118:
5115:
5113:
5110:
5108:
5105:
5103:
5100:
5098:
5095:
5091:
5088:
5086:
5083:
5081:
5078:
5077:
5076:
5073:
5072:
5070:
5066:
5063:
5061:
5057:
5045:
5042:
5040:
5037:
5035:
5032:
5030:
5027:
5026:
5025:
5022:
5020:
5017:
5015:
5012:
5010:
5007:
5005:
5002:
5000:
4997:
4995:
4992:
4990:
4987:
4986:
4984:
4980:
4974:
4971:
4969:
4966:
4964:
4961:
4959:
4956:
4954:
4951:
4949:
4946:
4944:
4941:
4939:
4936:
4934:
4931:
4929:
4926:
4924:
4921:
4920:
4918:
4914:
4911:
4909:
4905:
4895:
4892:
4890:
4887:
4883:
4880:
4878:
4875:
4873:
4870:
4868:
4865:
4863:
4860:
4859:
4858:
4855:
4851:
4848:
4847:
4846:
4843:
4839:
4836:
4834:
4831:
4830:
4829:
4826:
4824:
4821:
4819:
4816:
4815:
4813:
4810:
4806:
4800:
4797:
4795:
4794:Speech coding
4792:
4790:
4789:Sound quality
4787:
4785:
4782:
4780:
4777:
4775:
4772:
4770:
4767:
4765:
4764:Dynamic range
4762:
4760:
4757:
4755:
4752:
4748:
4745:
4743:
4740:
4738:
4735:
4734:
4733:
4730:
4729:
4727:
4723:
4720:
4718:
4714:
4704:
4701:
4697:
4694:
4692:
4689:
4687:
4684:
4683:
4681:
4677:
4674:
4672:
4669:
4667:
4664:
4662:
4659:
4657:
4654:
4653:
4652:
4649:
4645:
4642:
4641:
4640:
4637:
4636:
4634:
4630:
4622:
4619:
4617:
4614:
4612:
4609:
4608:
4607:
4604:
4602:
4599:
4597:
4594:
4590:
4587:
4585:
4582:
4581:
4580:
4577:
4576:
4574:
4572:
4568:
4565:
4563:
4559:
4547:
4544:
4543:
4541:
4536:
4534:
4531:
4530:
4529:LZ77 + Range
4528:
4524:
4521:
4520:
4518:
4514:
4511:
4510:
4508:
4504:
4501:
4500:
4498:
4494:
4491:
4490:
4488:
4484:
4481:
4479:
4476:
4474:
4471:
4470:
4468:
4467:
4465:
4461:
4455:
4452:
4450:
4447:
4445:
4442:
4440:
4437:
4435:
4432:
4428:
4425:
4423:
4420:
4419:
4418:
4415:
4413:
4410:
4408:
4405:
4401:
4398:
4397:
4396:
4393:
4391:
4388:
4386:
4383:
4381:
4378:
4377:
4375:
4371:
4363:
4360:
4358:
4355:
4353:
4350:
4348:
4345:
4343:
4340:
4338:
4335:
4333:
4330:
4328:
4325:
4323:
4320:
4319:
4318:
4315:
4313:
4310:
4309:
4307:
4305:
4301:
4293:
4290:
4288:
4285:
4283:
4280:
4278:
4275:
4274:
4273:
4270:
4268:
4265:
4263:
4260:
4258:
4255:
4253:
4250:
4248:
4245:
4243:
4240:
4236:
4233:
4231:
4228:
4226:
4223:
4222:
4221:
4218:
4216:
4213:
4211:
4208:
4206:
4203:
4201:
4198:
4197:
4195:
4193:
4189:
4186:
4184:
4180:
4175:
4168:
4163:
4161:
4156:
4154:
4149:
4148:
4145:
4139:
4136:
4133:
4131:
4127:
4121:
4116:
4115:
4102:
4098:
4094:
4090:
4086:
4080:
4075:
4074:
4065:
4057:
4051:
4047:
4043:
4036:
4030:, p. 310
4029:
4024:
4015:
4010:
4006:
3999:
3992:
3988:
3977:
3974:
3971:
3968:
3962:
3959:
3956:
3955:Decorrelation
3953:
3950:
3947:
3944:
3941:
3940:
3934:
3917:
3911:
3888:
3882:
3879:
3876:
3868:
3852:
3849:
3846:
3826:
3823:
3820:
3812:
3808:
3804:
3785:
3779:
3771:
3761:
3759:
3754:
3740:
3737:
3734:
3725:
3723:
3705:
3701:
3666:
3663:
3660:
3657:
3654:
3644:
3641:
3636:
3633:
3622:
3619:
3616:
3613:
3610:
3600:
3597:
3594:
3591:
3586:
3580:
3572:
3568:
3564:
3558:
3550:
3546:
3538:
3534:
3528:
3522:
3515:
3514:
3513:
3511:
3501:
3486:
3483:
3480:
3476:
3466:
3464:
3458:
3456:
3451:
3426:
3421:
3416:
3412:
3408:
3405:
3395:
3392:
3383:
3378:
3374:
3370:
3367:
3364:
3361:
3351:
3345:
3341:
3335:
3330:
3326:
3319:
3314:
3310:
3304:
3301:
3293:
3288:
3282:
3276:
3269:
3268:
3267:
3265:
3261:
3257:
3256:
3251:
3235:
3213:
3209:
3201:
3197:
3181:
3167:
3147:
3144:
3133:
3129:
3125:
3120:
3116:
3109:
3103:
3100:
3092:
3088:
3084:
3079:
3075:
3071:
3066:
3062:
3051:
3047:
3043:
3038:
3034:
3029:
3022:
3008:
3007:
3006:
2984:
2980:
2976:
2971:
2967:
2960:
2948:
2941:
2937:
2933:
2928:
2924:
2919:
2908:
2905:
2900:
2894:
2886:
2882:
2874:
2873:
2872:
2852:
2844:
2840:
2828:
2820:
2814:
2808:
2801:
2800:
2799:
2795:
2793:
2789:
2784:
2782:
2778:
2755:
2749:
2746:
2740:
2734:
2731:
2725:
2719:
2712:
2711:
2710:
2708:
2703:
2701:
2697:
2694:
2690:
2685:
2681:
2676:
2660:
2656:
2632:
2629:
2626:
2618:
2615:
2612:
2608:
2584:
2576:
2572:
2551:
2528:
2525:
2522:
2518:
2515:
2509:
2501:
2498:
2495:
2486:
2478:
2474:
2467:
2464:
2461:
2453:
2450:
2447:
2443:
2429:
2425:
2411:
2407:
2403:
2400:
2397:
2393:
2390:
2384:
2376:
2373:
2370:
2361:
2358:
2355:
2347:
2344:
2341:
2337:
2323:
2319:
2305:
2301:
2297:
2292:
2288:
2280:
2279:
2278:
2276:
2272:
2268:
2249:
2246:
2243:
2235:
2232:
2229:
2225:
2204:
2184:
2162:
2158:
2135:
2131:
2121:
2104:
2098:
2095:
2089:
2083:
2080:
2074:
2071:
2068:
2062:
2042:
2039:
2033:
2030:
2027:
2021:
2001:
1981:
1961:
1958:
1952:
1949:
1946:
1940:
1917:
1911:
1908:
1902:
1899:
1896:
1890:
1882:
1877:
1862:
1858:
1855:
1852:
1848:
1844:
1821:
1818:
1815:
1809:
1801:
1797:
1792:
1775:
1772:
1769:
1763:
1760:
1757:
1749:
1745:
1730:
1727:
1724:
1716:
1712:
1705:
1697:
1694:
1691:
1683:
1680:
1677:
1673:
1660:
1659:
1658:
1656:
1637:
1634:
1631:
1627:
1624:
1614:
1611:
1608:
1600:
1597:
1594:
1590:
1583:
1578:
1574:
1567:
1559:
1555:
1548:
1545:
1542:
1534:
1531:
1528:
1524:
1510:
1506:
1492:
1488:
1484:
1481:
1475:
1472:
1469:
1463:
1456:
1455:
1440:
1437:
1427:
1419:
1415:
1408:
1403:
1399:
1392:
1384:
1380:
1366:
1362:
1358:
1355:
1349:
1343:
1336:
1335:
1334:
1332:
1328:
1309:
1306:
1303:
1297:
1274:
1268:
1241:
1238:
1235:
1229:
1226:
1220:
1214:
1211:
1205:
1202:
1199:
1193:
1186:
1185:
1184:
1170:
1150:
1142:
1141:
1121:
1118:
1115:
1107:
1103:
1082:
1062:
1054:
1051:
1032:
1029:
1026:
1018:
1015:
1012:
1008:
984:
979:
975:
971:
966:
962:
950:
947:
944:
936:
932:
923:
920:
917:
909:
906:
903:
899:
886:
885:
884:
860:
855:
845:
839:
836:
832:
827:
815:
809:
806:
800:
793:
792:
759:
753:
750:
740:
727:
721:
718:
708:
702:
697:
685:
679:
676:
670:
663:
662:
656:
636:
613:
599:
597:
596:normalization
593:
589:
585:
581:
577:
573:
569:
565:
561:
557:
553:
549:
545:
541:
537:
532:
528:
524:
519:
517:
512:
491:
481:
455:
451:
428:
418:
392:
388:
365:
361:
338:
334:
311:
307:
297:
288:
286:
282:
278:
274:
270:
260:
257:
249:
238:
235:
231:
228:
224:
221:
217:
214:
210:
207: –
206:
202:
201:Find sources:
195:
191:
185:
184:
179:This article
177:
173:
168:
167:
157:
152:
150:
145:
143:
138:
137:
135:
134:
129:
126:
124:
121:
119:
116:
114:
111:
110:
109:
108:
103:
100:
98:
95:
94:
93:
92:
87:
84:
82:
79:
77:
74:
72:
69:
67:
64:
62:
59:
57:
56:Joint entropy
54:
52:
49:
47:
44:
42:
39:
38:
37:
36:
32:
28:
27:
24:
21:
20:
5300:Hutter Prize
5268:
5264:Quantization
5169:Compensation
4963:Quantization
4686:Compensation
4252:Shannon–Fano
4192:Entropy type
4129:
4125:
4123:
4072:
4064:
4045:
4035:
4023:
4004:
3991:
3866:
3810:
3802:
3769:
3767:
3755:
3726:
3720:denotes the
3692:
3507:
3467:
3459:
3452:
3449:
3260:uncorrelated
3259:
3253:
3173:
3165:
3004:
2870:
2796:
2785:
2780:
2776:
2774:
2704:
2677:
2543:
2217:for a given
2122:
1880:
1878:
1799:
1795:
1793:
1790:
1652:
1326:
1260:
1138:
999:
882:
605:
592:quantization
550:and perhaps
530:
522:
520:
513:
509:
291:Introduction
284:
280:
268:
267:
252:
243:
233:
226:
219:
212:
200:
188:Please help
183:verification
180:
101:
81:Entropy rate
5259:Prefix code
5112:Frame types
4933:Color space
4759:Convolution
4489:LZ77 + ANS
4400:Incremental
4373:Other types
4292:Levenshtein
3976:White noise
1183:defined as
1050:conditional
590:weighting (
556:probability
5364:Categories
5316:Mark Adler
5274:Redundancy
5191:Daubechies
5174:Estimation
5107:Frame rate
5029:Daubechies
4989:Chain code
4948:Macroblock
4754:Companding
4691:Estimation
4611:Daubechies
4317:Lempel–Ziv
4277:Exp-Golomb
4205:Arithmetic
4014:1901.07821
3983:References
3463:quantizers
3255:memoryless
2689:continuous
2680:analytical
598:) matrix.
568:estimation
552:aesthetics
548:perception
531:distortion
246:March 2012
216:newspapers
5293:Community
5117:Interlace
4503:Zstandard
4282:Fibonacci
4272:Universal
4230:Canonical
4093:75-148254
3880:−
3850:−
3664:−
3620:−
3601:≤
3595:≤
3565:−
3413:σ
3375:σ
3371:≤
3365:≤
3327:σ
3320:
3210:σ
3145:≤
3072:∣
3044:∣
2949:∈
2934:∣
2835:∞
2832:→
2747:−
2732:≥
2661:∗
2630:∣
2616:∣
2499:−
2465:∣
2451:∣
2438:∞
2433:∞
2430:−
2426:∫
2420:∞
2415:∞
2412:−
2408:∫
2374:−
2332:∞
2327:∞
2324:−
2320:∫
2314:∞
2309:∞
2306:−
2302:∫
2271:amplitude
2247:∣
2233:∣
2163:∗
2031:∣
1900:∣
1819:∣
1770:≤
1695:∣
1681:∣
1612:∣
1598:∣
1584:
1546:∣
1532:∣
1519:∞
1514:∞
1511:−
1507:∫
1501:∞
1496:∞
1493:−
1489:∫
1485:−
1473:∣
1409:
1375:∞
1370:∞
1367:−
1363:∫
1359:−
1307:∣
1239:∣
1227:−
1030:∣
1016:∣
980:∗
972:≤
921:∣
907:∣
849:^
840:−
819:^
763:^
754:≠
731:^
689:^
640:^
485:^
422:^
5279:Symmetry
5247:Timeline
5230:FM-index
5075:Bit rate
5068:Concepts
4916:Concepts
4779:Sampling
4732:Bit rate
4725:Concepts
4427:Sequitur
4262:Tunstall
4235:Modified
4225:Adaptive
4183:Lossless
3937:See also
3839:), then
3402:if
3358:if
3200:variance
3196:Gaussian
2700:function
1329:and the
1143:between
747:if
715:if
5237:Entropy
5186:Wavelet
5165:Motion
5024:Wavelet
5004:Fractal
4999:Deflate
4982:Methods
4769:Latency
4682:Motion
4606:Wavelet
4523:LHA/LZH
4473:Deflate
4422:Re-Pair
4417:Grammar
4247:Shannon
4220:Huffman
4176:methods
3813:(where
1655:infimum
1137:is the
230:scholar
41:Entropy
5348:codecs
5309:People
5212:Theory
5179:Vector
4696:Vector
4513:Brotli
4463:Hybrid
4362:Snappy
4215:Golomb
4101:156968
4099:
4091:
4081:
4052:
3693:where
2871:where
2775:where
2696:convex
1261:where
1095:, and
580:Vorbis
232:
225:
218:
211:
203:
5139:parts
5137:Codec
5102:Frame
5060:Video
5044:SPIHT
4953:Pixel
4908:Image
4862:ACELP
4833:ADPCM
4823:μ-law
4818:A-law
4811:parts
4809:Codec
4717:Audio
4656:ACELP
4644:ADPCM
4621:SPIHT
4562:Lossy
4546:bzip2
4537:LZHAM
4493:LZFSE
4395:Delta
4287:Gamma
4267:Unary
4242:Range
4009:arXiv
4001:(PDF)
3194:is a
1000:Here
544:music
237:JSTOR
223:books
5151:DPCM
4958:PSNR
4889:MDCT
4882:WLPC
4867:CELP
4828:DPCM
4676:WLPC
4661:CELP
4639:DPCM
4589:MDCT
4533:LZMA
4434:LDCT
4412:DPCM
4357:LZWL
4347:LZSS
4342:LZRW
4332:LZJB
4130:beta
4097:OCLC
4089:LCCN
4079:ISBN
4050:ISBN
3824:<
3645:>
3409:>
3248:are
3005:and
2786:The
2698:(U)
2197:and
2150:and
2055:and
1933:and
1290:and
1163:and
588:MPEG
586:and
584:JPEG
570:and
527:bits
523:rate
209:news
5196:DWT
5146:DCT
5090:VBR
5085:CBR
5080:ABR
5039:EZW
5034:DWT
5019:RLE
5009:KLT
4994:DCT
4877:LSP
4872:LAR
4857:LPC
4850:FFT
4747:VBR
4742:CBR
4737:ABR
4671:LSP
4666:LAR
4651:LPC
4616:DWT
4601:FFT
4596:DST
4584:DCT
4483:LZS
4478:LZX
4454:RLE
4449:PPM
4444:PAQ
4439:MTF
4407:DMC
4385:CTW
4380:BWT
4352:LZW
4337:LZO
4327:LZ4
4322:842
3809:is
3741:0.5
3648:min
3604:min
3311:log
2915:inf
2825:lim
2678:An
2277:):
1669:inf
1575:log
1400:log
895:inf
578:or
576:MP3
558:in
192:by
5366::
5014:LP
4845:FT
4838:DM
4390:CM
4122:.
4095:.
4087:.
4044:.
4003:.
3933:.
3753::
3724:.
2691:,
2120:.
1881:no
1876:.
594:,
287:.
5350:)
5346:(
4166:e
4159:t
4152:v
4132:.
4126:R
4103:.
4058:.
4017:.
4011::
3921:)
3918:D
3915:(
3912:R
3892:)
3889:D
3886:(
3883:R
3877:H
3867:D
3853:C
3847:H
3827:H
3821:C
3811:C
3803:H
3789:)
3786:D
3783:(
3780:R
3770:D
3738:=
3735:p
3706:b
3702:H
3670:)
3667:p
3661:1
3658:,
3655:p
3652:(
3642:D
3637:,
3634:0
3626:)
3623:p
3617:1
3614:,
3611:p
3608:(
3598:D
3592:0
3587:,
3584:)
3581:D
3578:(
3573:b
3569:H
3562:)
3559:p
3556:(
3551:b
3547:H
3539:{
3535:=
3532:)
3529:D
3526:(
3523:R
3487:)
3484:D
3481:(
3477:R
3427:.
3422:2
3417:x
3406:D
3396:,
3393:0
3384:2
3379:x
3368:D
3362:0
3352:,
3349:)
3346:D
3342:/
3336:2
3331:x
3323:(
3315:2
3305:2
3302:1
3294:{
3289:=
3286:)
3283:D
3280:(
3277:R
3236:X
3214:2
3182:X
3151:}
3148:D
3142:]
3139:)
3134:n
3130:Y
3126:,
3121:n
3117:X
3113:(
3110:d
3107:[
3104:E
3101::
3098:)
3093:0
3089:X
3085:,
3080:n
3076:X
3067:n
3063:Y
3059:(
3052:n
3048:X
3039:n
3035:Y
3030:Q
3026:{
3023:=
3018:Q
2990:)
2985:n
2981:X
2977:,
2972:n
2968:Y
2964:(
2961:I
2954:Q
2942:n
2938:X
2929:n
2925:Y
2920:Q
2909:n
2906:1
2901:=
2898:)
2895:D
2892:(
2887:n
2883:R
2856:)
2853:D
2850:(
2845:n
2841:R
2829:n
2821:=
2818:)
2815:D
2812:(
2809:R
2781:D
2779:(
2777:h
2759:)
2756:D
2753:(
2750:h
2744:)
2741:X
2738:(
2735:h
2729:)
2726:D
2723:(
2720:R
2657:D
2636:)
2633:x
2627:y
2624:(
2619:X
2613:Y
2609:Q
2588:)
2585:x
2582:(
2577:X
2573:P
2552:X
2529:.
2526:y
2523:d
2519:x
2516:d
2510:2
2506:)
2502:y
2496:x
2493:(
2490:)
2487:x
2484:(
2479:X
2475:P
2471:)
2468:x
2462:y
2459:(
2454:X
2448:Y
2444:Q
2404:=
2401:y
2398:d
2394:x
2391:d
2385:2
2381:)
2377:y
2371:x
2368:(
2365:)
2362:y
2359:,
2356:x
2353:(
2348:Y
2345:,
2342:X
2338:P
2298:=
2293:Q
2289:D
2273:-
2253:)
2250:x
2244:y
2241:(
2236:X
2230:Y
2226:Q
2205:Y
2185:X
2159:D
2136:Q
2132:D
2108:)
2105:Y
2102:(
2099:H
2096:=
2093:)
2090:X
2087:(
2084:H
2081:=
2078:)
2075:X
2072:;
2069:Y
2066:(
2063:I
2043:0
2040:=
2037:)
2034:X
2028:Y
2025:(
2022:H
2002:X
1982:Y
1962:0
1959:=
1956:)
1953:X
1950:;
1947:Y
1944:(
1941:I
1921:)
1918:Y
1915:(
1912:H
1909:=
1906:)
1903:X
1897:Y
1894:(
1891:H
1863:)
1859:X
1856:;
1853:Y
1849:(
1845:I
1825:)
1822:X
1816:Y
1813:(
1810:H
1800:Y
1798:(
1796:H
1776:.
1773:R
1767:)
1764:X
1761:;
1758:Y
1755:(
1750:Q
1746:I
1737:]
1734:]
1731:Y
1728:,
1725:X
1722:[
1717:Q
1713:D
1709:[
1706:E
1701:)
1698:x
1692:y
1689:(
1684:X
1678:Y
1674:Q
1638:.
1635:y
1632:d
1628:x
1625:d
1621:)
1618:)
1615:x
1609:y
1606:(
1601:X
1595:Y
1591:Q
1587:(
1579:2
1571:)
1568:x
1565:(
1560:X
1556:P
1552:)
1549:x
1543:y
1540:(
1535:X
1529:Y
1525:Q
1482:=
1479:)
1476:X
1470:Y
1467:(
1464:H
1441:y
1438:d
1434:)
1431:)
1428:y
1425:(
1420:Y
1416:P
1412:(
1404:2
1396:)
1393:y
1390:(
1385:Y
1381:P
1356:=
1353:)
1350:Y
1347:(
1344:H
1327:Y
1313:)
1310:X
1304:Y
1301:(
1298:H
1278:)
1275:Y
1272:(
1269:H
1245:)
1242:X
1236:Y
1233:(
1230:H
1224:)
1221:Y
1218:(
1215:H
1212:=
1209:)
1206:X
1203:;
1200:Y
1197:(
1194:I
1171:X
1151:Y
1125:)
1122:X
1119:;
1116:Y
1113:(
1108:Q
1104:I
1083:X
1063:Y
1036:)
1033:x
1027:y
1024:(
1019:X
1013:Y
1009:Q
985:.
976:D
967:Q
963:D
954:)
951:X
948:;
945:Y
942:(
937:Q
933:I
927:)
924:x
918:y
915:(
910:X
904:Y
900:Q
861:2
856:)
846:x
837:x
833:(
828:=
825:)
816:x
810:,
807:x
804:(
801:d
760:x
751:x
741:1
728:x
722:=
719:x
709:0
703:{
698:=
695:)
686:x
680:,
677:x
674:(
671:d
637:x
614:x
506:.
492:n
482:X
456:n
452:X
429:n
419:X
393:n
389:g
366:n
362:Y
339:n
335:X
312:n
308:f
285:D
281:R
259:)
253:(
248:)
244:(
234:·
227:·
220:·
213:·
186:.
155:e
148:t
141:v
Text is available under the Creative Commons Attribution-ShareAlike License. Additional terms may apply.