View source for JPEG

{{short description|Lossy compression method for reducing the size of digital images}}
{{other uses|JPEG (disambiguation)}}
{{redirect2|JPG|Jpg|other uses|JPG (disambiguation)}}
{{Pp-pc|small=yes}}
{{use dmy dates|date=November 2023}}
{{Infobox file format
| name          = JPEG
| logo          = JPEG format logo.svg
| screenshot    = [[File:Felis silvestris silvestris small gradual decrease of quality - JPEG compression.jpg|border|200px]]
| caption       = A photo of a [[European wildcat]] with the compression rate, and associated losses, decreasing from left to right
| extension     = {{code|.jpg}}, {{code|.jpeg}}, {{code|.jpe}}<br />{{code|.jif}}, {{code|.jfif}}, {{code|.jfi}}
| mime          = image/jpeg
| typecode      = {{code|JPEG}}
| uniform type  = public.jpeg
| released      = {{start date and age|df=yes|1992|9|18}}
| owner         = [[Joint Photographic Experts Group]], [[IBM]], [[Mitsubishi Electric]], [[AT&T]], [[Canon Inc.]]<ref name="t81"/>
| genre         = [[Lossy compression|Lossy]] [[image compression]] [[image format|format]]
| container for = 
| contained by  = 
| extended to   = [[JPEG 2000]]
| magic         = {{code|ff d8 ff}}
| standard      = ISO/IEC 10918, ITU-T T.81, ITU-T T.83, ITU-T T.84, ITU-T T.86
| url           = {{Official URL}}
}}
[[File:Continuously varied JPEG compression for an abdominal CT scan - 1471-2342-12-24-S1.ogv|thumb|Continuously varied JPEG compression (between Q=100 and Q=1) for an [[abdominal]] [[CT scan]]]]

'''JPEG''' ({{IPAc-en|ˈ|dʒ|eɪ|p|ɛ|ɡ}} {{respell|JAY|peg}}, short for '''Joint Photographic Experts Group''' and sometimes retroactively referred to as '''JPEG 1''')<ref>{{cite web|url=http://www.collinsdictionary.com/dictionary/english/jpeg|title=Definition of "JPEG"|work=[[Collins English Dictionary]]|access-date=2013-05-23|archive-date=2013-09-21|archive-url=https://web.archive.org/web/20130921054412/http://www.collinsdictionary.com/dictionary/english/jpeg|url-status=live}}</ref><ref name="JPEG-homepage">{{cite web|url=http://jpeg.org/jpeg|title=Overview of JPEG 1|website=Joint Photographic Experts Group|access-date=2025-02-05|archive-date=2025-01-30|archive-url=https://web.archive.org/web/20250130141949/https://jpeg.org/jpeg/|url-status=live}}</ref> is a commonly used method of [[lossy compression]] for [[digital image]]s, particularly for those images produced by [[digital photography]]. The degree of compression can be adjusted, allowing a selectable trade off between storage size and [[image quality]]. JPEG typically achieves 10:1 compression with noticeable, but widely agreed to be acceptable perceptible loss in image quality.<ref>{{cite tech report |url=https://ntrs.nasa.gov/search.jsp?R=19920024689 |last1=Haines |first1=Richard F. |last2=Chuang |first2=Sherry L. |title=The effects of video compression on acceptability of images for monitoring life sciences experiments |number=NASA-TP-3239, A-92040, NAS 1.60:3239  |institution=[[NASA]] |date=1 July 1992  |access-date=2016-03-13 |quote=The JPEG still-image-compression levels, even with the large range of 5:1 to 120:1 in this study, yielded equally high levels of acceptability}}</ref> Since its introduction in 1992, JPEG has been the most widely used [[image compression]] standard in the world,<ref name="Hudson"/><ref>{{cite web |title=The JPEG Image Format Explained |url=https://home.bt.com/tech-gadgets/photography/what-is-a-jpeg-11364206889349 |website=[[BT.com]] |publisher=[[BT Group]] |access-date=5 August 2019 |date=31 May 2018 |archive-date=5 August 2019 |archive-url=https://web.archive.org/web/20190805194553/https://home.bt.com/tech-gadgets/photography/what-is-a-jpeg-11364206889349|first=Joe|last=Svetlik}}</ref> and the most widely used digital [[image format]], with several billion JPEG images produced every day as of 2015.<ref>{{cite web |last=Baraniuk |first=Chris |title=Copy Protections Could Come to JPEGs |url=https://www.bbc.co.uk/news/technology-34538705 |access-date=13 September 2019 |website=[[BBC News]] |agency=[[BBC]] |date=15 October 2015 |archive-date=9 October 2019 |archive-url=https://web.archive.org/web/20191009193610/https://www.bbc.co.uk/news/technology-34538705 |url-status=live }}</ref>

The [[Joint Photographic Experts Group]] created the standard in 1992,<ref name="heise">{{cite news |date=7 October 2016 |title=JPEG: 25 Jahre und kein bisschen alt |language=de |work=[[:de:Heise online]] |url=https://www.heise.de/newsticker/meldung/JPEG-25-Jahre-und-kein-bisschen-alt-3342519.html |access-date=5 September 2019 |archive-date=5 September 2019 |archive-url=https://web.archive.org/web/20190905025359/https://www.heise.de/newsticker/meldung/JPEG-25-Jahre-und-kein-bisschen-alt-3342519.html |url-status=live |first=Andrea |last=Trinkwalder |trans-title=JPEG: 25 years (old) and not a bit old}}</ref> based on the [[discrete cosine transform]] (DCT) algorithm.<ref name="t812">{{cite web |date=September 1992 |title=T.81 – DIGITAL COMPRESSION AND CODING OF CONTINUOUS-TONE STILL IMAGES – REQUIREMENTS AND GUIDELINES |url=https://www.w3.org/Graphics/JPEG/itu-t81.pdf |accessdate=12 July 2019 |publisher=[[CCITT]]}}</ref><ref name="heise3">{{cite news |date=October 2016 |title=JPEG: 25 Jahre und kein bisschen alt |url=https://www.heise.de/newsticker/meldung/JPEG-25-Jahre-und-kein-bisschen-alt-3342519.html |accessdate=5 September 2019 |work=[[:de:Heise online|Heise online]] |language=de}}</ref> JPEG was largely responsible for the proliferation of digital images and [[digital photo]]s across the Internet and later [[social media]].<ref name="Atlantic">{{cite web |title=What Is a JPEG? The Invisible Object You See Every Day |url=https://www.theatlantic.com/technology/archive/2013/09/what-is-a-jpeg-the-invisible-object-you-see-every-day/279954/ |access-date=13 September 2019 |website=[[The Atlantic]] |date=24 September 2013 |archive-date=9 October 2019 |archive-url=https://web.archive.org/web/20191009054159/https://www.theatlantic.com/technology/archive/2013/09/what-is-a-jpeg-the-invisible-object-you-see-every-day/279954/ |url-status=live |first=Paul |last=Caplan|url-access=limited}}</ref>{{Circular reference|date=November 2023}}<!--The Atlantic link to the legal issues section of this Wikipedia article in the open part of their article, makes one suspicious that their article is largely based on Wikipedia and thus invalid as a source.--> JPEG compression is used in a number of [[image file formats]]. JPEG/[[Exif]] is the most common image format used by [[digital camera]]s and other photographic image capture devices; along with JPEG/[[JFIF]], it is the most common format for storing and transmitting [[photographic image]]s on the [[World Wide Web]].<ref>{{cite web|url=http://httparchive.org/interesting.php#imageformats|title=HTTP Archive – Interesting Stats|website=httparchive.org|access-date=2016-04-06}}</ref> These format variations are often not distinguished and are simply called JPEG.

The [[Internet media type|MIME media type]] for JPEG is "image/jpeg", except in older [[Internet Explorer]] versions, which provide a MIME type of "image/pjpeg" when uploading JPEG images.<ref>{{cite web|url=https://learn.microsoft.com/en-us/previous-versions/windows/internet-explorer/ie-developer/platform-apis/ms775147(v=vs.85)|title=MIME Type Detection in Internet Explorer|date=13 July 2016 |publisher=Microsoft|access-date=2 November 2022|archive-date=30 October 2022|archive-url=https://web.archive.org/web/20221030181306/https://learn.microsoft.com/en-us/previous-versions/windows/internet-explorer/ie-developer/platform-apis/ms775147(v=vs.85)|url-status=live}}</ref> JPEG files usually have a [[filename extension]] of "jpg" or "jpeg". JPEG/JFIF supports a maximum image size of 65,535×65,535 pixels,<ref>{{cite web|url=http://www.jpeg.org/public/jfif.pdf|title=JPEG File Interchange Format|date=3 September 2014|access-date=16 October 2017|url-status=bot: unknown|archive-url=https://web.archive.org/web/20140903080533/http://www.jpeg.org/public/jfif.pdf|archive-date=3 September 2014}}</ref> hence up to 4 gigapixels for an [[aspect ratio (image)|aspect ratio]] of 1:1. In 2000, the JPEG group introduced a format intended to be a successor, [[JPEG 2000]], but it was unable to replace the original JPEG as the dominant image standard.<ref>{{cite web |title=Why JPEG 2000 Never Took Off |url=https://blog.ansi.org/2018/07/why-jpeg-2000-never-used-standard-iso-iec/ |website=[[American National Standards Institute]] |access-date=13 September 2019 |date=10 July 2018 |archive-date=16 December 2018 |archive-url=https://web.archive.org/web/20181216031019/https://blog.ansi.org/2018/07/why-jpeg-2000-never-used-standard-iso-iec/ |url-status=live }}</ref>

==History==
===Background===
The original JPEG specification published in 1992 implements processes from various earlier [[Academic paper|research papers]] and [[patent]]s cited by the [[CCITT]] (now [[ITU-T]]) and Joint Photographic Experts Group.<ref name="t81">{{cite web |title=T.81 – DIGITAL COMPRESSION AND CODING OF CONTINUOUS-TONE STILL IMAGES – REQUIREMENTS AND GUIDELINES |url=https://www.w3.org/Graphics/JPEG/itu-t81.pdf |publisher=[[CCITT]] |date=September 1992 |access-date=12 July 2019 |archive-date=30 December 2019 |archive-url=https://web.archive.org/web/20191230093239/http://www.w3.org/Graphics/JPEG/itu-t81.pdf |url-status=live }}</ref>

The basis for JPEG's [[lossy compression]] algorithm is the [[discrete cosine transform]] (DCT),<ref name="t812" /><ref name="heise3" /> which was first proposed by [[Nasir Ahmed (engineer)|Nasir Ahmed]] as an [[image compression]] technique in 1972.<ref name="Ahmed2">{{cite journal |last=Ahmed |first=Nasir |author-link=N. Ahmed |date=January 1991 |title=How I Came Up With the Discrete Cosine Transform |url=https://www.scribd.com/doc/52879771/DCT-History-How-I-Came-Up-with-the-Discrete-Cosine-Transform |journal=[[Digital Signal Processing (journal)|Digital Signal Processing]] |volume=1 |issue=1 |pages=4–5 |doi=10.1016/1051-2004(91)90086-Z|bibcode=1991DSP.....1....4A }}</ref><ref name="heise3" /> Ahmed published the DCT algorithm with T. Natarajan and [[K. R. Rao]] in a 1974 paper,<ref name="pubDCT">{{Citation |last1=Ahmed |first1=Nasir |title=Discrete Cosine Transform |date=January 1974 |journal=IEEE Transactions on Computers |volume=C-23 |issue=1 |pages=90–93 |doi=10.1109/T-C.1974.223784 |s2cid=149806273 |last2=Natarajan |first2=T. |last3=Rao |first3=K. R. |author1-link=N. Ahmed |author3-link=K. R. Rao}}</ref> which is cited in the JPEG specification.<ref name="t812" />

The JPEG specification cites patents from several companies. The following patents provided the basis for its [[arithmetic coding]] algorithm.<ref name="t81"/>
* [[IBM]]
** {{US patent|4652856}} {{ndash}} 4 February 1986 {{ndash}} Kottappuram M. A. Mohiuddin and [[Jorma Rissanen|Jorma J. Rissanen]] {{ndash}} Multiplication-free multi-alphabet arithmetic code
** {{US patent|4905297}} {{ndash}} 27 February 1990 {{ndash}} G. Langdon, [[Joan L. Mitchell|J. L. Mitchell]], W. B. Pennebaker, and Jorma J. Rissanen {{ndash}} Arithmetic coding encoder and decoder system
** {{US patent|4935882}} {{ndash}} 19 June 1990 {{ndash}} W. B. Pennebaker and J. L. Mitchell {{ndash}} Probability adaptation for arithmetic coders
* [[Mitsubishi Electric]]
** {{Patent|JP|H02202267}} ([https://patents.google.com/patent/JPH02202267A 1021672]) {{ndash}} 21 January 1989 {{ndash}} Toshihiro Kimura, Shigenori Kino, Fumitaka Ono, Masayuki Yoshida {{ndash}} Coding system
** {{Patent|JP|H03247123}} ([https://patents.google.com/patent/JPH0834434B2/en 2-46275]) {{ndash}} 26 February 1990 {{ndash}} Tomohiro Kimura, Shigenori Kino, Fumitaka Ono, and Masayuki Yoshida {{ndash}} Coding apparatus and coding method

The JPEG specification also cites three other patents from IBM. Other companies cited as patent holders include [[AT&T]] (two patents) and [[Canon Inc.]]<ref name="t81"/> Absent from the list is {{US patent|4698672}}, filed by [[Compression Labs, Inc.|Compression Labs]]' Wen-Hsiung Chen and Daniel J. Klenke in October 1986. The patent describes a DCT-based image compression algorithm, and would later be a cause of controversy in 2002 (see ''[[#Patent controversy|Patent controversy]]'' below).<ref name="cnet">{{cite news |last1=Lemos |first1=Robert |title=Finding patent truth in JPEG claim |url=https://www.cnet.com/news/finding-patent-truth-in-jpeg-claim/ |access-date=13 July 2019 |work=[[CNET]] |date=23 July 2002 |archive-date=13 July 2019 |archive-url=https://web.archive.org/web/20190713165715/https://www.cnet.com/news/finding-patent-truth-in-jpeg-claim/ |url-status=live }}</ref> However, the JPEG specification did cite two earlier research papers by Wen-Hsiung Chen, published in 1977 and 1984.<ref name="t81"/>

===JPEG standard===
"JPEG" stands for [[Joint Photographic Experts Group]], the name of the committee that created the JPEG standard and other still picture coding standards. The "Joint" stood for [[ISO]] TC97 WG8 and [[CCITT]] SGVIII. Founded in 1986, the group developed the JPEG standard during the late 1980s. The group published the JPEG standard in 1992.<ref name="Hudson">{{cite journal |last1=Hudson |first1=Graham |last2=Léger |first2=Alain |last3=Niss |first3=Birger |last4=Sebestyén |first4=István |last5=Vaaben |first5=Jørgen |title=JPEG-1 standard 25 years: past, present, and future reasons for a success |journal=[[Journal of Electronic Imaging]] |date=31 August 2018 |volume=27 |issue=4 |page=1 |doi=10.1117/1.JEI.27.4.040901|s2cid=52164892 |doi-access=free }}</ref>

In 1987, ISO TC 97 became ISO/IEC JTC 1 and, in 1992, CCITT became ITU-T. Currently on the JTC1 side, JPEG is one of two sub-groups of [[International Organization for Standardization|ISO]]/[[International Electrotechnical Commission|IEC]] [[ISO/IEC JTC 1|Joint Technical Committee 1]], Subcommittee 29, Working Group 1 ([[ISO/IEC JTC 1/SC 29]]/WG 1)&nbsp;– titled as ''Coding of still pictures''.<ref name="sc29-wg1">{{cite web |url=http://kikaku.itscj.ipsj.or.jp/sc29/29w12901.htm |title=ISO/IEC JTC 1/SC 29/WG 1&nbsp;– Coding of Still Pictures (SC 29/WG 1 Structure) |author=ISO/IEC JTC 1/SC 29 |date=2009-05-07 |access-date=2009-11-11  |archive-url=https://web.archive.org/web/20131231055215/http://kikaku.itscj.ipsj.or.jp/sc29/29w12901.htm |archive-date=2013-12-31 }}</ref><ref name="jpeg-sc29">{{cite web |url=http://kikaku.itscj.ipsj.or.jp/sc29/29w42901.htm |title=Programme of Work, (Allocated to SC 29/WG 1) |author=ISO/IEC JTC 1/SC 29 |access-date=2009-11-07  |archive-url=https://web.archive.org/web/20131231012300/http://kikaku.itscj.ipsj.or.jp/sc29/29w42901.htm |archive-date=2013-12-31 }}</ref><ref>{{cite web |url=http://www.iso.org/iso/standards_development/technical_committees/list_of_iso_technical_committees/iso_technical_committee.htm?commid=45316 |title=JTC 1/SC 29&nbsp;– Coding of audio, picture, multimedia and hypermedia information |author=ISO |access-date=2009-11-11 |archive-date=2010-07-03 |archive-url=https://web.archive.org/web/20100703085542/http://www.iso.org/iso/standards_development/technical_committees/list_of_iso_technical_committees/iso_technical_committee.htm?commid=45316 |url-status=live }}</ref> On the ITU-T side, ITU-T SG16 is the respective body. The original JPEG Group was organized in 1986,<ref name="jpeg-jpeg">{{cite web | url=http://www.jpeg.org/jpeg/index.html | title=Joint Photographic Experts Group, JPEG Homepage | author=JPEG | access-date=2009-11-08 | archive-date=2009-09-27 | archive-url=https://web.archive.org/web/20090927035043/http://www.jpeg.org/jpeg/index.html | url-status=live }}</ref> issuing the first JPEG standard in 1992, which was approved in September 1992 as '''[[ITU-T]] Recommendation T.81'''<ref>{{cite web|url=http://www.itu.int/rec/T-REC-T.81|title=T.81: Information technology&nbsp;– Digital compression and coding of continuous-tone still images&nbsp;– Requirements and guidelines|website=Itu.int|access-date=2009-11-07|archive-date=2012-11-06|archive-url=https://web.archive.org/web/20121106222712/http://www.itu.int/rec/T-REC-T.81|url-status=live}}</ref> and, in 1994, as '''[[International Organization for Standardization|ISO]]/[[International Electrotechnical Commission|IEC]] 10918-1'''.

The JPEG standard specifies the [[codec]], which defines how an image is compressed into a stream of [[byte]]s and decompressed back into an image, but not the file format used to contain that stream.<ref>
{{cite book
 | title = JPEG still image data compression standard
 | edition = 3rd
 |author1=William B. Pennebaker |author2=Joan L. Mitchell|author2-link=Joan L. Mitchell | publisher = Springer
 | year = 1993
 | isbn = 978-0-442-01272-4
 | page = 291
 | url = https://books.google.com/books?id=AepB_PZ_WMkC&dq=JPEG+%22did+not+specify+a+file+format%22&pg=PA291
 }}</ref>
The [[Exif]] and [[JFIF]] standards define the commonly used file formats for interchange of JPEG-compressed images.

JPEG standards are formally named as ''Information technology&nbsp;– Digital compression and coding of continuous-tone still images''. ISO/IEC 10918 consists of the following parts:

{| class="wikitable sortable" style="width:100%;"
|+Digital compression and coding of continuous-tone still images&nbsp;–&nbsp;Parts<ref name="jpeg-sc29" /><ref name="jpeg-jpeg" /><ref>{{cite web |url=http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_tc_browse.htm?commid=45316 |title=JTC 1/SC 29&nbsp;– Coding of audio, picture, multimedia and hypermedia information |author=ISO |access-date=2009-11-07 |archive-date=2010-07-03 |archive-url=https://web.archive.org/web/20100703224947/http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_tc_browse.htm?commid=45316 |url-status=live }}</ref>
|-
! style="width:6%;"| Part
! style="width:15%;"| ISO/IEC standard
! style="width:10%;"| ITU-T Rec.
! style="width:8%;"| First public release date
! style="width:8%;"| Latest amendment
! style="width:30%;"| Title
! Description
|-
| Part 1
| [https://www.iso.org/standard/18902.html ISO/IEC 10918-1:1994]
| [http://www.itu.int/rec/T-REC-T.81 T.81 (09/92)]
| {{dts|abbr=on|1992-09-18}}
|
| Requirements and guidelines
|
|-
| Part 2
| [https://www.iso.org/standard/20689.html ISO/IEC 10918-2:1995]
| [http://www.itu.int/rec/T-REC-T.83 T.83 (11/94)]
| {{dts|abbr=on|1994-11-11}}
|
| Compliance testing
| Rules and checks for software conformance (to Part 1).
|-
| Part 3
| [https://www.iso.org/standard/25037.html ISO/IEC 10918-3:1997]
| [http://www.itu.int/rec/T-REC-T.84 T.84 (07/96)]
| {{dts|abbr=on|1996-07-03}}
| {{dts|abbr=on|1999-04-01}}
| Extensions
| Set of extensions to improve the Part 1, including the '''Still Picture Interchange File Format''' (SPIFF).<ref name="spiffloc" >{{cite web|url=https://www.loc.gov/preservation/digital/formats/fdd/fdd000019.shtml|title=SPIFF, Still Picture Interchange File Format |date=30 January 2012 |publisher= [[Library of Congress]] |access-date=2018-07-31 |archive-url=https://web.archive.org/web/20180731093543/https://www.loc.gov/preservation/digital/formats/fdd/fdd000019.shtml |archive-date=2018-07-31 |url-status=live}}</ref>
|-
| Part 4
| [https://www.iso.org/standard/85634.html ISO/IEC 10918-4:2024]
| [http://www.itu.int/rec/T-REC-T.86 T.86 (06/98)]
| {{dts|abbr=on|1998-06-18}}
|
| Appn Markers
| methods for registering some of the parameters used to extend JPEG
|-
| Part 5
| [https://www.iso.org/standard/54989.html ISO/IEC 10918-5:2013]
| [http://www.itu.int/rec/T-REC-T.871 T.871 (05/11)]
| {{dts|abbr=on|2011-05-14}}
|
| JPEG File Interchange Format (JFIF)
| A popular format which has been the de facto file format for images encoded by the JPEG standard. In 2009, the JPEG Committee formally established an Ad Hoc Group to standardize JFIF as JPEG Part 5.<ref name="jfif-development">{{cite press release|url=http://www.jpeg.org/newsrel25.html |title=JPEG XR enters FDIS status JPEG File Interchange Format (JFIF) to be standardized as JPEG Part 5 |author=Louis Sharpe|date=2009-04-24 |access-date=2009-11-09  |archive-url=https://web.archive.org/web/20091008041637/http://www.jpeg.org/newsrel25.html |archive-date=2009-10-08 }}</ref>
|-
| Part 6
| [https://www.iso.org/standard/59634.html ISO/IEC 10918-6:2013]
| [http://www.itu.int/rec/T-REC-T.872 T.872 (06/12)]
| {{dts|abbr=on|2012-06}}
|
| Application to printing systems
| Specifies a subset of features and application tools for the interchange of images encoded according to the ISO/IEC 10918-1 for printing.
|-
|Part 7
|[https://www.iso.org/standard/85635.html ISO/IEC 10918-7:2023]
|[http://www.itu.int/rec/T-REC-T.873 T.873 (06/21)]
|May 2019
|November 2023
|Reference Software
|Provides reference implementations of the JPEG core coding system
|}

[[Ecma International]] <abbr title="Technical Report">TR</abbr>/98 specifies the JPEG File Interchange Format (JFIF); the first edition was published in June 2009.<ref>{{cite web
|url=http://www.ecma-international.org/publications/techreports/E-TR-098.htm
|year=2009
|title=JPEG File Interchange Format (JFIF)
|access-date=2011-08-01
|work=ECMA TR/98 1st ed.
|publisher=[[Ecma International]]
|archive-date=2021-01-14
|archive-url=https://web.archive.org/web/20210114101520/http://www.ecma-international.org/publications/techreports/E-TR-098.htm
|url-status=live
}}</ref>

===Patent controversy===
<!-- Courtesy note per [[MOS:LINK2SECT]]: [[Patent troll]] links here -->

In 2002, [[Forgent Networks]] asserted that it owned and would enforce patent rights on the JPEG technology, arising from a patent that had been filed on 27 October 1986, and granted on 6 October 1987: {{US patent|4698672}} by [[Compression Labs, Inc.|Compression Labs]]' Wen-Hsiung Chen and Daniel J. Klenke.<ref name="cnet"/><ref>{{cite news |title=Forgent's JPEG Patent |url=https://pmt.sourceforge.io/SVG-patents/jpeg.html |access-date=13 July 2019 |work=[[SourceForge]] |date=2002 |archive-date=13 May 2019 |archive-url=https://web.archive.org/web/20190513043628/https://pmt.sourceforge.io/SVG-patents/jpeg.html |url-status=live }}</ref> While Forgent did not own Compression Labs at the time, Chen later sold Compression Labs to Forgent, before Chen went on to work for [[Cisco]]. This led to Forgent acquiring ownership over the patent.<ref name="cnet"/> Forgent's 2002 announcement created a furor reminiscent of [[Unisys]]' attempts to assert its rights over the GIF image compression standard.

The JPEG committee investigated the patent claims in 2002 and were of the opinion that they were invalidated by [[prior art]],<ref>{{cite web |url=http://www.jpeg.org/newsrel1.html |title=Concerning recent patent claims |website=Jpeg.org |date=2002-07-19 |access-date=2011-05-29  |archive-url=https://web.archive.org/web/20070714232941/http://www.jpeg.org/newsrel1.html |archive-date=2007-07-14 }}</ref> a view shared by various experts.<ref name="cnet"/><ref>{{cite web|url=http://www.algovision-luratech.com/company/news/patentquarrel.jsp?OnlineShopId=164241031081525276 |title=JPEG and JPEG2000&nbsp;– Between Patent Quarrel and Change of Technology |access-date=2017-04-16 |url-status=bot: unknown |archive-url=https://web.archive.org/web/20040817154508/http://www.algovision-luratech.com/company/news/patentquarrel.jsp?OnlineShopId=164241031081525276 |archive-date=17 August 2004 }}</ref>

Between 2002 and 2004, Forgent was able to obtain about US$105 million by licensing their patent to some 30 companies. In April 2004, Forgent sued 31 other companies to enforce further license payments. In July of the same year, a consortium of 21 large computer companies filed a countersuit, with the goal of invalidating the patent. In addition, Microsoft launched a separate lawsuit against Forgent in April 2005.<ref>{{cite web|title=Graphics patent suit fires back at Microsoft|publisher=[[CNET News]]|first=Dawn|last=Kawamoto|date=22 April 2005|url=https://www.cnet.com/culture/graphics-patent-suit-fires-back-at-microsoft/|access-date=2023-01-20|archive-date=2023-01-20|archive-url=https://web.archive.org/web/20230120140214/https://www.cnet.com/culture/graphics-patent-suit-fires-back-at-microsoft/|url-status=live}}</ref> In February 2006, the [[United States Patent and Trademark Office]] agreed to re-examine Forgent's JPEG patent at the request of the Public Patent Foundation.<ref name="reexam">{{cite web |title=Trademark Office Re-examines Forgent JPEG Patent |website=Publish.com |date=3 February 2006 |url=http://www.publish.com/c/a/Graphics-Tools/Trademark-Office-Reexamines-Forgent-JPEG-Patent/ |access-date=2009-01-28 |archive-date=2016-05-15 |archive-url=http://arquivo.pt/wayback/20160515113624/http://www.publish.com/c/a/Graphics-Tools/Trademark-Office-Reexamines-Forgent-JPEG-Patent/  }}</ref> On 26 May 2006, the USPTO found the patent invalid based on prior art. The USPTO also found that Forgent knew about the prior art, yet it intentionally avoided telling the Patent Office. This made any appeal to reinstate the patent highly unlikely to succeed.<ref>{{cite web |title=USPTO: Broadest Claims Forgent Asserts Against JPEG Standard Invalid |website=[[Groklaw.net]] |date=26 May 2006 |url=http://www.groklaw.net/article.php?story=20060526105754880 |access-date=2007-07-21 |archive-date=2019-05-16 |archive-url=https://web.archive.org/web/20190516201650/http://www.groklaw.net/article.php?story=20060526105754880 |url-status=usurped }}</ref>

Forgent also possesses a similar patent granted by the [[European Patent Office]] in 1994, though it is unclear how enforceable it is.<ref name="ep266049">{{cite web |url=http://gauss.ffii.org/PatentView/EP266049 |title=Coding System for Reducing Redundancy |website=Gauss.ffii.org |access-date=2011-05-29 |archive-date=2011-06-12 |archive-url=https://web.archive.org/web/20110612114555/http://gauss.ffii.org/PatentView/EP266049  }}</ref>

As of 27 October 2006, the U.S. patent's 20-year term appears to have expired, and in November 2006, Forgent agreed to abandon enforcement of patent claims against use of the JPEG standard.<ref name="surrendered">{{cite web |title=JPEG Patent Claim Surrendered |publisher=Public Patent Foundation |date=2 November 2006 |url=http://www.pubpat.org/jpegsurrendered.htm |access-date=2006-11-03 |archive-date=2007-01-02 |archive-url=https://web.archive.org/web/20070102145054/http://www.pubpat.org/jpegsurrendered.htm |url-status=live }}</ref>

The JPEG committee has as one of its explicit goals that their standards (in particular their baseline methods) be implementable without payment of license fees, and they have secured appropriate license rights for their [[JPEG 2000]] standard from over 20 large organizations.

Beginning in August 2007, another company, Global Patent Holdings, LLC claimed that its patent ({{US patent|5,253,341}}) issued in 1993, is infringed by the downloading of JPEG images on either a website or through e-mail. If not invalidated, this patent could apply to any website that displays JPEG images. The patent was under reexamination by the U.S. Patent and Trademark Office from 2000 to 2007; in July 2007, the Patent Office revoked all of the original claims of the patent but found that an additional claim proposed by Global Patent Holdings (claim 17) was valid.<ref name="GPHLaw1">{{Cite web|url=http://www.uspto.gov/web/patents/patog/week30/OG/html/1320-4/US05253341-20070724.html|archive-url=https://web.archive.org/web/20080602141045/http://www.uspto.gov/web/patents/patog/week30/OG/html/1320-4/US05253341-20070724.html|title=Ex Parte Reexamination Certificate for U.S. Patent No. 5,253,341|archive-date=2 June 2008}}</ref> Global Patent Holdings then filed a number of lawsuits based on claim 17 of its patent.

In its first two lawsuits following the reexamination, both filed in Chicago, Illinois, Global Patent Holdings sued the [[Green Bay Packers]], [[CDW]], [[Motorola]], [[Apple Inc.|Apple]], [[Orbitz]], [[Officemax]], [[Caterpillar Inc.|Caterpillar]], [[Kraft]] and [[Peapod]] as defendants. A third lawsuit was filed on 5 December 2007, in South Florida against [[ADT Security Services]], [[AutoNation]], [[Fanjul brothers|Florida Crystals]] Corp., HearUSA, [[MovieTickets.com]], [[Ocwen Financial Corp.]] and [[Tire Kingdom]], and a fourth lawsuit on 8 January 2008, in South Florida against the [[Boca Raton Resort & Club]]. A fifth lawsuit was filed against Global Patent Holdings in Nevada. That lawsuit was filed by [[Zappos.com]], Inc., which was allegedly threatened by Global Patent Holdings, and sought a judicial declaration that the '341 patent is invalid and not infringed.

Global Patent Holdings had also used the '341 patent to sue or threaten outspoken critics of broad software patents, including Gregory Aharonian<ref name="GPHLaw2">{{cite web|author=Workgroup |url=http://eupat.ffii.org/pikta/xrani/rozmanith/index.en.html |title=Rozmanith: Using Software Patents to Silence Critics |website=Eupat.ffii.org |access-date=2011-05-29  |archive-url=https://web.archive.org/web/20110716123228/http://eupat.ffii.org/pikta/xrani/rozmanith/index.en.html |archive-date=2011-07-16 }}</ref> and the anonymous operator of a website blog known as the "[[Patent Troll Tracker]]."<ref name="GPHLaw3">{{cite web |url=http://www.law.com/jsp/article.jsp?id=1196762670106 |title=A Bounty of $5,000 to Name Troll Tracker: Ray Niro Wants To Know Who Is saying All Those Nasty Things About Him |website=Law.com |access-date=2011-05-29 |archive-date=2010-11-21 |archive-url=https://web.archive.org/web/20101121012911/http://www.law.com/jsp/article.jsp?id=1196762670106 |url-status=live }}</ref> On 21 December 2007, patent lawyer Vernon Francissen of Chicago asked the U.S. Patent and Trademark Office to reexamine the sole remaining claim of the '341 patent on the basis of new prior art.<ref name="GPHLaw4">{{cite web |last=Reimer |first=Jeremy |url=https://arstechnica.com/news.ars/post/20080205-hunting-trolls-uspto-asked-to-reexamine-broad-image-patent.html |title=Hunting trolls: USPTO asked to reexamine broad image patent |website=[[Arstechnica.com]] |date=2008-02-05 |access-date=2011-05-29 |archive-date=2008-12-08 |archive-url=https://web.archive.org/web/20081208145656/http://arstechnica.com/news.ars/post/20080205-hunting-trolls-uspto-asked-to-reexamine-broad-image-patent.html |url-status=live }}</ref>

On 5 March 2008, the U.S. Patent and Trademark Office agreed to reexamine the '341 patent, finding that the new prior art raised substantial new questions regarding the patent's validity.<ref name="GPHLaw11">[[commons:Image:2008-03-05 USPTO Determination Granting Reexam of 5,253,341 C1.pdf|U.S. Patent Office&nbsp;– Granting Reexamination on 5,253,341 C1]]</ref> In light of the reexamination, the accused infringers in four of the five pending lawsuits have filed motions to suspend (stay) their cases until completion of the U.S. Patent and Trademark Office's review of the '341 patent. On 23 April 2008, a judge presiding over the two lawsuits in Chicago, Illinois granted the motions in those cases.<ref name="GPHLaw55">{{cite web |url=http://www.techdirt.com/articles/20080427/143205960.shtml |title=Judge Puts JPEG Patent On Ice |website=Techdirt.com |date=2008-04-30 |access-date=2011-05-29 |archive-date=2011-11-14 |archive-url=https://web.archive.org/web/20111114020657/http://www.techdirt.com/articles/20080427/143205960.shtml |url-status=live }}</ref> On 22 July 2008, the Patent Office issued the first "Office Action" of the second reexamination, finding the claim invalid based on nineteen separate grounds.<ref name="GPHLaw66">{{cite web |url=http://techdirt.com/articles/20080731/0337491852.shtml |title=JPEG Patent's Single Claim Rejected (And Smacked Down For Good Measure) |website=[[Techdirt.com]] |date=2008-08-01 |access-date=2011-05-29 |archive-date=2019-11-28 |archive-url=https://web.archive.org/web/20191128051716/https://www.techdirt.com/articles/20080731/0337491852.shtml |url-status=live }}</ref> On 24 November 2009, a Reexamination Certificate was issued cancelling all claims.{{fact|date=August 2025}}

Beginning in 2011 and continuing as of early 2013, an entity known as Princeton Digital Image Corporation,<ref name="PDICTroll">{{cite web |author=Workgroup |url=http://www.princetondigitalimage.com/ |title=Princeton Digital Image Corporation Home Page |access-date=2013-05-01 |archive-date=2013-04-11 |archive-url=https://archive.today/20130411084058/http://www.princetondigitalimage.com/  }}</ref> based in Eastern Texas, began suing large numbers of companies for alleged infringement of {{US patent|4813056}}. Princeton claims that the JPEG image compression standard infringes the '056 patent and has sued large numbers of websites, retailers, camera and device manufacturers and resellers. The patent was originally owned and assigned to General Electric. The patent expired in December 2007, but Princeton has sued large numbers of companies for "past infringement" of this patent. (Under U.S. patent laws, a patent owner can sue for "past infringement" up to six years before the filing of a lawsuit, so Princeton could theoretically have continued suing companies until December 2013.) As of March 2013, Princeton had suits pending in New York and Delaware against more than 55 companies. General Electric's involvement in the suit is unknown, although court records indicate that it assigned the patent to Princeton in 2009 and retains certain rights in the patent.<ref name="PDICTroll2">{{cite web|author=Workgroup |url=http://patentlaw.jmbm.com/2013/04/hps-motion-to-dismiss-for-lack.html/ |archive-url=https://web.archive.org/web/20160309142416/http://patentlaw.jmbm.com/2013/04/hps-motion-to-dismiss-for-lack.html  |archive-date=2016-03-09 |title=Article on Princeton Court Ruling Regarding GE License Agreement |date=3 April 2013 |access-date=2013-05-01 }}</ref>

==Typical use==
The JPEG compression algorithm operates at its best on photographs and paintings of realistic scenes with smooth variations of tone and color. For web usage, where reducing the amount of data used for an image is important for responsive presentation, JPEG's compression benefits make JPEG popular. JPEG/[[Exif]] is also the most common format saved by digital cameras.

However, JPEG is not well suited for line drawings and other textual or iconic graphics, where the sharp contrasts between adjacent pixels can cause noticeable artifacts.<ref name=":0">{{Cite journal |last=Kaur |first=Rajandeep |date=May 2016 |title=A Review of Image Compression Techniques |url=https://www.researchgate.net/publication/303325238 |journal=International Journal of Computer Applications |volume=142 |issue=1 |pages=8–11 |doi=10.5120/ijca2016909658|doi-access=free }}</ref> Such images are better saved in a [[Lossless data compression#Raster graphics|lossless graphics format]] such as [[TIFF]], [[Graphics Interchange Format|GIF]], [[Portable Network Graphics|PNG]], or a [[raw image format]]. The JPEG standard includes a lossless coding mode, but that mode is not supported in most products.

As the typical use of JPEG is a [[lossy compression]] method, which reduces the image fidelity, it is inappropriate for exact reproduction of imaging data (such as some scientific and medical imaging applications and certain technical [[image processing]] work).<ref name=":0" />

JPEG is also not well suited to files that will undergo multiple edits, as some image quality is lost each time the image is recompressed, particularly if the image is cropped or shifted, or if encoding parameters are changed&nbsp;– see [[digital generation loss]] for details. To prevent image information loss during sequential and repetitive editing, the first edit can be saved in a lossless format, subsequently edited in that format, then finally published as JPEG for distribution.

==JPEG compression==
JPEG uses a lossy form of compression based on the [[Discrete cosine transform|discrete cosine transform (DCT)]]. This mathematical operation converts each frame/field of the video source from the spatial (2D) domain into the frequency domain (a.k.a. transform domain). A perceptual model based loosely on how the human psychovisual system discards high-frequency information, i.e. sharp transitions in intensity, and color hue. In the transform domain, the process of reducing information is called quantization. In simpler terms, quantization is a method for optimally reducing a large number scale (with different occurrences of each number) into a smaller one, and the transform-domain is a convenient representation of the image because the high-frequency coefficients, which contribute less to the overall picture than other coefficients, are characteristically small-values with high compressibility. The quantized coefficients are then sequenced and losslessly packed into the output bitstream. Nearly all software implementations of JPEG permit user control over the compression ratio (as well as other optional parameters), allowing the user to trade off picture-quality for smaller file size. In embedded applications (such as miniDV, which uses a similar DCT-compression scheme), the parameters are pre-selected and fixed for the application.

The compression method is usually [[Lossy compression|lossy]], meaning that some original image information is lost and cannot be restored, possibly affecting image quality. There is an optional [[Lossless JPEG|lossless]] mode defined in the JPEG standard. However, this mode is not widely supported in products.

There is also an [[interlace (bitmaps)|interlaced]] ''progressive'' JPEG format, in which data is compressed in multiple passes of progressively higher detail. This is ideal for large images that will be displayed while downloading over a slow connection, allowing a reasonable preview after receiving only a portion of the data. However, support for progressive JPEGs is not universal. When progressive JPEGs are received by programs that do not support them (such as versions of [[Internet Explorer for Mac|Internet Explorer]] before [[Windows 7]])<ref>{{cite web
 | url = http://msdn.microsoft.com/en-us/library/ee720036(v=vs.85).aspx
 | title = Progressive Decoding Overview
 | access-date = 2012-03-23
 | work = Microsoft Developer Network
 | publisher = Microsoft
 | archive-date = 2012-11-19
 | archive-url = https://web.archive.org/web/20121119054723/http://msdn.microsoft.com/en-us/library/ee720036(v=vs.85).aspx
 | url-status = live
 }}</ref> the software displays the image only after it has been completely downloaded.

There are also many medical imaging, traffic and camera applications that create and process 12-bit JPEG images both grayscale and color. 12-bit JPEG format is included in an Extended part of the JPEG specification. The libjpeg codec supports 12-bit JPEG and there even exists a high-performance version.<ref>{{Cite web | last = Fastvideo | title = 12-bit JPEG encoder on GPU | date = May 2019 | url = https://www.fastcompression.com/products/12-bit-jpeg-codec.htm | access-date = 2019-05-06 | archive-date = 2019-05-06 | archive-url = https://web.archive.org/web/20190506142226/https://www.fastcompression.com/products/12-bit-jpeg-codec.htm | url-status = live }}</ref>

===Lossless editing===

Several alterations to a JPEG image can be performed losslessly (that is, without recompression and the associated quality loss) as long as the image size is a multiple of 1 MCU block (Minimum Coded Unit) (usually 16 pixels in both directions, for 4:2:0 [[chroma subsampling]]). Utilities that implement this include:
* [[jpegtran]] and its GUI, Jpegcrop.
* [[IrfanView]] using "JPG Lossless Crop (PlugIn)" and "JPG Lossless Rotation (PlugIn)", which require installing the JPG_TRANSFORM plugin.
* [[FastStone Image Viewer]] using "Lossless Crop to File" and "JPEG Lossless Rotate".
* [[XnViewMP]] using "JPEG lossless transformations".
* [[ACDSee]] supports lossless rotation (but not lossless cropping) with its "Force lossless JPEG operations" option.

Blocks can be rotated in 90-degree increments, flipped in the horizontal, vertical and diagonal axes and moved about in the image. Not all blocks from the original image need to be used in the modified one.

The top and left edge of a JPEG image must lie on an 8&nbsp;×&nbsp;8 pixel block boundary (or 16&nbsp;×&nbsp;16 pixel for larger MCU sizes), but the bottom and right edge need not do so. This limits the possible '''lossless crop''' operations, and prevents flips and rotations of an image whose bottom or right edge does not lie on a block boundary for all channels (because the edge would end up on top or left, where&nbsp;– as aforementioned&nbsp;– a block boundary is obligatory).

Rotations where the image is not a multiple of 8 or 16, which value depends upon the chroma subsampling, are not lossless. Rotating such an image causes the blocks to be recomputed which results in loss of quality.<ref>{{cite web|url=http://petapixel.com/2012/08/14/why-you-should-always-rotate-original-jpeg-photos-losslessly/|title=Why You Should Always Rotate Original JPEG Photos Losslessly|date=14 August 2012|website=Petapixel.com|access-date=16 October 2017|archive-date=17 October 2017|archive-url=https://web.archive.org/web/20171017093502/https://petapixel.com/2012/08/14/why-you-should-always-rotate-original-jpeg-photos-losslessly/|url-status=live}}</ref>

When using lossless cropping, if the bottom or right side of the crop region is not on a block boundary, then the rest of the data from the partially used blocks will still be present in the cropped file and can be recovered. It is also possible to transform between baseline and progressive formats without any loss of quality, since the only difference is the order in which the coefficients are placed in the file.

Furthermore, several JPEG images can be losslessly joined, as long as they were saved with the same quality and the edges coincide with block boundaries.

==JPEG files==
<!-- This section is targeted by the redirect [[JPEG Interchange Format]] -->

The [[file format]] known as "JPEG Interchange Format" (JIF) is specified in Annex B of the standard. However, this "pure" file format is rarely used, primarily because of the difficulty of programming encoders and decoders that fully implement all aspects of the standard and because of certain shortcomings of the standard:
* Color space definition
* Component sub-sampling registration
* Pixel aspect ratio definition.

Several additional standards have evolved to address these issues. The first of these, released in 1992, was the [[JPEG File Interchange Format]] (JFIF), followed in recent years by [[Exchangeable image file format]] (Exif) and [[International Color Consortium|ICC]] [[ICC Profile|color profiles]]. Both of these formats use the actual JIF byte layout, consisting of different ''markers'', but in addition, employ one of the JIF standard's extension points, namely the ''application markers'': JFIF uses APP0, while Exif uses APP1. Within these segments of the file that were left for future use in the JIF standard and are not read by it, these standards add specific metadata.

Thus, in some ways, JFIF is a cut-down version of the JIF standard in that it specifies certain constraints (such as not allowing all the different encoding modes), while in other ways, it is an extension of JIF due to the added metadata. The documentation for the original JFIF standard states:<ref>{{cite web
| url = http://www.w3.org/Graphics/JPEG/jfif3.pdf
| title = JFIF File Format as PDF
| access-date = 2006-06-19
| archive-date = 2021-01-13
| archive-url = https://web.archive.org/web/20210113062300/http://www.w3.org/Graphics/JPEG/jfif3.pdf
| url-status = live
}}</ref>

{{Blockquote|JPEG File Interchange Format is a minimal file format which enables JPEG bitstreams to be exchanged between a wide variety of platforms and applications. This minimal format does not include any of the advanced features found in the TIFF JPEG specification or any application specific file format. Nor should it, for the only purpose of this simplified format is to allow the exchange of JPEG compressed images.}}

Image files that employ JPEG compression are commonly called "JPEG files", and are stored in variants of the JIF image format. Most image capture devices (such as digital cameras) that output JPEG are actually creating files in the [[Exif]] format, the format that the camera industry has standardized on for metadata interchange. On the other hand, since the Exif standard does not allow color profiles, most image editing software stores JPEG in [[JFIF]] format, and includes the APP1 segment from the Exif file to include the metadata in an almost-compliant way; the JFIF standard is interpreted somewhat flexibly.<ref>{{cite web|url=http://www.faqs.org/faqs/jpeg-faq/part1/|title=JPEG image compression FAQ|author=Tom Lane|date=1999-03-29|access-date=2007-09-11|archive-date=2010-11-10|archive-url=https://web.archive.org/web/20101110054244/http://www.faqs.org/faqs/jpeg-faq/part1/|url-status=live}} (q. 14: "Why all the argument about file formats?")</ref>

Strictly speaking, the JFIF and Exif standards are incompatible, because each specifies that its marker segment (APP0 or APP1, respectively) appear first. In practice, most JPEG files contain a JFIF marker segment that precedes the Exif header. This allows older readers to correctly handle the older format JFIF segment, while newer readers also decode the following Exif segment, being less strict about requiring it to appear first.

===JPEG filename extensions===
The most common [[filename extension]]s for files employing JPEG compression are '''<code>.jpg</code>''' and '''<code>.jpeg</code>''', though <code>.jpe</code>, <code>.jfif</code> and <code>.jif</code> are also used.<ref>{{Cite web |title=Everything you need to know about JPEG files {{!}} Adobe |url=https://www.adobe.com/creativecloud/file-types/image/raster/jpeg-file.html |access-date=2023-08-18 |website=www.adobe.com |language=en-US}}</ref> It is also possible for JPEG data to be embedded in other file types&nbsp;– [[Tag Image File Format|TIFF]] encoded files often embed a JPEG image as a [[thumbnail]] of the main image; and [[MP3]] files can contain a JPEG of [[cover art]] in the [[ID3v2]] tag.

===Color profile===
Many JPEG files embed an [[ICC profile|ICC color profile]] ([[color space]]). Commonly used color profiles include [[sRGB color space|sRGB]] and [[Adobe RGB color space|Adobe RGB]]. Because these color spaces use a non-linear transformation, the [[dynamic range]] of an 8-bit JPEG file is about 11 [[f-number|stops]]; see [[gamma curve]].

If the image doesn't specify color profile information (''untagged''), the color space is assumed to be sRGB for the purposes of display on webpages.<ref name="auto">{{Cite web|url=https://www.w3.org/Graphics/Color/sRGB.html|title=A Standard Default Color Space for the Internet - sRGB|website=www.w3.org|access-date=2022-02-18|archive-date=2022-02-18|archive-url=https://web.archive.org/web/20220218083217/https://www.w3.org/Graphics/Color/sRGB.html|url-status=live}}</ref><ref name="auto1">{{Cite web|url=https://webstore.iec.ch/publication/6168|title=IEC 61966-2-1:1999/AMD1:2003 &#124; IEC Webstore|website=webstore.iec.ch|access-date=2022-02-18|archive-date=2022-02-18|archive-url=https://web.archive.org/web/20220218083451/https://webstore.iec.ch/publication/6168|url-status=live}}</ref>

==Syntax and structure==
A JPEG image consists of a sequence of ''segments'', each beginning with a ''marker'', each of which begins with a 0xFF byte, followed by a byte indicating what kind of marker it is. Some markers consist of just those two bytes; others are followed by two bytes (high then low), indicating the length of marker-specific payload data that follows. (The length includes the two bytes for the length, but not the two bytes for the marker.) Some markers are followed by [[Entropy encoding|entropy-coded]] data; the length of such a marker does not include the entropy-coded data. Note that consecutive 0xFF bytes are used as fill bytes for [[Data padding|padding]] purposes, although this fill byte padding should only ever take place for markers immediately following entropy-coded scan data (see JPEG specification section B.1.1.2 and E.1.2 for details; specifically "In all cases where markers are appended after the compressed data, optional 0xFF fill bytes may precede the marker").

Within the entropy-coded data, after any 0xFF byte, a 0x00 byte is inserted by the encoder before the next byte, so that there does not appear to be a marker where none is intended, preventing framing errors. Decoders must skip this 0x00 byte. This technique, called [[byte stuffing]] (see JPEG specification section F.1.2.3), is only applied to the entropy-coded data, not to marker payload data. Note however that entropy-coded data has a few markers of its own; specifically the Reset markers (0xD0 through 0xD7), which are used to isolate independent chunks of entropy-coded data to allow parallel decoding, and encoders are free to insert these Reset markers at regular intervals (although not all encoders do this).

{| class="wikitable"
|+ Common JPEG markers<ref>{{cite web | url = http://www.digicamsoft.com/itu/itu-t81-36.html | title = ISO/IEC 10918-1: 1993(E) p.36 | access-date = 2007-11-30 | archive-date = 2011-08-01 | archive-url = https://web.archive.org/web/20110801031604/http://www.digicamsoft.com/itu/itu-t81-36.html | url-status = live }}</ref>
! Short name
! Bytes
! Payload
! Name
! Comments
|-
! SOI
| 0xFF, 0xD8 || ''none'' || Start Of Image
|
|-
! SOF0
| 0xFF, 0xC0 || ''variable size'' || Start Of Frame (baseline [[Discrete cosine transform|DCT]])
| Indicates that this is a baseline DCT-based JPEG, and specifies the width, height, number of components, and component subsampling (e.g., 4:2:0).
|-
! SOF2
| 0xFF, 0xC2 || ''variable size'' || Start Of Frame (progressive DCT)
| Indicates that this is a progressive DCT-based JPEG, and specifies the width, height, number of components, and component subsampling (e.g., 4:2:0).
|-
! DHT
| 0xFF, 0xC4 || ''variable size'' || Define Huffman Table(s)
| Specifies one or more Huffman tables.
|-
! DQT
| 0xFF, 0xDB || ''variable size'' || Define Quantization Table(s)
| Specifies one or more quantization tables.
|-
! DRI
| 0xFF, 0xDD || 4 bytes || Define Restart Interval
| Specifies the interval between RST''n'' markers, in Minimum Coded Units (MCUs). This marker is followed by two bytes indicating the fixed size so it can be treated like any other variable size segment.
|-
! SOS
| 0xFF, 0xDA || ''variable size'' || Start Of Scan
| Begins a top-to-bottom scan of the image. In baseline DCT JPEG images, there is generally a single scan. Progressive DCT JPEG images usually contain multiple scans. This marker specifies which slice of data it will contain, and is immediately followed by entropy-coded data.
|-
! RST''n''
| 0xFF, 0xD''n'' (''n''=0..7) || ''none'' || Restart
| Inserted every ''r'' macroblocks, where ''r'' is the restart interval set by a DRI marker. Not used if there was no DRI marker. The low three bits of the marker code cycle in value from 0 to 7.
|-
! APP''n''
| 0xFF, 0xE''n'' || ''variable size'' || Application-specific
| For example, an [[Exif]] JPEG file uses an APP1 marker to store metadata, laid out in a structure based closely on [[TIFF]].
|-
! COM
| 0xFF, 0xFE || ''variable size'' || Comment
| Contains a text comment.
|-
! EOI
| 0xFF, 0xD9 || ''none'' || End Of Image
|
|}

There are other ''Start Of Frame'' markers that introduce other kinds of JPEG encodings.

Since several vendors might use the same APP''n'' marker type, application-specific markers often begin with a standard or vendor name (e.g., "Exif" or "Adobe") or some other identifying string.

At a restart marker, block-to-block predictor variables are reset, and the bitstream is synchronized to a byte boundary. Restart markers provide means for recovery after bitstream error, such as transmission over an unreliable network or file corruption. Since the runs of macroblocks between restart markers may be independently decoded, these runs may be decoded in parallel.

==JPEG codec example==
Although a JPEG file can be encoded in various ways, most commonly it is done with JFIF encoding. The encoding process consists of several steps:

# The representation of the colors in the image is converted from [[RGB color model|RGB]] to [[YCbCr|{{YCbCr}}]], consisting of one [[Luma (video)|luma]] component (Y'), representing brightness, and two [[chrominance|chroma]] components, (C<sub>B</sub> and C<sub>R</sub>), representing color. This step is sometimes skipped.
# The resolution of the chroma data is reduced, usually by a factor of 2 or 3. This reflects the fact that the eye is less sensitive to fine color details than to fine brightness details.
# The image is split into blocks of 8×8 pixels, and for each block, each of the Y, C<sub>B</sub>, and C<sub>R</sub> data undergoes the [[discrete cosine transform]] (DCT). A DCT is similar to a [[Fourier transform]] in the sense that it produces a kind of spatial frequency spectrum.
# The amplitudes of the frequency components are [[Quantization (image processing)|quantized]]. Human vision is much more sensitive to small variations in color or brightness over large areas than to the strength of high-frequency brightness variations. Therefore, the magnitudes of the high-frequency components are stored with a lower accuracy than the low-frequency components. The quality setting of the encoder (for example 50 or 95 on a scale of 0–100 in the Independent JPEG Group's library<ref>{{cite web | url = http://apodeline.free.fr/DOC/libjpeg/libjpeg-3.html | title = Advanced Features: Compression parameter selection | work = Using the IJG JPEG Library | author = Thomas G. Lane. | access-date = 2008-10-08 | archive-date = 2001-11-26 | archive-url = https://web.archive.org/web/20011126012948/http://apodeline.free.fr/DOC/libjpeg/libjpeg-3.html | url-status = live }}</ref>) affects to what extent the resolution of each frequency component is reduced. If an excessively low quality setting is used, the high-frequency components are discarded altogether.
# The resulting data for all 8×8 blocks is further compressed with a lossless algorithm, a variant of [[Huffman encoding]].
The decoding process reverses these steps, except the ''quantization'' because it is irreversible. In the remainder of this section, the encoding and decoding processes are described in more detail.

===Encoding===
Many of the options in the JPEG standard are not commonly used, and as mentioned above, most image software uses the simpler JFIF format when creating a JPEG file, which among other things specifies the encoding method. Here is a brief description of one of the more common methods of encoding when applied to an input that has 24 [[bits per pixel]] (eight each of [[RGB color model|red, green, and blue]]). This particular option is a [[lossy data compression]] method. They are represented in [[Matrix (mathematics)|matrices]] below.

====Color space transformation====
First, the image should be converted from RGB (by default sRGB,<ref name="auto"/><ref name="auto1"/> but other [[color space]]s are possible) into a different color space called [[YCbCr#JPEG conversion|{{YCbCr}}]] (or, informally, YCbCr). It has three components Y', C<sub>B</sub> and C<sub>R</sub>: the Y' component represents the brightness of a pixel, and the C<sub>B</sub> and C<sub>R</sub> components represent the [[chrominance]] (split into blue and red components). This is basically the same color space as used by [[digital television|digital color television]] as well as digital video including [[DVD Video|video DVDs]]. The {{YCbCr}} color space conversion allows greater compression without a significant effect on perceptual image quality (or greater perceptual image quality for the same compression). The compression is more efficient because the brightness information, which is more important to the eventual perceptual quality of the image, is confined to a single channel. This more closely corresponds to the perception of color in the human visual system. The color transformation also improves compression by statistical [[decorrelation]].

A particular conversion to {{YCbCr}} is specified in the JFIF standard, and should be performed for the resulting JPEG file to have maximum compatibility. However, some JPEG implementations in "highest quality" mode do not apply this step and instead keep the color information in the RGB color model,{{Citation needed|date=February 2008}} where the image is stored in separate channels for red, green and blue brightness components. This results in less efficient compression, and would not likely be used when file size is especially important.

====Downsampling====
Due to the densities of color- and brightness-sensitive receptors in the human eye, humans can see considerably more fine detail in the brightness of an image (the Y' component) than in the hue and color saturation of an image (the Cb and Cr components). Using this knowledge, encoders can be designed to compress images more efficiently.

The transformation into the [[YCbCr|{{YCbCr}} color model]] enables the next usual step, which is to reduce the spatial resolution of the Cb and Cr components (called "[[downsampling]]" or "[[chroma subsampling]]"). The ratios at which the downsampling is ordinarily done for JPEG images are [[YUV 4:4:4|4:4:4]] (no downsampling), [[YUV 4:2:2|4:2:2]] (reduction by a factor of 2 in the horizontal direction), or (most commonly) [[YUV 4:2:0|4:2:0]] (reduction by a factor of 2 in both the horizontal and vertical directions). For the rest of the compression process, Y', Cb and Cr are processed separately and in a very similar manner.

====Block splitting====
After [[Chroma subsampling|subsampling]], each [[Channel (digital image)|channel]] must be split into 8×8 blocks. Depending on chroma subsampling, this yields Minimum Coded Unit (MCU) blocks of size 8×8 (4:4:4&nbsp;– no subsampling), 16×8 (4:2:2), or most commonly 16×16 (4:2:0). In [[video compression]] MCUs are called [[macroblock]]s.

If the data for a channel does not represent an integer number of blocks then the encoder must fill the remaining area of the incomplete blocks with some form of dummy data. Filling the edges with a fixed color (for example, black) can create [[ringing artifact]]s along the visible part of the border;
repeating the edge pixels is a common technique that reduces (but does not necessarily eliminate) such artifacts, and more sophisticated border filling techniques can also be applied.

====Discrete cosine transform====
[[File:JPEG example subimage.svg|thumb|256px|The 8×8 sub-image shown in 8-bit grayscale]]

Next, each 8×8 block of each component (Y, Cb, Cr) is converted to a [[frequency-domain]] representation, using a normalized, two-dimensional type-II discrete cosine transform (DCT), see Citation 1 in discrete cosine transform. The DCT is sometimes referred to as "type-II DCT" in the context of a family of transforms as in [[Discrete cosine transform#DCT-II|discrete cosine transform]], and the corresponding inverse (IDCT) is denoted as "type-III DCT".

As an example, one such 8×8 8-bit subimage might be:

:<math>
\left[
\begin{array}{rrrrrrrr}
 52 & 55 & 61 & 66 & 70 & 61 & 64 & 73 \\
 63 & 59 & 55 & 90 & 109 & 85 & 69 & 72 \\
 62 & 59 & 68 & 113 & 144 & 104 & 66 & 73 \\
 63 & 58 & 71 & 122 & 154 & 106 & 70 & 69 \\
 67 & 61 & 68 & 104 & 126 & 88 & 68 & 70 \\
 79 & 65 & 60 & 70 & 77 & 68 & 58 & 75 \\
 85 & 71 & 64 & 59 & 55 & 61 & 65 & 83 \\
 87 & 79 & 69 & 68 & 65 & 76 & 78 & 94
\end{array}
\right].
</math>

Before computing the DCT of the 8×8 block, its values are shifted from a positive range to one centered on zero. For an 8-bit image, each entry in the original block falls in the range <math>[0, 255]</math>. The midpoint of the range (in this case, the value 128) is subtracted from each entry to produce a data range that is centered on zero, so that the modified range is <math>[-128, 127]</math>. This step reduces the dynamic range requirements in the DCT processing stage that follows.

This step results in the following values:

:<math>g=
\begin{array}{c}
x \\
\longrightarrow \\
\left[
\begin{array}{rrrrrrrr}
 -76 & -73 & -67 & -62 & -58 & -67 & -64 & -55 \\
 -65 & -69 & -73 & -38 & -19 & -43 & -59 & -56 \\
 -66 & -69 & -60 & -15 & 16 & -24 & -62 & -55 \\
 -65 & -70 & -57 & -6 & 26 & -22 & -58 & -59 \\
 -61 & -67 & -60 & -24 & -2 & -40 & -60 & -58 \\
 -49 & -63 & -68 & -58 & -51 & -60 & -70 & -53 \\
 -43 & -57 & -64 & -69 & -73 & -67 & -63 & -45 \\
 -41 & -49 & -59 & -60 & -63 & -52 & -50 & -34
\end{array}
\right]
\end{array}
\Bigg\downarrow y.
</math>

[[File:Dctjpeg.png|thumb|The DCT transforms an 8×8 block of input values to a [[linear combination]] of these 64 patterns. The patterns are referred to as the two-dimensional DCT ''basis functions'', and the output values are referred to as ''transform coefficients''. The horizontal index is <math>u</math> and the vertical index is <math>v</math>.]]

The next step is to take the two-dimensional DCT, which is given by:

:<math>\ G_{u,v} =
 \frac{1}{4}
 \alpha(u)
 \alpha(v)
 \sum_{x=0}^7
 \sum_{y=0}^7
 g_{x,y}
 \cos \left[\frac{(2x+1)u\pi}{16} \right]
 \cos \left[\frac{(2y+1)v\pi}{16} \right]
</math>

where
* <math>\ u</math> is the horizontal [[spatial frequency]], for the integers <math>\ 0 \leq u < 8</math>.
* <math>\ v</math> is the vertical spatial frequency, for the integers <math>\ 0 \leq v < 8</math>.
* <math>\alpha(u)</math> and <math>\alpha(v)</math> are normalizing scale factors to make the transformation [[orthonormal]] with <math>
\alpha(i) =
\begin{cases}
 \frac{1}{\sqrt{2}}, & \mbox{if }i=0 \\
 1, & \mbox{otherwise}
\end{cases}
</math> 
* <math>\ g_{x,y}</math> is the pixel value at coordinates <math>\ (x,y)</math>
* <math>\ G_{u,v}</math> is the DCT coefficient at coordinates <math>\ (u,v).</math>

If we perform this transformation on our matrix above, we get the following (rounded to the nearest two digits beyond the decimal point):

:<math>G=
\begin{array}{c}
u \\
\longrightarrow \\
\left[
\begin{array}{rrrrrrrr}
-415.38 & -30.19 & -61.20 & 27.24 & 56.12 & -20.10 & -2.39 & 0.46 \\
4.47 & -21.86 & -60.76 & 10.25 & 13.15 & -7.09 & -8.54 & 4.88 \\
-46.83 & 7.37 & 77.13 & -24.56 & -28.91 & 9.93 & 5.42 & -5.65 \\
-48.53 & 12.07 & 34.10 & -14.76 & -10.24 & 6.30 & 1.83 & 1.95 \\
12.12 & -6.55 & -13.20 & -3.95 & -1.87 & 1.75 & -2.79 & 3.14 \\
-7.73 & 2.91 & 2.38 & -5.94 & -2.38 & 0.94 & 4.30 & 1.85 \\
-1.03 & 0.18 & 0.42 & -2.42 & -0.88 & -3.02 & 4.12 & -0.66 \\
-0.17 & 0.14 & -1.07 & -4.19 & -1.17 & -0.10 & 0.50 & 1.68
\end{array}
\right]
\end{array}
\Bigg\downarrow v.
</math>

Note the top-left corner entry with the rather large magnitude. This is the [[DC bias|DC]] coefficient (also called the constant component), which defines the basic hue for the entire block. The remaining 63 coefficients are the AC coefficients (also called the alternating components).<ref>{{cite web|url=http://forum.doom9.org/showthread.php?p=184647#post184647|title=DC / AC Frequency Questions - Doom9's Forum|website=forum.doom9.org|access-date=16 October 2017|archive-date=17 October 2017|archive-url=https://web.archive.org/web/20171017042422/http://forum.doom9.org/showthread.php?p=184647#post184647|url-status=live}}</ref> The advantage of the DCT is its tendency to aggregate most of the signal in one corner of the result, as may be seen above. The quantization step to follow accentuates this effect while simultaneously reducing the overall size of the DCT coefficients, resulting in a signal that is easy to compress efficiently in the entropy stage.

The DCT temporarily increases the bit-depth of the data, since the DCT coefficients of an 8-bit/component image take up to 11 or more bits (depending on fidelity of the DCT calculation) to store. This may force the codec to temporarily use 16-bit numbers to hold these coefficients, doubling the size of the image representation at this point; these values are typically reduced back to 8-bit values by the quantization step. The temporary increase in size at this stage is not a performance concern for most JPEG implementations, since typically only a very small part of the image is stored in full DCT form at any given time during the image encoding or decoding process.

====Quantization====
The human eye is good at seeing small differences in [[brightness]] over a relatively large area, but not so good at distinguishing the exact strength of a high frequency brightness variation. This allows one to greatly reduce the amount of information in the high frequency components. This is done by simply dividing each component in the frequency domain by a constant for that component, and then rounding to the nearest integer. This rounding operation is the only lossy operation in the whole process (other than chroma subsampling) if the DCT computation is performed with sufficiently high precision. As a result of this, it is typically the case that many of the higher frequency components are rounded to zero, and many of the rest become small positive or negative numbers, which take many fewer bits to represent.

The elements in the [[quantization matrix]] control the compression ratio, with larger values producing greater compression. A typical quantization matrix (for a quality of 50% as specified in the original JPEG Standard), is as follows:

:<math>Q=
\begin{bmatrix}
 16 & 11 & 10 & 16 & 24 & 40 & 51 & 61 \\
 12 & 12 & 14 & 19 & 26 & 58 & 60 & 55 \\
 14 & 13 & 16 & 24 & 40 & 57 & 69 & 56 \\
 14 & 17 & 22 & 29 & 51 & 87 & 80 & 62 \\
 18 & 22 & 37 & 56 & 68 & 109 & 103 & 77 \\
 24 & 35 & 55 & 64 & 81 & 104 & 113 & 92 \\
 49 & 64 & 78 & 87 & 103 & 121 & 120 & 101 \\
 72 & 92 & 95 & 98 & 112 & 100 & 103 & 99
\end{bmatrix}.
</math>

The quantized DCT coefficients are computed with

:<math>B_{j,k} = \mathrm{round} \left( \frac{G_{j,k}}{Q_{j,k}} \right) \mbox{ for } j=0,1,2,\ldots,7; k=0,1,2,\ldots,7</math>

where <math>G</math> is the unquantized DCT coefficients; <math>Q</math> is the quantization matrix above; and <math>B</math> is the quantized DCT coefficients.

Using this quantization matrix with the DCT coefficient matrix from above results in:
[[File:idct-animation.gif|frame|Left: a final image is built up from a series of basis functions. Right: each of the DCT basis functions that comprise the image, and the corresponding weighting coefficient. Middle: the basis function, after multiplication by the coefficient: this component is added to the final image. For clarity, the 8×8 macroblock in this example is magnified by 10x using bilinear interpolation.]]
:<math>B=
\left[
\begin{array}{rrrrrrrr}
-26 & -3 & -6 & 2 & 2 & -1 & 0 & 0 \\
0 & -2 & -4 & 1 & 1 & 0 & 0 & 0 \\
-3 & 1 & 5 & -1 & -1 & 0 & 0 & 0 \\
-3 & 1 & 2 & -1 & 0 & 0 & 0 & 0 \\
1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0
\end{array}
\right].
</math>

For example, using −415 (the DC coefficient) and rounding to the nearest integer

:<math>
\mathrm{round}
\left(
 \frac{-415.37}{16}
\right)
=
\mathrm{round}
\left(
 -25.96
\right)
=
-26.
</math>

Notice that most of the higher-frequency elements of the sub-block (i.e., those with an ''x'' or ''y'' spatial frequency greater than 4) are quantized into zero values.

====Entropy coding====
{{Main|Entropy encoding}}
[[File:JPEG ZigZag.svg|thumb|right|Zigzag ordering of JPEG image components]]

Entropy coding is a special form of [[lossless data compression]]. It involves arranging the image components in a "[[zigzag]]" order employing [[run-length encoding]] (RLE) algorithm that groups similar frequencies together, inserting length coding zeros, and then using [[Huffman coding]] on what is left.

The JPEG standard also allows, but does not require, decoders to support the use of [[arithmetic coding]], which is mathematically superior to Huffman coding. However, this feature has rarely been used, as it was historically covered by [[patent]]s requiring royalty-bearing licenses, and because it is slower to encode and decode compared to Huffman coding. Arithmetic coding typically makes files about 5–7% smaller.<ref>{{Cite journal |last1=Maini |first1=Raman |last2=Mehra |first2=Suruchi |date=December 2010 |title=A Review on JPEG2000 Image Compression |url=https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=bcdc5083bcf5c3120e7450273bd380f61071afcb |journal=[[International Journal of Computer Applications]] |volume=11 |issue=9 |pages=43–47 |doi=10.5120/1607-2159 |via=CiteSeerX}}</ref>

The previous quantized DC coefficient is used to predict the current quantized DC coefficient. The difference between the two is encoded rather than the actual value. The encoding of the 63 quantized AC coefficients does not use such prediction differencing.

The zigzag sequence for the above quantized coefficients are shown below. (The format shown is just for ease of understanding/viewing.)

:{| style="text-align: right"
|-
|style="width: 2em"| −26 || style="width: 2em"| || style="width: 2em"| || style="width: 2em"| || style="width: 2em"| || style="width: 2em"| || style="width: 2em"| || style="width: 2em"|
|-
| −3 || 0
|-
| −3 || −2 || −6
|-
| 2 || −4 || 1 || −3
|-
| 1 || 1 || 5 || 1 || 2
|-
| −1 || 1 || −1 || 2 || 0 || 0
|-
| 0 || 0 || 0 || −1 || −1 || 0 || 0
|-
| 0 || 0 || 0 || 0 || 0 || 0 || 0 || 0
|-
| 0 || 0 || 0 || 0 || 0 || 0 || 0
|-
| 0 || 0 || 0 || 0 || 0 || 0
|-
| 0 || 0 || 0 || 0 || 0
|-
| 0 || 0 || 0 || 0
|-
| 0 || 0 || 0
|-
| 0 || 0
|-
| 0
|}

If the ''i''-th block is represented by <math>B_i</math> and positions within each block are represented by <math>(p,q)</math> where <math>p = 0, 1, ..., 7</math> and <math>q = 0, 1, ..., 7</math>, then any coefficient in the DCT image can be represented as <math>B_i (p,q)</math>. Thus, in the above scheme, the order of encoding pixels (for the {{mvar|i}}-th block) is <math>B_i (0,0)</math>, <math>B_i (0,1)</math>, <math>B_i (1,0)</math>, <math>B_i (2,0)</math>, <math>B_i (1,1)</math>, <math>B_i (0,2)</math>, <math>B_i (0,3)</math>, <math>B_i (1,2)</math> and so on.

[[File:JPEG process.svg|thumb|upright=1.35|''Baseline sequential'' JPEG encoding and decoding processes]]

This encoding mode is called baseline ''sequential'' encoding. Baseline JPEG also supports ''progressive'' encoding. While sequential encoding encodes coefficients of a single block at a time (in a zigzag manner), progressive encoding encodes similar-positioned batch of coefficients of all blocks in one go (called a ''scan''), followed by the next batch of coefficients of all blocks, and so on. For example, if the image is divided into N 8×8 blocks <math>B_0,B_1,B_2,...,B_{n-1}</math>, then a 3-scan progressive encoding encodes DC component, <math>B_i (0,0)</math> for all blocks, i.e., for all <math>i = 0, 1, 2, ..., N-1</math>, in first scan. This is followed by the second scan which encoding a few more components (assuming four more components, they are <math>B_i (0,1)</math> to <math>B_i (1,1)</math>, still in a zigzag manner) coefficients of all blocks (so the sequence is: <math>B_0 (0,1),B_0 (1,0),B_0 (2,0),B_0 (1,1),B_1 (0,1),B_1 (1,0),...,B_N (2,0),B_N (1,1)</math>), followed by all the remained coefficients of all blocks in the last scan.

Once all similar-positioned coefficients have been encoded, the next position to be encoded is the one occurring next in the zigzag traversal as indicated in the figure above. It has been found that ''baseline progressive'' JPEG encoding usually gives better compression as compared to ''baseline sequential'' JPEG due to the ability to use different Huffman tables (see below) tailored for different frequencies on each "scan" or "pass" (which includes similar-positioned coefficients), though the difference is not too large.

In the rest of the article, it is assumed that the coefficient pattern generated is due to sequential mode.

In order to encode the above generated coefficient pattern, JPEG uses Huffman encoding. The JPEG standard provides general-purpose Huffman tables; encoders may also choose to generate Huffman tables optimized for the actual frequency distributions in images being encoded.

The process of encoding the zig-zag quantized data begins with a run-length encoding explained below, where:
* {{mvar|x}} is the non-zero, quantized AC coefficient.
* ''RUNLENGTH'' is the number of zeroes that came before this non-zero AC coefficient.
* ''SIZE'' is the number of bits required to represent {{mvar|x}}.
* ''AMPLITUDE'' is the bit-representation of {{mvar|x}}.

The run-length encoding works by examining each non-zero AC coefficient {{mvar|x}} and determining how many zeroes came before the previous AC coefficient. With this information, two symbols are created:

:{| style="text-align: center" class="wikitable"
|-
! Symbol 1 || Symbol 2
|-
| (RUNLENGTH, SIZE) || (AMPLITUDE)
|}

Both ''RUNLENGTH'' and ''SIZE'' rest on the same byte, meaning that each only contains four bits of information. The higher bits deal with the number of zeroes, while the lower bits denote the number of bits necessary to encode the value of {{mvar|x}}.

This has the immediate implication of ''Symbol 1'' being only able store information regarding the first 15 zeroes preceding the non-zero AC coefficient. However, JPEG defines two special Huffman code words. One is for ending the sequence prematurely when the remaining coefficients are zero (called "End-of-Block" or "EOB"), and another when the run of zeroes goes beyond 15 before reaching a non-zero AC coefficient. In such a case where 16 zeroes are encountered before a given non-zero AC coefficient, ''Symbol 1'' is encoded "specially" as: (15, 0)(0).

The overall process continues until "EOB"{{snd}} denoted by (0, 0){{snd}} is reached.

With this in mind, the sequence from earlier becomes:

:(0, 2)(-3);(1, 2)(-3);(0, 2)(-2);(0, 3)(-6);(0, 2)(2);(0, 3)(-4);(0, 1)(1);(0, 2)(-3);(0, 1)(1);(0, 1)(1);
:(0, 3)(5);(0, 1)(1);(0, 2)(2);(0, 1)(-1);(0, 1)(1);(0, 1)(-1);(0, 2)(2);(5, 1)(-1);(0, 1)(-1);(0, 0);

(The first value in the matrix, −26, is the DC coefficient; it is not encoded the same way. See above.)

From here, frequency calculations are made based on occurrences of the coefficients. In our example block, most of the quantized coefficients are small numbers that are not preceded immediately by a zero coefficient. These more-frequent cases will be represented by shorter code words.

===Compression ratio and artifacts===
[[File:Lichtenstein jpeg difference.png|256px|thumb|This image shows the pixels that are different between a non-compressed image and the same image JPEG compressed with a quality setting of 50. Darker means a larger difference. Note especially the changes occurring near sharp edges and having a block-like shape.]]
[[File:Lichtenstein img processing test.png|256px|thumb|The original image]]
[[File:Jpegvergroessert.jpg|thumb|192px|The compressed 8×8 squares are visible in the scaled-up picture, together with other visual artifacts of the [[lossy compression]].]]

The resulting compression ratio can be varied according to need by being more or less aggressive in the divisors used in the quantization phase. Ten to one compression usually results in an image that cannot be distinguished by eye from the original. A compression ratio of 100:1 is usually possible, but will look distinctly artifacted compared to the original. The appropriate level of compression depends on the use to which the image will be put.

{{external media
|image1=[http://i.cmpnet.com/videsignline/2006/02/algolith-fig2.jpg Illustration of edge busyness]<ref name="dinhpatry" />
}}

Those who use the [[World Wide Web]] may be familiar with the irregularities known as [[compression artifact]]s that appear in JPEG images, which may take the form of noise around contrasting edges (especially curves and corners), or "blocky" images. These are due to the quantization step of the JPEG algorithm. They are especially noticeable around sharp corners between contrasting colors (text is a good example, as it contains many such corners). The analogous artifacts in [[MPEG]] video are referred to as ''[[mosquito noise]],'' as the resulting "edge busyness" and spurious dots, which change over time, resemble mosquitoes swarming around the object.<ref name="dinhpatry">Phuc-Tue Le Dinh and Jacques Patry. [http://www.videsignline.com/howto/180207350 Video compression artifacts and MPEG noise reduction] {{Webarchive|url=https://web.archive.org/web/20060314201405/http://www.videsignline.com/howto/180207350 |date=2006-03-14 }}. Video Imaging DesignLine. 24 February 2006. Retrieved 28 May 2009.</ref><ref>"'''3.9 mosquito noise:''' Form of edge busyness distortion sometimes associated with movement, characterized by moving artifacts and/or blotchy noise patterns superimposed over the objects (resembling a mosquito flying around a person's head and shoulders)."
[http://eu.sabotage.org/www/ITU/P/P0930e.pdf ITU-T Rec. P.930 (08/96) Principles of a reference impairment system for video] {{webarchive|url=https://web.archive.org/web/20100216033245/http://eu.sabotage.org/www/ITU/P/P0930e.pdf |date=2010-02-16 }}</ref>

These artifacts can be reduced by choosing a lower level of [[image compression|compression]]; they may be completely avoided by saving an image using a lossless file format, though this will result in a larger file size. The images created with [[Ray tracing (graphics)|ray-tracing]] programs have noticeable blocky shapes on the terrain. Certain low-intensity compression artifacts might be acceptable when simply viewing the images, but can be emphasized if the image is subsequently processed, usually resulting in unacceptable quality. Consider the example below, demonstrating the effect of lossy compression on an [[edge detection]] processing step.

{| class="wikitable" style="text-align:center"
|-
! Image !! Lossless compression !! Lossy compression
|-
! Original
| [[File:Lossless-circle.png|100px]] || [[File:Lossy-circle.jpg|100px]]
|-
! Processed by<br />[[Canny edge detector]]
| [[File:Lossless-circle-canny.png|100px]] || [[File:Lossy-circle-canny.png|100px]]
|}

Some programs allow the user to vary the amount by which individual blocks are compressed. Stronger compression is applied to areas of the image that show fewer artifacts. This way it is possible to manually reduce JPEG file size with less loss of quality.

Since the quantization stage ''always'' results in a loss of information, JPEG standard is always a lossy compression codec. (Information is lost both in quantizing and rounding of the floating-point numbers.) Even if the quantization matrix is a [[matrix of ones]], information will still be lost in the rounding step.
{{clear}}

===Decoding===
Decoding to display the image consists of doing all the above in reverse.

Taking the DCT coefficient matrix (after adding the difference of the DC coefficient back in)

:<math>
\left[
\begin{array}{rrrrrrrr}
-26 & -3 & -6 & 2 & 2 & -1 & 0 & 0 \\
0 & -2 & -4 & 1 & 1 & 0 & 0 & 0 \\
-3 & 1 & 5 & -1 & -1 & 0 & 0 & 0 \\
-3 & 1 & 2 & -1 & 0 & 0 & 0 & 0 \\
1 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0
\end{array}
\right]
</math>

and taking the [[Hadamard product (matrices)|entry-for-entry product]] with the quantization matrix from above results in

:<math>
\left[
\begin{array}{rrrrrrrr}
-416 & -33 & -60 & 32 & 48 & -40 & 0 & 0 \\
0 & -24 & -56 & 19 & 26 & 0 & 0 & 0 \\
-42 & 13 & 80 & -24 & -40 & 0 & 0 & 0 \\
-42 & 17 & 44 & -29 & 0 & 0 & 0 & 0 \\
18 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0 \\
0 & 0 & 0 & 0 & 0 & 0 & 0 & 0
\end{array}
\right]
</math>

which closely resembles the original DCT coefficient matrix for the top-left portion.

The next step is to take the two-dimensional inverse DCT (a 2D type-III DCT), which is given by:

<math>f_{x,y} =
 \frac{1}{4}
 \sum_{u=0}^7
 \sum_{v=0}^7
 \alpha(u) \alpha(v) F_{u,v}
 \cos \left[\frac{(2x+1)u\pi}{16} \right]
 \cos \left[\frac{(2y+1)v\pi}{16} \right]
</math>

where
* <math>\ x</math> is the pixel row, for the integers <math>\ 0 \leq x < 8</math>.
* <math>\ y</math> is the pixel column, for the integers <math>\ 0 \leq y < 8</math>.
* <math>\ \alpha(u)</math> is the normalizing scale factor defined earlier, for the integers <math>\ 0 \leq u < 8</math>.
* <math>\ F_{u,v}</math> is the approximated DCT coefficient at coordinates <math>\ (u,v).</math>
* <math>\ f_{x,y}</math> is the reconstructed pixel value at coordinates <math>\ (x,y)</math>

Rounding the output to integer values (since the original had integer values) results in an image with values (still shifted down by 128)

{{Multiple image|align=right|direction=vertical|image1=JPEG example image.jpg|alt1=|image2=JPEG example image decompressed.jpg|alt2=|width=180|footer=Slight differences are noticeable between the original (top) and decompressed image (bottom), which is most readily seen in the bottom-left corner.}}

:<math>
\left[
\begin{array}{rrrrrrrr}
-66 & -63 & -71 & -68 & -56 & -65 & -68 & -46 \\
-71 & -73 & -72 & -46 & -20 & -41 & -66 & -57 \\
-70 & -78 & -68 & -17 & 20 & -14 & -61 & -63 \\
-63 & -73 & -62 & -8 & 27 & -14 & -60 & -58 \\
-58 & -65 & -61 & -27 & -6 & -40 & -68 & -50 \\
-57 & -57 & -64 & -58 & -48 & -66 & -72 & -47 \\
-53 & -46 & -61 & -74 & -65 & -63 & -62 & -45 \\
-47 & -34 & -53 & -74 & -60 & -47 & -47 & -41
\end{array}
\right]
</math>

and adding 128 to each entry

:<math>
\left[
\begin{array}{rrrrrrrr}
62 & 65 & 57 & 60 & 72 & 63 & 60 & 82 \\
57 & 55 & 56 & 82 & 108 & 87 & 62 & 71 \\
58 & 50 & 60 & 111 & 148 & 114 & 67 & 65 \\
65 & 55 & 66 & 120 & 155 & 114 & 68 & 70 \\
70 & 63 & 67 & 101 & 122 & 88 & 60 & 78 \\
71 & 71 & 64 & 70 & 80 & 62 & 56 & 81 \\
75 & 82 & 67 & 54 & 63 & 65 & 66 & 83 \\
81 & 94 & 75 & 54 & 68 & 81 & 81 & 87
\end{array}
\right].
</math>

This is the decompressed subimage. In general, the decompression process may produce values outside the original input range of <math>[0, 255]</math>. If this occurs, the decoder needs to clip the output values so as to keep them within that range to prevent overflow when storing the decompressed image with the original bit depth.

The decompressed subimage can be compared to the original subimage (also see images to the right) by taking the difference (original − uncompressed) results in the following error values:

:<math>
\left[
\begin{array}{rrrrrrrr}
-10 & -10 & 4 & 6 & -2 & -2 & 4 & -9 \\
6 & 4 & -1 & 8 & 1 & -2 & 7 & 1 \\
4 & 9 & 8 & 2 & -4 & -10 & -1 & 8 \\
-2 & 3 & 5 & 2 & -1 & -8 & 2 & -1 \\
-3 & -2 & 1 & 3 & 4 & 0 & 8 & -8 \\
8 & -6 & -4 & -0 & -3 & 6 & 2 & -6 \\
10 & -11 & -3 & 5 & -8 & -4 & -1 & -0 \\
6 & -15 & -6 & 14 & -3 & -5 & -3 & 7
\end{array}
\right]
</math>

with an average absolute error of about 5 values per pixels (i.e., <math>\frac{1}{64} \sum_{x=0}^7 \sum_{y=0}^7 |e(x,y)| = 4.8750</math>).

The error is most noticeable in the bottom-left corner where the bottom-left pixel becomes darker than the pixel to its immediate right.

===Required precision===
The required implementation precision of a JPEG codec is implicitly defined through the requirements formulated for compliance to the JPEG standard. These requirements are specified in ITU.T Recommendation T.83 | ISO/IEC 10918-2. Unlike MPEG standards and many later JPEG standards, the above document defines both required implementation precisions for the encoding and the decoding process of a JPEG codec by means of a maximal tolerable error of the forwards and inverse DCT in the DCT domain as determined by reference test streams. For example, the output of a decoder implementation must not exceed an error of one quantization unit in the DCT domain when applied to the reference testing codestreams provided as part of the above standard. While unusual, and unlike many other and more modern standards, ITU.T T.83 | ISO/IEC 10918-2 does not formulate error bounds in the image domain.

==Effects of JPEG compression==
JPEG compression artifacts blend well into photographs with detailed non-uniform textures, allowing higher compression ratios. Notice how a higher compression ratio first affects the high-frequency textures in the upper-left corner of the image, and how the contrasting lines become more fuzzy. The very high compression ratio severely affects the quality of the image, although the overall colors and image form are still recognizable. However, the precision of colors suffer less (for a human eye) than the precision of contours (based on luminance). This justifies the fact that images should be first transformed in a color model separating the luminance from the chromatic information, before subsampling the chromatic planes (which may also use lower quality quantization) in order to preserve the precision of the luminance plane with more information bits.

===Sample photographs===

[[File:Visual impact of a jpeg compression on Photoshop.jpg|thumb|Visual impact of a jpeg compression on Photoshop on a picture of 4480x4480 pixels]]

For information, the uncompressed 24-bit RGB bitmap image below (73,242 pixels) would require 219,726 bytes (excluding all other information headers). The filesizes indicated below include the internal JPEG information headers and some [[metadata]]. For highest quality images (Q=100), about 8.25 bits per color pixel is required. On grayscale images, a minimum of 6.5 bits per pixel is enough (a comparable Q=100 quality color information requires about 25% more encoded bits). The highest quality image below (Q=100) is encoded at nine bits per color pixel, the medium quality image (Q=25) uses one bit per color pixel. For most applications, the quality factor should not go below 0.75 bit per pixel (Q=12.5), as demonstrated by the low quality image. The image at lowest quality uses only 0.13 bit per pixel, and displays very poor color. This is useful when the image will be displayed in a significantly scaled-down size. A method for creating better quantization matrices for a given image quality using [[PSNR]] instead of the Q factor is described in Minguillón & Pujol (2001).<ref>{{cite journal|author=Julià Minguillón, Jaume Pujol|title=JPEG standard uniform quantization error modeling with applications to sequential and progressive operation modes|journal=Electronic Imaging|date=April 2001|volume=10|number=2|pages=475–485|hdl=10609/6263|bibcode=2001JEI....10..475M|doi=10.1117/1.1344592|s2cid=16629522 |url=http://openaccess.uoc.edu/webapps/o2/bitstream/10609/6263/6/jei-jpeg.pdf|access-date=2019-09-23|archive-date=2020-08-03|archive-url=https://web.archive.org/web/20200803042010/http://openaccess.uoc.edu/webapps/o2/bitstream/10609/6263/6/jei-jpeg.pdf|url-status=live}}</ref>

::{| class="wikitable"
|+ align="bottom"| Note: The above images are not [[Institute of Electrical and Electronics Engineers|IEEE]] / [[ITU-R|CCIR]] / [[European Broadcasting Union|EBU]]&nbsp;[[standard test image|test images]], and the encoder settings are not specified or available.
|-
! Image !! Quality !! Size (bytes) !! Compression ratio !! Comment
|-
| [[File:JPEG example JPG RIP 100.jpg|120px]]
| Highest quality (Q&nbsp;=&nbsp;100)
| 81,447
| 2.7:1
| Extremely minor artifacts
|-
| [[File:JPEG example JPG RIP 050.jpg|120px]]
| High quality (Q&nbsp;=&nbsp;50)
| 14,679
| 15:1
| Initial signs of subimage artifacts
|-
| [[File:JPEG example JPG RIP 025.jpg|120px]]
| Medium quality (Q&nbsp;=&nbsp;25)
| 9,407
| 23:1
| Stronger artifacts; loss of high frequency information
|-
| [[File:JPEG example JPG RIP 010.jpg|120px]]
| Low quality (Q&nbsp;=&nbsp;10)
| 4,787
| 46:1
| Severe high frequency loss leads to obvious artifacts on subimage boundaries ("macroblocking")
|-
| [[File:JPEG example JPG RIP 001.jpg|120px]]
| Lowest quality (Q&nbsp;=&nbsp;1)
| 1,523
| 144:1
| Extreme loss of color and detail; the leaves are nearly unrecognizable.
|}

The medium quality photo uses only 4.3% of the storage space required for the uncompressed image, but has little noticeable loss of detail or visible artifacts. However, once a certain threshold of compression is passed, compressed images show increasingly visible defects. See the article on [[rate–distortion theory]] for a mathematical explanation of this threshold effect. A particular limitation of JPEG in this regard is its non-overlapped 8×8 block transform structure. More modern designs such as [[JPEG 2000]] and [[JPEG XR]] exhibit a more graceful degradation of quality as the bit usage decreases&nbsp;– by using transforms with a larger spatial extent for the lower frequency coefficients and by using overlapping transform basis functions.

==Lossless further compression==
From 2004 to 2008, new research emerged on ways to further compress the data contained in JPEG images without modifying the represented image.<ref name="Bauermann">I. Bauermann and E. Steinbacj. Further Lossless Compression of JPEG Images. Proc. of Picture Coding Symposium (PCS 2004), San Francisco, US, 15–17 December 2004.</ref><ref name="Ponomarenko">N. Ponomarenko, K. Egiazarian, V. Lukin and J. Astola. Additional Lossless Compression of JPEG Images, Proc. of the 4th Intl. Symposium on Image and Signal Processing and Analysis (ISPA 2005), Zagreb, Croatia, pp. 117–120, 15–17 September 2005.</ref><ref name="stirner">M. Stirner and G. Seelmann. Improved Redundancy Reduction for JPEG Files. Proc. of Picture Coding Symposium (PCS 2007), Lisbon, Portugal, 7–9 November 2007</ref><ref name="matsuda">Ichiro Matsuda, Yukio Nomoto, Kei Wakabayashi and Susumu Itoh. Lossless Re-encoding of JPEG images using block-adaptive intra prediction. Proceedings of the 16th European Signal Processing Conference (EUSIPCO 2008).</ref> This has applications in scenarios where the original image is only available in JPEG format, and its size needs to be reduced for archiving or transmission. Standard general-purpose compression tools cannot significantly compress JPEG files.

Typically, such schemes take advantage of improvements to the naive scheme for coding DCT coefficients, which fails to take into account:
* Correlations between magnitudes of adjacent coefficients in the same block;
* Correlations between magnitudes of the same coefficient in adjacent blocks;
* Correlations between magnitudes of the same coefficient/block in different channels;
* The DC coefficients when taken together resemble a downscale version of the original image multiplied by a scaling factor. Well-known schemes for lossless coding of continuous-tone images can be applied, achieving somewhat better compression than the [[Huffman code]]d [[DPCM]] used in JPEG.

Some standard but rarely used options already exist in JPEG to improve the efficiency of coding DCT coefficients: the arithmetic coding option, and the progressive coding option (which produces lower bitrates because values for each coefficient are coded independently, and each coefficient has a significantly different distribution). Modern methods have improved on these techniques by reordering coefficients to group coefficients of larger magnitude together;<ref name="Bauermann" /> using adjacent coefficients and blocks to predict new coefficient values;<ref name="stirner" /> dividing blocks or coefficients up among a small number of independently coded models based on their statistics and adjacent values;<ref name="Ponomarenko" /><ref name="stirner" /> and most recently, by decoding blocks, predicting subsequent blocks in the spatial domain, and then encoding these to generate predictions for DCT coefficients.<ref name="matsuda" />

Typically, such methods can compress existing JPEG files between 15 and 25 percent, and for JPEGs compressed at low-quality settings, can produce improvements of up to 65%.<ref name="stirner" /><ref name="matsuda" />

A freely available tool called packJPG is based on the 2007 paper "Improved Redundancy Reduction for JPEG Files." As of version 2.5k of 2016, it reports a typical 20% reduction by transcoding.<ref>{{cite web |last1=Stirner |first1=Matthias |title=packjpg/packJPG |website=[[GitHub]] |url=https://github.com/packjpg/packJPG |date=19 February 2023 |access-date=2 March 2023 |archive-date=2 March 2023 |archive-url=https://web.archive.org/web/20230302013300/https://github.com/packjpg/packJPG |url-status=live }}</ref> [[JPEG XL]] (ISO/IEC 18181) of 2018 reports a similar reduction in its transcoding.

==Derived formats for stereoscopic 3D==

===JPEG Stereoscopic===
[[File:JPS-sample.jpg|thumb|An example of a stereoscopic .JPS file]]
JPS is a stereoscopic JPEG image used for creating 3D effects from 2D images. It contains two static images, one for the left eye and one for the right eye; encoded as two side-by-side images in a single JPEG file.
JPEG Stereoscopic (JPS, extension .jps) is a JPEG-based format for [[stereoscopy|stereoscopic]] images.<ref>{{cite web |author=J. Siragusa |author2=D. C. Swift |url=http://vrex.com/developer/sterdesc.pdf |title=General Purpose Stereoscopic Data Descriptor |work=VRex, Inc., Elmsford, New York, US |date=1997 |archive-url=https://web.archive.org/web/20111030182549/http://vrex.com/developer/sterdesc.pdf |archive-date=2011-10-30}}</ref><ref>Tim Kemp, [http://ephehm.com/jps/ JPS files] {{Webarchive|url=https://web.archive.org/web/20090118200346/http://ephehm.com/jps/ |date=2009-01-18 }}</ref> It has a range of configurations stored in the JPEG APP3 marker field, but usually contains one image of double width, representing two images of identical size in cross-eyed (i.e. left frame on the right half of the image and vice versa) side-by-side arrangement. This file format can be viewed as a JPEG without any special software, or can be processed for rendering in other modes.

===JPEG Multi-Picture Format===
<!-- This section is linked from [[MPO]] -->
{{infobox file format
|name = JPEG Multi-Picture
|extension = {{mono|.mpo}}
|mime = hide
|uniform type = public.mpo-image<ref>{{cite web |url=https://www.mbsplugins.eu/CGImageSourceSupportedTypes.shtml |title=CGImageSource.SupportedTypes |work=Claris FileMaker MBS Plug-in |publisher=MonkeyBread Software |access-date=2023-05-21 |archive-date=2020-12-30 |archive-url=https://web.archive.org/web/20201230132850/https://mbsplugins.eu/CGImageSourceSupportedTypes.shtml |url-status=live }}</ref>
}}
JPEG Multi-Picture Format (MPO, extension .mpo) is a JPEG-based format for storing multiple images in a single file. It contains two or more JPEG files concatenated together.<ref>{{cite web |url=http://www.cipa.jp/std/documents/e/DC-007_E.pdf |title=Multi-Picture Format |date=2009 |access-date=2015-12-30 |archive-date=2016-04-05 |archive-url=https://web.archive.org/web/20160405200235/http://cipa.jp/std/documents/e/DC-007_E.pdf  }}</ref><ref>{{citation|url=http://www.mtbs3d.com/phpbb/viewtopic.php?f=3&t=4124&start=15|title=MPO2Stereo: Convert Fujifilm MPO files to JPEG stereo pairs|website=Mtbs3d.com|access-date=12 January 2010|archive-date=31 May 2010|archive-url=https://web.archive.org/web/20100531020644/http://www.mtbs3d.com/phpbb/viewtopic.php?f=3&t=4124&start=15|url-status=live}}</ref> It also defines a JPEG APP2 marker segment for image description. Various devices use it to store 3D images, such as [[Fujifilm FinePix Real 3D W1]], [[HTC Evo 3D]], JVC GY-HMZ1U AVCHD/MVC extension camcorder, [[Nintendo 3DS]], [[Panasonic Lumix DMC-TZ20]], [[Panasonic Lumix DMC-TZ30|DMC-TZ30]], [[Panasonic Lumix DMC-TZ60|DMC-TZ60]], DMC-TS4 (FT4), and [[Sony]] DSC-HX7V. Other devices use it to store "preview images" that can be displayed on a TV.

In the last few years, due to the growing use of stereoscopic images, much effort has been spent by the scientific community to develop algorithms for stereoscopic image compression.<ref>{{citation|url=http://iplab.dmi.unict.it/publication/503|title=A new fast matching method for adaptive compression of stereoscopic images|journal=Three-Dimensional Image Processing|volume=9393|pages=93930K|author1=Alessandro Ortis|author2=Sebastiano Battiato|editor1-first=Robert|editor1-last=Sitnik|editor2-first=William|editor2-last=Puech|publisher=SPIE - Three-Dimensional Image Processing, Measurement (3DIPM), and Applications 2015|access-date=30 April 2015|bibcode=2015SPIE.9393E..0KO|year=2015|doi=10.1117/12.2086372|series=Three-Dimensional Image Processing, Measurement (3DIPM), and Applications 2015|s2cid=18879942|archive-date=3 March 2016|archive-url=https://web.archive.org/web/20160303232457/http://iplab.dmi.unict.it/publication/503|url-status=live}}</ref><ref>{{citation |url=http://iplab.dmi.unict.it/publication/49 |title=Adaptive Compression of Stereoscopic Images |author1=Alessandro Ortis |author2=Francesco Rundo |author3=Giuseppe Di Giore |author4=Sebastiano Battiato |publisher=International Conference on Image Analysis and Processing (ICIAP) 2013 |access-date=30 April 2015 |archive-date=3 March 2016 |archive-url=https://web.archive.org/web/20160303205755/http://iplab.dmi.unict.it/publication/49 |url-status=live }}</ref>

==Implementations==
A very important implementation of a JPEG codec is the free programming library '''[[libjpeg]]''' of the Independent JPEG Group. It was first published in 1991 and was key for the success of the standard. This library was used in countless applications.<ref name="JPEG-homepage" /> The development went quiet in 1998; when libjpeg resurfaced with the 2009 version 7, it broke [[Application binary interface|ABI compatibility]] with previous versions. Version 8 of 2010 introduced non-standard extensions, a decision criticized by the original IJG leader Tom Lane.<ref name=TomSmartScale>Tom Lane, 16 January 2013: [https://lists.fedoraproject.org/pipermail/devel/2013-January/176400.html jpeg-9, API/ABI compatibility, and the future role of this project] {{Webarchive|url=https://web.archive.org/web/20181204231539/https://lists.fedoraproject.org/pipermail/devel/2013-January/176400.html |date=2018-12-04 }}</ref>

'''libjpeg-turbo''', forked from the 1998 libjpeg 6b, improves on libjpeg with [[SIMD]] optimizations. Originally seen as a maintained fork of libjpeg, it has become more popular after the incompatible changes of 2009.<ref name="turbo-software">[http://libjpeg-turbo.virtualgl.org/About/Software Software That Uses or Provides libjpeg-turbo] {{Webarchive|url=https://web.archive.org/web/20170318000736/http://libjpeg-turbo.virtualgl.org/About/Software |date=2017-03-18 }}. 9 February 2012.</ref><ref name="Chromium">[http://code.google.com/p/chromium/issues/detail?id=48789#c19 Issue 48789 – chromium – Use libjpeg-turbo instead of libjpeg] {{Webarchive|url=https://web.archive.org/web/20150801035040/http://code.google.com/p/chromium/issues/detail?id=48789#c19 |date=2015-08-01 }}. 14 April 2011.</ref> In 2019, it became the ITU|ISO/IEC reference implementation as ISO/IEC 10918-7 and ITU-T T.873.<ref name="refimpl">{{cite web |title=ISO/IEC 10918-7: 2023 Information technology — Digital compression and coding of continuous-tone still images — Part 7: Reference software |url=https://www.iso.org/standard/85635.html |access-date=24 June 2025 |website=ISO |language=en}}{{cite web |title=T.873 (05/19): Information technology - Digital compression and coding of continuous-tone still images: Reference software |url=https://www.itu.int/rec/T-REC-T.873-201905-S/en |url-status=live |archive-url=https://web.archive.org/web/20220602173200/https://www.itu.int/rec/T-REC-T.873-201905-S/en |archive-date=2022-06-02 |access-date=2023-03-01 |website=www.itu.int}}</ref>

ISO/IEC Joint Photographic Experts Group maintains the other reference software implementation under the [[JPEG XT]] heading. It can encode both base JPEG (ISO/IEC 10918-1 and 18477–1) and [[JPEG XT]] extensions (ISO/IEC 18477 Parts 2 and 6–9), as well as [[JPEG-LS]] (ISO/IEC 14495).<ref>{{cite web|url=https://jpeg.org/jpegxt/software.html|title=JPEG - JPEG XT|website=jpeg.org|access-date=2018-03-03|archive-date=2018-03-04|archive-url=https://web.archive.org/web/20180304055159/https://jpeg.org/jpegxt/software.html|url-status=live}}</ref> In 2016, "JPEG on steroids" was introduced as an option for the ISO JPEG XT reference implementation.<ref>{{cite book |last1=Richter |first1=Thomas |title=2016 IEEE International Conference on Image Processing (ICIP) |chapter=JPEG on STEROIDS: Common optimization techniques for JPEG image compression |date=September 2016 |pages=61–65 |doi=10.1109/ICIP.2016.7532319 |isbn=978-1-4673-9961-6 |s2cid=14922251}}</ref>

There is persistent interest in encoding JPEG in unconventional ways that maximize image quality for a given file size. In 2014, [[Mozilla]] created '''MozJPEG''' from libjpeg-turbo, a slower but higher-quality encoder intended for web images.<ref>{{cite web |title=Introducing the 'mozjpeg' Project |url=https://research.mozilla.org/2014/03/05/introducing-the-mozjpeg-project/ |website=Mozilla Research |date=5 March 2014 |access-date=2023-03-01 |archive-date=2023-03-01 |archive-url=https://web.archive.org/web/20230301034809/https://research.mozilla.org/2014/03/05/introducing-the-mozjpeg-project/ |url-status=live }}</ref> In March 2017, Google released the open source project [[Guetzli]], which trades off a much longer encoding time for smaller file size (similar to what [[Zopfli]] does for PNG and other lossless data formats).<ref>{{cite web|url=https://research.googleblog.com/2017/03/announcing-guetzli-new-open-source-jpeg.html|title=Announcing Guetzli: A New Open Source JPEG Encoder|website=Research.googleblog.com|date=16 March 2017 |access-date=16 October 2017|archive-date=6 October 2017|archive-url=https://web.archive.org/web/20171006012117/https://research.googleblog.com/2017/03/announcing-guetzli-new-open-source-jpeg.html|url-status=live}}</ref>

In April 2024, Google introduced '''Jpegli''', a new JPEG coding library that offers enhanced capabilities and a 35% compression ratio improvement at high quality compression settings, while the coding speed is comparable with MozJPEG.<ref>{{cite web |title=Introducing Jpegli: A New JPEG Coding Library |url=https://opensource.googleblog.com/2024/04/introducing-jpegli-new-jpeg-coding-library.html |publisher=Google Open Source Blog |access-date=4 April 2024 |archive-url=https://web.archive.org/web/20240403181607/https://opensource.googleblog.com/2024/04/introducing-jpegli-new-jpeg-coding-library.html |archive-date=3 April 2024 |date=3 April 2024}}</ref>

== Successors ==
The Joint Photographic Experts Group has developed several newer standards meant to complement or replace the functionality of the original JPEG format.

=== JPEG LS ===
{{main|JPEG LS}}
Originating in 1993 and published as ISO-14495-1/ITU-T.87, JPEG LS offers a low-complexity lossless file format which was more efficient than JPEG's original lossless implementation. It also features a lossy mode close to lossless. Its functionality is largely limited to that, and largely shares the same limitations of the original JPEG in other aspects.

=== JPEG 2000 ===
{{main|JPEG 2000}}
JPEG 2000 was published as ISO/IEC 15444 in December 2000. It is based on a [[discrete wavelet transform]] (DWT) and was designed to completely replace the original JPEG standard and exceed it in every way. It allows up to 38 bits per colour channel and 16384 channels, more than any other format, with a multitude of colour spaces, and thus high dynamic range (HDR). Furthermore, it supports alpha transparency coding, billions-by-billions pixel images, which is also more than any other format, and lossless compression. It has significantly improved lossy compression ratio with significantly less visible artefacts at strong compression levels.<ref>{{cite web|url=https://cloudinary.com/blog/time_for_next_gen_codecs_to_dethrone_jpeg|title=It's High Time to Replace JPEG With a Next-Generation Image Codec|website=Cloudinary|first=Jon|last=Sneyers |date=22 February 2021|access-date=14 November 2023}}</ref>

=== JPEG XT ===
{{main|JPEG XT}}

JPEG XT (ISO/IEC 18477) was published in June 2015; it extends base JPEG format with support for higher integer bit depths (up to 16 bit), high dynamic range imaging and floating-point coding, lossless coding, and alpha channel coding. Extensions are backward compatible with the base JPEG/JFIF file format and 8-bit lossy compressed image. JPEG XT uses an extensible file format based on JFIF. Extension layers are used to modify the JPEG 8-bit base layer and restore the high-resolution image. Existing software is forward compatible and can read the JPEG XT binary stream, though it would only decode the base 8-bit layer.<ref>{{cite web|url=https://jpeg.org/jpegxt/|title=JPEG - JPEG XT|website=jpeg.org|access-date=2018-03-03|archive-date=2018-03-04|archive-url=https://web.archive.org/web/20180304055205/https://jpeg.org/jpegxt/|url-status=live}}</ref>

=== JPEG XL ===
{{main|JPEG XL}}
JPEG XL (ISO/IEC 18181) was published in 2021&ndash;2022. It replaces the JPEG format with a new DCT-based [[royalty-free]] format and allows efficient transcoding as a storage option for traditional JPEG images.<ref name=spie_jpegxl>{{cite book|chapter=JPEG XL next-generation image compression architecture and coding tools|chapter-url=https://www.spiedigitallibrary.org/conference-proceedings-of-spie/11137/111370K/JPEG-XL-next-generation-image-compression-architecture-and-coding-tools/10.1117/12.2529237.full?SSO=1|first1=Jyrki|last1=Alakuijala|first2=Ruud|last2=van Asseldonk|first3=Sami|last3=Boukortt|first4=Martin|last4=Bruse|first5=Iulia-Maria|last5=Comșa|first6=Moritz|last6=Firsching|first7=Thomas|last7=Fischbacher|first8=Evgenii|last8=Kliuchnikov|first9=Sebastian|last9=Gomez|first10=Robert|last10=Obryk|first11=Krzysztof|last11=Potempa|first12=Alexander|last12=Rhatushnyak|first13=Jon|last13=Sneyers|first14=Zoltan|last14=Szabadka|first15=Lode|last15=Vandervenne|first16=Luca|last16=Versari|first17=Jan|last17=Wassenberg|editor1-first=Andrew G|editor1-last=Tescher|editor2-first=Touradj|editor2-last=Ebrahimi|title=Applications of Digital Image Processing XLII|date=2019-09-06| volume=11137 |page=20|doi=10.1117/12.2529237|bibcode=2019SPIE11137E..0KA|isbn=978-1-5106-2967-7|s2cid=202785129|access-date=2021-12-26|archive-date=2021-12-26|archive-url=https://web.archive.org/web/20211226010413/https://www.spiedigitallibrary.org/conference-proceedings-of-spie/11137/111370K/JPEG-XL-next-generation-image-compression-architecture-and-coding-tools/10.1117/12.2529237.full?SSO=1|url-status=live}}</ref> The new format is designed to exceed the still image compression performance shown by [[HEIF]] HM, [[Daala]] and [[WebP]]. It supports billion-by-billion pixel images, up to 32-bit-per-component [[high dynamic range]] with the appropriate transfer functions ([[Perceptual quantizer|PQ]] and [[Hybrid log–gamma|HLG]]), patch encoding of synthetic images such as bitmap fonts and gradients, animated images, alpha channel coding, and a choice of RGB/YCbCr/[[ICtCp]] color encoding.<ref name="jpegxl_committeedraft">{{cite arXiv|title=Committee Draft of JPEG XL Image Coding System|eprint = 1908.03565|last1 = Rhatushnyak|first1 = Alexander|last2 = Wassenberg|first2 = Jan|last3 = Sneyers|first3 = Jon|last4 = Alakuijala|first4 = Jyrki|last5 = Vandevenne|first5 = Lode|last6 = Versari|first6 = Luca|last7 = Obryk|first7 = Robert|last8 = Szabadka|first8 = Zoltan|last9 = Kliuchnikov|first9 = Evgenii|last10 = Comsa|first10 = Iulia-Maria|last11 = Potempa|first11 = Krzysztof|last12 = Bruse|first12 = Martin|last13 = Firsching|first13 = Moritz|last14 = Khasanova|first14 = Renata|author15 = Ruud van Asseldonk|last16 = Boukortt|first16 = Sami|last17 = Gomez|first17 = Sebastian|last18 = Fischbacher|first18 = Thomas|class = eess.IV|year = 2019}}</ref><ref name="jpegxl_cfp">{{cite web|url=https://jpeg.org/downloads/jpegxl/jpegxl-cfp.pdf|title=N79010 Final Call for Proposals for a Next-Generation Image Coding Standard (JPEG XL)|work=ISO/IEC JTC 1/SC 29/WG 1 (ITU-T SG16)|access-date=29 May 2018|archive-date=31 October 2022|archive-url=https://web.archive.org/web/20221031184400/https://jpeg.org/downloads/jpegxl/jpegxl-cfp.pdf|url-status=live}}</ref><ref>{{Cite ISO standard |title=ISO/IEC 18181-1:2022 Information technology — JPEG XL image coding system — Part 1: Core coding system |csnumber=77977 }}</ref><ref>{{Cite ISO standard |title=ISO/IEC 18181-2:2021 Information technology — JPEG XL image coding system — Part 2: File format |csnumber=80617 }}</ref>

==See also==
* [[AVIF]]
* [[Better Portable Graphics]], a format based on intra-frame encoding of the HEVC
* [[C-Cube]], an early implementer of JPEG in chip form
* [[Comparison of graphics file formats]]
* [[Deblocking filter (video)]], the similar deblocking methods could be applied to JPEG
* [[Design rule for Camera File system]] (DCF)
* [[FELICS]], a lossless image codec
* [[File extensions]]
* [[Graphics editing program]]
* [[High Efficiency Image File Format]], image container format for [[High Efficiency Video Coding|HEVC]] and other image coding formats
* [[Lenna (test image)]], the traditional standard image used to test image processing algorithms
* [[Motion JPEG]]
* [[Ultra HDR]], adds [[high dynamic range]] (HDR) support to JPEG with backwards compatibility
* [[WebP]]

==References==
{{Reflist}}

==External links==
{{Commons category|JPEG compression}}
* {{Official website}}
* [https://www.w3.org/Graphics/JPEG/itu-t81.pdf JPEG Standard (JPEG ISO/IEC 10918-1 ITU-T Recommendation T.81)] at W3.org
* [https://www.w3.org/Graphics/JPEG/jfif3.pdf JPEG File Interchange Format (JFIF), Version 1.02, Sept. 1992] at W3.org
* [https://www.loc.gov/preservation/digital/formats/fdd/fdd000017.shtml Format description of JPEG Image Coding Family] from [[Library of Congress]]
* [https://www.visengi.com/products/jpeg_hardware_encoder Example images over the full range of quantization levels from 1 to 100] at visengi.com<!--Use real title "JPEG Encoder" and move former text after) and say something on "This IP core"?-->

{{Compression formats}}
{{Graphics file formats}}

[[Category:JPEG| ]]
[[Category:Articles containing video clips]]
[[Category:Computer-related introductions in 1992]]
[[Category:Discovery and invention controversies]]
[[Category:IEC standards]]
[[Category:Image compression]]
[[Category:ISO standards]]
[[Category:ITU-T recommendations]]
[[Category:Lossy compression algorithms]]
[[Category:Open formats]]
[[Category:Raster graphics file formats]]