Tiêu chuẩn iso ts 19264 1 2017

These an othe charact eris ics of a digital imag e can be as es ed b ima ing sy st ems q alty analy sis.. To this en , ima ing sy st ems q alty analy sis can as is those develo ing or ac

Photog raphy — A rchi ving systems —

Part 1:

Ph to ra hie — S stème s d’ar hiva e — An lyse de la q alité

d’ima e —

Partie 1: Document réfléchissant

Fir t edition

2 17-04

COPYRIGHT PROTECTED DOCUMENT

© ISO 2017, P blshed in Sw itz rlan

A ll rig hts r eserved Unles otherw ise spe ified, nopar of this p blc tion ma y be r epr od c d or utilz d otherw ise in an form

or b an me ns, ele tr onic or me hanic l, inclu in p oto opying , or postin on the internet or an intranet , w ithout prior

written permis ion Permis ion c n be req esed from either ISO at the ad r es below or ISO’s member bod y in the c u try of

F reword i v

Introduction v

1 Sc ope 1

2 Nor mati ve r eferenc es 1

3 Terms and definitions 1

4 System setup and calbration 10

4.1 General 1

4.2 Sys em config uration 1

4.3 Camera/scan e seting s 1

4.4 Ex posur e 1

4.5 White b lancing 1

4.6 IC Pr ofi ng 1

4.7 Focusing 1

4.8 Colour encoding 1

4.9 Reprod ction scale 1

5 Imag ing system qual ty anal ysis proc edure 12

6 Imag ing systems q alty characteristic and metric 13

6.1 General 1

6.2 Tones an noise 1

6.3 Colour 1

6.4 Detais 1

6.5 Geometry 2

7 Repor ting r esults 21 7.1 General 2

7.2 Example r eport for tone repr od ction results 2

7.3 Gain modulation 2

7.4 Dynamic rang e .24

7.5 Noise 24

7.6 Banding 2

7.7 Defe t pixels 2

7.8 White b lanc 2

7.9 Colour r epr od ction 2

7.1 Colour mis-r eg is ration 2

7.1 Sampl ng rate .2

7.1 Resolution 2

7.1 MTF 5 /MTF 1 2

7.14 Sharpening 2

7.1 A cutanc 2

7.1 I luminanc non-u iformity 2

7.1 Chrominanc non-uniformity 2

7.1 Dis ortion 2

7.1 Reprod ction scale 2

A nne x A (normative) Test char t req ir ements 30 A nne x B (normative) Guidelnes for imag ing per for mance aims and toleranc es 32 A nne x C (informative) Example of multipat er n char t:Uni ver sal Test Tar g et (UTT) 34 Biblog raphy 55

ISO (he Int ernational Org nization for Stan ardization) is a worldwidefede ation of national s an ards

b dies (ISO membe b dies) The work of pr p ring Int ernational Stan ards is normaly car ied out

through ISO t ech ical committ ees Each membe b dy int er st ed in a subje t for w hich a t ech ical

committ ee has be n es a lshed has the right t o be r pr sent ed on that committ ee Int ernational

org nizations, g ove nmental an non-g ove nmental, in laison with ISO, also take part in the work

ISO cola orat es closely with the Int ernational Ele trot ech ical C mmis ion (IEC) on al matt ers of

ele trot ech ical s an ardization

The proc d r s used t o develo this document an those int en ed for it furthe maint enanc ar

desc ibed in the ISO/IEC Dir ctives, Part 1 In p rticular the dife ent a pro al c it eria ne ded for the

dife ent ty es of ISO document should be not ed This document was draft ed in ac ordanc with the

edit orial rules of the ISO/IEC Dir ctives, Part 2 www iso org dir ctives

A tt ention is drawn t o the p s ibi ity that some of the element of this document ma be the subje t of

p t ent right ISO shal not be held r sponsible for identifying any or al such p t ent right Detais of

any p t ent right identified d ring the develo ment of the document wi be in the Introd ction an / r

on the ISO ls of p t ent de larations r c ived www iso org pat ent

Any trade name used in this document is information given for the convenienc of use s an does not

cons itut e an en orsement

F or an ex lanation on the meaning of ISO spe ific t erms an ex r s ions r lat ed t o conformity

as es ment, as wel as information a out ISO’ s adhe enc t o the World Trade Org nization (WTO)

principles in the Te h ical Bar ie s t o Trade (TBT), se the folowing URL: ht p:/ www iso org iso /

for word html

The committ ee r sp nsible for thisdocument is ISO/TC42, Ph to ra h

Ele tronic ima ing sy st ems, such asscan e s and came as, can be used for digitizing physical r cords,

e.g document , pictur s, ma s The r sulting digital imag es can bemor or les ac urat e in t erms of how

wel they r prod c the original r cord t ones, colours, detai s, et c These an othe charact eris ics of

a digital imag e can be as es ed b ima ing sy st ems q alty analy sis In g ene al, the achiev ble ac uracy

of digital r prod ctions depen son the natur of the original r cord an the digitization, espe ial y the

pe formanc of the ima ing sy st em an the a pled sy st em set ings

In some org nizations, e.g within the ar hiving an cultural he itag e field, w he e conside a le

r sour es ar put int o digitization proje t , it is key t o ensur that the r q ir d ima ing sy st ems

q alty is met an that it is consist ent To this en , ima ing sy st ems q alty analy sis can as is those

develo ing or acq iring ima ing sy st ems with the as es ment an ve if ication of sy st em pe formanc ,

such as the spe ified r solution an dynamic rang e of a scanne , an the comp rative pe formanc of

dife ent ima ing sy st ems Ima ing sy st ems q alty analysis is also used for set ing up an calbrating

ima ing sy st ems as wel as for enhancing their pe formanc Final y, ima ing syst ems q al ty analy sis

is used for as es ing ac uracy and control ng ima ing consist ency o e time Not e, that w hie the

ne d t o ensur ima ing sy st ems q alty is g ene ic, the r q ir d level of ima ing syst ems q alty an

ac uracy is use-case spe ific F or ex mple, w hen digitizing wat er olours it is usualy es ential t o r ach

a high degr e of ac uracy in the ca tur of the colour information, w hie this is not normaly eq aly

c itical w hen digitizing new spa e s Also, some imag e proc s ing pro rams, such as Optical Charact er

R eco nition (OCR), ar mor ac urat e if the contras is enhanc d d ring ima ing

In practic , ima ing sy st ems q alty is analysed b digitizing a phy sical r fe enc targ et (t es chart)

with k own (measur d) v lues an comp ring these r fe enc v lues t o the cor espon ing ca tur d

v lues r pr sent ed in the digital imag e file ( e Figur 1)

The use of a t es chart ensur s that the ima ing sy st ems q alty charact eris ics can be det ermined

o je tively Howeve , t o be usa le the q alty of the targ et ne ds t o ex ce d the pe formanc of the

ima ing sy st em F or ex mple, t o det ermine the r solution of an ima ing sy st em, the targ et ne ds t o

ha e a t ech ical p tt ern with mor detai s than the sy st em is ca a le of r solving Ima ing sy st ems

q alty analy sis r p rt how ac urat ely the ima ing sy st em r prod c s the r fe enc targ et The efor ,

if the original r cord dife s signif icantly from the targ et, e.g with r spe t t o t one, t onal rang e, colours,

detai s, and lght r fle tance /a sorb nc , this ma , in spit e of a wel pe forming sy st em, compromise

the ac uracy of the r prod c d imag e S e also R efe enc s [2 ] an [2 ] Idealy, the targ et should

r semble the natur of the original mat erial Howeve , given the many dife ent ty es of original r cords

this is oft en not practical or t ech icaly impos ible Even though sy st ems ma pe form dife ently

on the dife ent ty es of originals this document pro ides t ools t o ve ify if a sy st em is ac urat ely

calbrat ed an in g ene al pe forms wel on a sele t ed ty e of original This is sufficient in mos cases

be a use sy st ems ar usualy designed t o han le v rious ty es of originals (being close t o the Luthe

con ition) Pe formanc on spe if ic ty es of originals howeve can only be verif ied if the t ools ar made

of that mat erial It is also important t o not e that an ac urat e r prod ction usual y r q ir s subseq ent

proc s ing t o r n e a visualy pleasing imag e

The e ar ISO s an ards for o je tively measuring dife ent pe formanc charact eris ics of ima ing

sy st ems, e.g r solution, noise, dynamic rang e, t one an colour r prod ction ( e Clause 2) This

document combines al of the s an ards that r lat e t o the ima ing sy st ems q alty analy sis for cultural

he itag e an defines a t ool set t o a ply them t o these devic s an workflow s These t ools ar b sed

on the use of a t es chart with multiple t echnical p tt erns coupled with sof war that alows the use

t o analy se several ima ing sy st ems q alty charact eris ics simultaneously an r c ive compr hensive

r sult Howeve , these t ools ar not b sed on a s an ardiz d imag e q alty analy sis method, w hich

has caused confusion among use s W ith the publcation of this spe ification ima ing sy st ems q alty

analy sist ools can r fe t o an ISO document

To sup ort this document a s andard with a glos ary inclu ing al r lev nt t erms an definitions

Photog raphy — A rchi ving systems — Imag e quality

Part 1:

This document desc ibes a method for analy sing ima ing sy st ems q alty in the ar a of cultural he itag e

ima ing The method desc ibed analy ses multiple ima ing sy st ems q al ty charact eris ics from a single

imag e of a spe if ied t es targ et The spe if ication s at es w hich charact eris ics ar measur d, how they

ar measur d, an how the r sult of the analy sis ne d t o be pr sent ed

This spe if ication a pl es t o scan e s an digital came as used for digitization of cultural he itag e

mat erial

NOTE T is document ad res es ima ing of reflective originals, a future part two wil ad res ima ing of

transparent originals

2 Normati ve r eferences

The folowing document ar r fe r d t o in the t ext in such a wa that some or al of their cont ent

cons itut es r q ir ment of this document F or dat ed r fe enc s, only the edition cit ed a ples F or

un at ed r fe enc s, the lat es edition of the r fe enc d document ( inclu ing any amen ment ) a ples

ISO 1 2 3, Ph to ra h — Elec tro ic s ti pic tur ima in — R es olutio a d s patial fr q ency r s po ses

ISO 145 4, Ph to ra h — Elec tro ic s ti -pic tur c ame as — Meth ds fr meas urin o to-elec tro ic

c on e s io f nc tio s (OE Fs)

ISO 1 7 9, Ph to ra h — Elec tro ic s ti -pic tur ima in — Noi s e meas ur ments

ISO 1 0 7-1, Ph to ra h — S atial r s olutio meas ur ment o elec tro ic s c an e s fr p oto ra hic

ima es — Part 1: Sc an e s fr r flec ti ve media

ISO 1 9 7, Ph to ra h — Digital c ame as — S adin meas ur ments

ISO 2 5 0, Ph to ra h — Elec tro ic s c an e s fr p oto ra hic ima es — Dy namic ra ge meas ur ments

CIE 1 , Colorimetr y

3 Terms and definitions

F or the purposes of this document, the folowing t ermsand definitions a ply

ISO an IEC maintain t erminolo ical data ases for use in s an ardization at the folowing ad r s es:

— ISO Onlne brow sing plat orm:a ai a le at ht p:/ www iso org o p

— IECEle tro edia: a aia le at ht p:/ www ele tro edia org

thr e-component colour imag e encoding def ined in Ado e RGB (1 98)colour imag e encoding

Not e 1 t o entry: Ado e RGB 1 9 Colour Ima e E coding can b fou d u der the fol owing URL ht ps:/ www

.ado e com/ digitalima / p fs/ A do eRGB1 9 p f

[ SOURCE:ISO 1 640-4:2 1 , 3.1, modif ied — ad ition of the Not e 1 t o entry.]

3.3

banding

ima ing

unwant ed s ripes or b n s that oc ur in a digital imag e

Not e 1 to entry: Note 1 t o entry: Ban s are usual y caused by f ix ed patt ern noise of sensor in s an er ,

interference pro lems b tween electronic parts of a camera, or by to -co r e q antization

[SOURCE: ISO 1 2 2:2 1 , 3.9, modif ied — ad ition of “or b to -co rse q antization” in the Note 1

chromaticnes , colourfulnes , of an ar a ju g ed as a pro ortion of the brightnes of a simi arly

i luminat ed ar a that a pears w hit e or highly transmit ing

[ SOURCE:ISO/IEC 8613- 2:1 9 , 3.1 ]

3.6

CIE AB colo r space

thr e-dimensional, a pro imat ely u iform colour sp c , prod c d b plot ing, in r ctangular

co rdinat es L*, a*, b*

Not e 1t o entry: T is colour space has b en designed t o b device in epen ent

[ SOURCE: CIE Publcation 1 an ISO/IEC 5 3 -1:2 1 , 3.6, modified — Not e 1 t o entry has be n

modified.]

3.7

colour

sensation r sulting from the visual pe c ption of radiation ofa given spe tral comp sition

[ SOURCE: ISO 46 18:2 14, 2.58, modif ied — def inition sl ghtly chang ed an Not e 1 and Not e 2 t o entry

ha e be n delet ed.]

3.8

colour dif erence

dis inction betwe n two colours o se ved or measur d u de s an ard con itions

[ SOURCE:ISO 1 63 - 2:2 08, 2.2 ]

3 9

colour encoding

g ene ic t erm for a q antiz d digital encoding of a colour sp c , encomp s ing b th colour sp c

encodings and colour imag e encodings

[ SOURCE:ISO/TS2 0 8-3:2 1 , 3.6]

3.1

colour mis e istration

colour-t o-colour spatial dislocation of othe wise spatialy coincident colour featur sof an imag ed o je t

digital f ile consis ing of pictur element (pix els) with one or mor digital code v lues pe pix el that

r pr sent a colour or t onal v lue

[ SOURCE:ISO 1 2 2:2 1 , 3.7 , modif ied — deletion of the Not e 1 t o entry.]

3.1

digital ima ing

proc s of c eating digital imag es

Note 1t o entry: T e term can also b used more g neraly t o inclu e digital imag e proces ing

[ SOURCE:ISO 1 2 2:2 1 , 3.74 ]

3.1

digital ima ing sy stem

sy st em that r cords an / r prod c s imag es using digital data

[ SOURCE:ISO 1 2 1:2 1 , 3.3 ]

3.17

digital stil camera

DSC

devic w hich incorp rat es an imag e sensor and prod c s a digital signal r pr senting a s i l pictur

Note 1 t o entry: A digital stil camera is ty icaly a porta le, han -held device T e digital signal is usual y

recorded on a remova le memory, such as a solid-state memory card or ma netic disk

digitizatio

act of g ene ating a digital (q antiz d)r pr sentation of a contin ous signal

[ SOURCE:ISO 2 9 8-1:2 0 , 2.7, modif ied — T e Not e 1 t o entry has be n delet ed.]

Not e 1 to entry: It b sicaly derives from variation of lat eral ma nif ication in the imag e f ield of a camera lens

an results in straight lines b ing ren ered as curves There are other factor t o in uce g ometric dist ortion, for

e ample rotational asymmetricity of a camera lens or position shif proces ing in a camera ima ing proces

[ SOURCE:ISO 1 2 2:2 1 , 3.8 ]

3.20

dy amic rang e

dife enc , o e a given luminanc rang e, between ma imum an minimum signal levels, ex r s ed in

de ibels, contras ratios or f-st ops

Not e 1t o entry: T e minimum signal level ne ds t o b gre ter than a specif ied usa le signal level

Not e 2 t o entry: This def inition is derived from IE 70 – 4– 3 but was altered to mat ch the ima ing an

ar hiving ap lication

[ SOURCE:ISO 1 2 2:2 1 , 3.8 ]

3.20.1

ISO DSC dy amic rang e

ratio of the ma imum luminanc level that a pears unclp ed t o the minimum luminanc level that can

be r prod c d with an inc emental signal-t o-t emporal-noise ratio of at leas 1, as det ermined ac ording

t o ISO 1 7 9

[ SOURCE:ISO 1 2 1:2 1 , 3.8 ]

3.20.2

ISO s anne dy amic rang e

dife enc of the ma imum density w he e the inc emental g in is highe than 0,5, as det ermined

ac ording t o ISO 2 5 0 t o the minimum density that a pears u clp ed

[ SOURCE:ISO 2 5 0:2 04, 3.1 ]

3.21

ex osure

H

< hotogra phic> total q antity of lght alowed to fal up n a photosensitive em ulsion or an ima ing sensor

Not e 1t o entry: T e e posure is me sured in lux per secon

[ SOURCE: ISO 1 9 4-1:2 0 , 2.5 , modif ied — A symb l, the field of a plcation an a not e t o entry

ha e be n ad ed.]

3.22

fast s an directio

scan dir ction cor espon ing t o the dir ction of the al gnment of the ad r s a le phot oelement in a

l near ar a imag e sensor

[ SOURCE:ISO 1 0 7-1:2 0 , 3.7]

g in mo ulatio

v riation of the g in o e the signal level

Note 1t o entry: One e ample for a g in mod lation is the ap lication of a g mma to an ima e

[ SOURCE:ISO 1 2 2:2 1 , 3.1 9]

3.24

gray s ale

grey s ale pattern

t es chart consis ing of t es patt ern based on spe tral y neutral or efe tively spe tral y neutral, an

consis s of a larg e n mbe of dife ent r fle tanc or transmit anc v lues in a pr sc ibed sp tial

ar ang ement

Note 1t o entry: Grey s ale pat erns are ty ical y used to me sure opt o-electronic conver ion fu ctions

3.25

horizo tal resolutio

r solution v lue measur d in the long er imag e dimension, cor espon ing t o the horizontal dir ction for

a “lan sca e” imag e orientation, ty icaly using a vertical y orient ed t es -chart featur

[ SOURCE:ISO 1 2 1:2 1 , 3.6 ]

3.26

IC profi e

Inte national Colour Consortium’ s f ile format, used to s or trans orms from one colour encoding

to anothe , e.g from devic colour co rdinates to prof ile conne tion spac , as part of a colour

mana ement ys em

[ SOURCE:ISO 2 0 8-1:2 1 , 3.24 ]

3.27

imag e q al ty

impr s ion of the o e al me it or ex celenc of an imag e, as pe c ived b an o se ve neithe as ociat ed

with the act of phot ogra hy, nor closely inv lved with the subje t matt er depict ed

Note 1 to entry: The p rpose of def ining ima e q ality in t erms of third-party (u involved) o server is to

eliminat e sour es of varia ility that arise from more idiosy c atic aspects of ima e per eption an pertain to

at ributes outside the control of ima ing sy st em designer

[ SOURCE:ISO 2 46 -1:2 0 , 3.5]

3.28

l miting resolutio

v lue of that portion of a spe ified r solution t es p tt ern, measur d in l ne widths pe pictur

height, w hich cor esp n s t o an a e ag e mod lation v lue eq al t o some spe if ied pe c ntag e of the

mod lation v lue at a spe if ied r fe enc fr q ency

Note 1 t o entry: T e limiting resolution could b the t est pat ern value, in line widths per picture height w

l/

p),

cor espon ing t o a camera outp t mod lation level of 1 % of the camera outp t mod lation level at a reference

freq ency of 1 w

l/

p

3.29

maximum modulatio

ma imum v lue of thespatial fr q ency r sp nse

Note 1t o entry: Ma imum mod lation is an in icator for ap lied sharpening

r lationship betwe n the lo of the input levels an the cor esp n ing digital output levels for an o t

o-ele tronic digital imag e ca tur sy st em

Not e 1 t o entry: If the in ut lo e posure points are very f inely spaced an the outp t noise is smal compared

t o the q antization interval, the OE F pos ibly has a st ep-like character Such b haviour is an artefact of the

q antization proces an ne ds to b removed by using an ap ropriat e smo thing alg rithm or by f it ing a

smo th curve t o the data

[ SOURCE:ISO 1 3 1-1:2 1 , 3.3]

3.33

original refer ed imag e state

s ene-refer ed

imag e s at e as ociat ed with imag e data that r pr sent the colour-spac co rdinat es ofthe element of

a two dimensional hardco y or soft co y imag e, ty icaly prod c d b scan ing artwork, phot ogra hic

transpar ncies or print , or phot ome hanical or othe r prod ctions

Not e 1 t o entry: When the p rase “original-refer ed” is used as a q alif ier to an o ject, it implies that he o ject

is in an original-refer ed ima e state F or e ample, original-refer ed ima e data are ima e data in an

original-refer ed ima e stat e

Not e 2 t o entry: Original-refer ed ima e data are related t o the colour -space co rdinat es of the original, ty ical y

me sured ac ording to ISO 1 65 , an do not inclu e an ad itional veiling glare or other flare

Not e 3 t o entry: The charact eristic of original-refer ed ima e data that most g neraly distinguish them from

s ene-refer ed ima e data are that hey refer to a two-dimensional surface, an the il umination incident on the

two-dimensional surface is as umed t o b u iform (or the ima e data cor ected for an non-u iformity in the

ilumination)

Not e 4 t o entry: There are clas es of originals that prod ce original-refer ed ima e data with diferent

characteristic Ex mples inclu e various ty es of artwork, p oto rap ic prints, p oto rap ic transparencies,

emis ive displays, etc When selecting a colour re-ren ering alg rithm, it is usualy neces ary to k ow the clas of

the original in order to determine the ap ropriate colour re-ren ering t o b ap lied For e ample, a colourimetric

intent is g neraly ap lied t o artwork, while diferent per eptual alg rithms are ap lied to prod ce p ot ograp ic

prints from transparencies, or newsprint reprod ctions from p oto rap ic prints In some cases the as umed

viewing con itions are also diferent b tween the original clas es, such as b tween p oto rap ic prints an

transparencies, an wil usualy b considered in wel -designed systems

Not e 5 to entry: In a few cases, it can b desira le to introd ce slight colourimetric er or in the prod ction of

original-refer ed ima e data, for e ample t o make the g mut of the original more closely f it he colour space, or

b cause of the way the imag e data were captured (such as a Status A densitometry-b sed s an er)

[ SOURCE:ISO 2 0 8-1:2 1 , 3.3 , modified — A t erm has be n slghtly modified an se on one ad ed.]

output-refer ed imag e state

imag e s at e as ociat ed with imag e data that r pr sent the colour-spac co rdinat es of the element of

an imag e that has un e g one colour-r n e ing a pro riat e for a spe ified r al or virtual output devic

an viewing con itions

Note 1 to entry: When the p rase “outp t-refer ed” is used as a q alif ier t o an o ject, it implies that he o ject is

in an outp t-refer ed ima e state For e ample, outp t-refer ed ima e data are ima e data in an outp t-refer ed

ima e state

Note 2 to entry: Outp t-refer ed ima e data are refer ed to the specif ied outp t device an viewing con itions A

single s ene can b colour -ren ered t o a variety of outp t-refer ed representations depen ing on the anticipated

outp t-viewing con itions, media limitations, an / r artistic int ents

Note 3 t o entry: Outp t-refer ed ima e data can b come the starting point for a subseq ent reprod ction proces

For e ample, sRGB outp t-refer ed ima e data are freq ently considered to b the starting point or the colour

re-ren ering performed by a print er designed to receive sRGB ima e data

al those plan ed an sy st ematic activities ne es ary t o pro ide con denc that a prod ct satisfies

given ac eptanc c it eria

ar ang ement of t es patt erns designed t o t es p rticular aspe t of an ima ing sy st em

Note 1t o entry: S e e amples in ISO 1 2 3:2 17, ISO 16067-1 an ISO 16067-2

[ SOURCE:ISO 1 2 2:2 1 , 3.2 7]

3.39

reprod ctio s ale

ratio of the siz of an o je t in a digital imag e an the siz of the original o je t

[ SOURCE:ISO 1 2 2:2 1 , 3.2 5]

3.40

reprographic iluminatio g eometry

ty ical ar ang ement of the i lumination in two dimensional r pro ra hic phot ogra hy w he e the lght

ar p sitioned on two sides of the original in a 45° angle t o the plane of original and t o the came a,

w hich is p sitioned pe pen icular t o the plane of the original

resolutio

theoretical resolutio

l miting resolutio

measur of the a i ty of a came a sy st em, or a comp nent of a came a sy st em, t o depict pictur detai

Not e 1t o entry: R esolution me surement metric inclu e resolving power, limiting resolution, special freq ency

ratio of the measur d l miting r solution an the Nyq is fr q ency

Not e 1t o entry: Both values ne d to have the same u it

[ SOURCE:ISO 1 2 2:2 1 , 3.2 0]

3.44

sampl ng rate

n mbe of samples pe unit of time, angle, r v lutions or othe me hanical, indepen ent v ria le for

uniformly sampled data

s ene refer ed imag e state

imag e s at e imag e s at e as ociat ed with imag e data that r pr sent es imat es of the colour-sp c

co rdinat es of the element of a sc ne

Not e 1 t o entry: When the p rase “ cene-refer ed” is used as a q alifier t o an o ject, it implies that the o ject is in a

s ene refer ed imag e stat e F or e ample, s ene-refer ed imag e data are imag e data in a s ene-refer ed imag e stat e

Not e 2to entry: Scene-refer ed ima e data can b det ermined from raw DSC ima e data b fore colour -ren ering

is performed General y, DSCs do not write s ene-refer ed ima e data in ima e f iles, but some do so in a special

mode inten ed for this p rpose T picaly, DSCs write stan ard outp t-refer ed ima e data where colour

-ren ering has alre d b en performed

Not e 3 t o entry: Scene-refer ed ima e data ty ical y represents relative s ene colourimetry estimates

Absolute s ene colourimetry estimates can b calculat ed using a s aling fact or T e s aling factor can b

derived from ad itional information such as the ima e OE F, F-n mb r or Aperture Value, an ExposureTime or

Sh tt erSpe dValue ta s

Not e 4 t o entry: Scene-refer ed ima e data can contain inac uracies d e t o the d namic rang limitations of the

capture device, noise from various sour es, q antization, optical blur ing an flare that are not cor ected for,

an colour analysis er or d e to capture device metamerism In some cases, these sour es of inac uracy can b

signif icant

Not e 5t o entry: T e transformation from raw DSC imag e data t o s ene-refer ed imag e data depen s on the relative

adopt ed w hit es select ed for the s ene an the colour space used t o encode the imag e data If the chosen s ene

adopt ed w hit e is inap ropriat e, ad itional er or wil b introd ced int o the s ene-refer ed imag e data T ese

er or can b cor ecta le if the transformation used t o prod ce the s ene-refer ed imag e data are k own, an the

colour encoding used for the incor ect s ene-refer ed imag e data has adeq at e precision an d namic rang e

Note 6 to entry: T e s ene can cor espon t o an actual view of the natural world, or b a comp t er -g enerated

virtual s ene simulating such a view It can also cor espon t o a modif ied s ene determined by ap lying

modif ications t o an original s ene t o prod ce some diferent desired s ene An such s ene modif ications ne d

t o le ve the ima e in a s ene refer ed ima e state, an ne d to b done in the conte t of an e pected colour

-ren ering transform

ampl fication of the SFR b means of imag e proc s ing t o achieve sharpe a pearing imag es

Note 1 t o entry: Also, a clas of ima e proces ing operations that en ances the contrast of selective spatial

freq encies, usual y visual y important ones

slow s an directio

dir ction in w hich the scan e mo es the phot o element (pe pen icular t o the lnes of active phot o

element in a l near ar a imag e sensor)

[ SOURCE:ISO 1 0 7-1:2 0 , 3.1 ]

3.5

spatial frequency respo se

SFR

measur d ampltu e r sp nse of an ima ing sy st em as a fu ction of r lative input sp tial fr q ency

Note 1 t o entry: T e SFR is normal y represented by a curve of the outp t response t o an in ut signal of u it

amplitu e, over a rang of spatial freq encies

Note 2t o entry: T e SFR is normalized t o yield a value of u ity at a spatial freq ency of 0

Note 3t o entry: In eq ations, the symb l RSFR rather than the a breviation SFR is used for clarity

test pattern

spe ified ar ang ement of spe tral r fle tanc or transmit anc charact eris ics used in measuring an

ima ing sy st ems q alty at ribut e

r solution v lue measur d in the short er imag e dimension, cor espon ing t o the ve tical dir ction for a

“lan sca e” imag e orientation, ty icaly using a horizontal y orient ed t es chart featur

[ SOURCE:ISO 1 2 3:2 1 , 3.2 ]

3.56

w hite balance

adjus ment of ele tronic s i pictur colour channel g ins or imag e proc s ing so that radiation with

r lative spe tral p we dis ribution eq al t o that of the sc ne i umination sour e is r n e ed as a

visual neutral

[ SOURCE:ISO 145 4:2 09, 3.1 ]

4 System setup and cal bration

The imag eca tur sy st em ne ds t o be car fuly set up t o ensur consist ent, r peata le, an high q alty

r sult Prior t o che king or conf irming the q alty of the sy st em, it alway s ne ds t o be ac urat ely

calbrat ed an adjust ed F or a mor detaied desc iption on how t o set up an calbrat e an ima ing

sy st em prior t o ima ing sy st ems q alty analy sis se ISO/TR 1 2 3

4.2 System c onfig uration

The came a ne ds t o be mount ed on a sold s an that does not mo e d ring ex osur Any ambient

l ght that does not originat e from the desir d i lumination shal be a oided

4.3 Camera/s anner set ing s

The lowes sensitivity an lowes imag e compr s ion rat e, i.e the highes imag e q alty, should be

sele t ed

4.4 Ex posur e

The ex osur shal be adjust ed so a difuse w hit e flat surfac (a t es chart ma be used for this) is

ca tur d an r corded using encoding v lues that ha e an L* v lue eq al t o the actual L* v lue of the

dif use w hit e flat surfac In the case of a thr e-dimensional original the plac ment an orientation

of the difuse w hit e flat surfac ar lef t o the phot ogra he , but should r sult in a r asona le imag e

a pearanc (w hen displa ed ac urat ely) comp r d t o viewing the original The use ne ds t o make

sur that the dark ar as ar also not clp ed If clp ing in the black ar as is encount er d, the use

ne ds t o ensur that the sy st em is a le t o ca tur the dynamic rang e of the original r fe ring t o the

measur ment desc ibed in ISO 2 5 0

4.5 White balancing

The w hit e b lanc shal be measur d on a gr y card or a w hit e card (without o tical bright ene s)

t o ensur cor e t an consist ent r sult This gr y r fe enc is r q ir d t o be spe traly neutral in

r fle tion an the sur oun ing shal not ha e a dominating colour These set ings shal be st or d an

used for prod ction aft erwards Thisproc s shal be r peat ed on a r gular b sis t o compensat e for the

spe tral chang e of the l ght sour eo e it lfetime Depen ing on the ty e of lght our e the int erv l in

w hich this ne ds t o be done v ries

Whit e b lanc pe formed on dife ent t onal levels can v ry Highl ght ar g ene aly mor sensitive t o

e rors To che k the v rianc s of a sy st em, it is bes t o use a gr y scale an try dife ent t onal levels

4.6 ICC Pr ofi ing

If the originals ar ca tur d using a colour ima ing sy st em, an IC prof ile should be c eat ed t o

charact eriz the sy st em F or the purp se of IC profilng, an ideal colour t es chart r fle t the ty e of

originals t o be digitiz d in t erms of mat ching mat erial an colourant

If the sof war does not sup ort IC colour manag ement, it is c itical t o det ermine if the sy st em sensor,

or any int ernal calbration, r aches ac urat e colour r prod ction in the desir d encoding befor you

de ide t o pur hase or use the sy st em

4.7 F cusing

The sy st em shal cor e tly be focused on the original It depen s on the t ools the sy st em has a aia le

how a g ood focus level can be achieved A ut o focus sy st ems ar oft en not r la le an ma ha e pro lems

focusing on c rtain originals without the introd ction of focus aids

4.8 Colour enc oding

The desir d colour encoding should be sele t ed b sed on the int en ed a plcation r q ir ment an

workflow pr fe enc s In ISO 2 0 8-1:2 1 , Annex B ls s the charact eris ics an sour e s an ards for

a n mbe of s an ard colour encodings an An ex C pro ides c it eria for sele tion of colour encodings

In g ene al, original- an sc ne-r fe r d encodings ar mos a pro riat e for digital ar hiving sy st ems

Ex mples of original-r fe r d imag es ar pro ided in ISO 1 640-3, an ex mples of sc ne-r fe r d

imag es ar pro ided in ISO 1 640-5 Howeve , at the time of the draf ing of this document, ve y few

scanne s an digital came as or raw proc s ing a plcations sup ort ed eithe original- or sc

ne-r fe r d encodings, making it ne es ary t o ada t output-r fe r d encodings t o this use

When ada ting output-r fe r d encodings for the purp se of digital ar hiving, seve al changes t o normal

practic should be made in the proc s ing, encoding, int erpr tation an displa of the imag e data:

a) When proc s ing the imag e data for encoding, any colour r n e ing should be turned of t o the

ext ent p s ible, so that the imag e colourimetry encoded ac urat ely r pr sent the colourimetry

of the original o je t, with chromatic ada tation t o the encoding w hit e point Particular att ention

should be paid t o proc s ing controls that a ply nonl near t one r prod ction, or black or w hit e

clp ing

b) If it is not p s ible t o turn of the colour r n e ing in the proc s ing, prof ilng should be used t o

un o it t o the ext ent p s ible, an ther sulting profile as igned t o the imag e, inst ead of the profile

normaly as ociat ed with the colour encoding used F or ex mple, if a came a ca tur s an A do e

RGB imag e, and the scan e or came a has be n profiled, w hich this document r commen s, the

prof ile as igned t o the imag e should be the a pro riat e scan e or came a profile rathe than the

Ado eRGB profile

It should be not ed that mos output-r fe r d RGB encodings use encoding 0 t o r pr sent the encoding

r fe enc medium black p int, as o p sed t o an a solut e black This can cause seve al pro lems:

— In some cases, conve ting t o the encoding could r sult in clp ing of t ones darke than the r fe enc

medium black p int, w hich should be avoided When original- or sc ne-r fe r d imag es ar st or d

using output-r fe r d encodings, the r fe enc medium should be ignor d, and encoding 0 should

be conside ed t o r pr sent a solut e black

— Likewise, in some cases conve ting from an output-r fe r d encoding could ca usea lght ening of the

dark t ones, if the code v lues ar int erpr t ed as r pr senting coloursa o e the encoding r fe enc

medium black, as o p sed t o a o e a solut e black

— When displa ing imag es, o e ating sy st ems an a pl cation sof war ma a ply black p int

compensation, w he e the encoding black p int is scaled t o the displa black p int This ma be

desira le in some cases, even with digital ar hive imag es, t o avoid clp ing of t ones darker than

thedispla black p int Howeve , it should be not ed that w hen black p int compensation is a pled,

thedark t ones wi be displa ed somew hat lght er and with lowe contras than they a pear on the

original Ap lcations with so his icat ed colour manag ement int erfac s ma ofe the o tion t o turn

black p int compensation on or of Howeve , it should also be not ed that many displa profiles set

thedispla black point t o 0, in w hich case turning black p int compensation of wi s i l not r sult

in cor e t r n ition of the dark t ones

Ho eful y in the futur sup ort for original- an sc ne-r fe r d colour encodings wi l be ome mor

widespr ad, a oiding mos of these is ues

4.9 Reproduction s ale

If a came a syst em with an ar a sensor is used the r prod ction scale depen s on the focal length as

wel as the o je t dis anc This ma ne d t o be adjust ed in the f inal imag e

5 Imag ing system qual ty anal ysis pr oc edur e

To det ermine the q alty of an ima ing sy st em ac ording t o this document one or mor t es chart

as desc ibed in the An ex es A an C ne ds t o be digitiz d with the sy st em that has be n set up an

calbrat ed ac ording t o the aspe t desc ibed in the pr vious p ra ra h

The digital imag e is then analy sed ac ording t o the in ivid al q alty aspe t mentioned in Clause 6

F or some of the measur ment proc d r s (e.g C lour r prod ction) r fe enc data for the t es chart is

r q ir d The ear comme cial y a aia le sof war t ools that can do the analy sis

Figure 1 — Schematic representation of ima ing sy stems qual ty analy sis procedure

In orde t o de ide w hethe or not the q alty of the sy st em is suff icient for the int en ed a plcation a

set of aims an t ole anc s is r q ir d An ex B desc ibes thr e t ole anc levels A, B an C, w hich can

be used or combined int o a cust omiz d l s for aimsan t ole anc s for a spe ific proje t or a plcation

6 Imag ing systems qual ty characteristics and metrics

The folowing ta les desc ibe the imag e charact eris ics that should be used t o analy se the q alty of

a digital imag e Each ta le pro ides a desc iption of the imag e charact eris ic, r fe enc s the r lat ed

s an ards used t o measur the charact eris ic, s at es a summary of the measuring method, an a

spe if ication of w hich t ech ical p tt erns of the r fe enc targ et should be used for the measur ment

The ta les also inclu e aim v lues for o timal ima ing syst ems q alty Annex B contains aims an

t ole anc s for prod cing ac urat e an ac epta le digital r prod ctions

The imag e charact eris ics ar grouped ac ording t o the b sic metric they r lat e t o: Tones an Noise

(t one r prod ction, g in mod lation, dynamic rang e, noise, b n ing, defe t pix els), C lour (w hit e

b lanc , colour r prod ction), Detai s ( amplng rat e, r solution, sharpening, acutanc ), and Geometry

( ilumination non-u iformity, colour mis-r gis ration, dist ortion, r prod ction scale)

Even though this document detai s a spe ific multi-p tt ern chart, these v lues can also be measur d

using the chart outlned in the dife ent s an ards, e.g in ISO 1 2 3, or othe chart that fulf il the

r q ir ment desc ibed in the s an ards

6.2 Tones and noise

Desc iption The o t o-ele tronic conve sion fu ction (OEC ) desc ibes the r lationship between

the input lght levels an the cor esp n ing digital output levels for an o t o-ele tronic

digital imag e ca tur sy st em To prod c an ac urate r prod ction it is important

that his curveis sha ed in a wa that he luminanc (CIE L*) levels of the original ar

trans e r d int o digital v lues that r pr sent the same L* v luesin the sele t ed colour

encoding of the imag e

Related s

and-ards

ISO 145 4 (came as)

ISO 2 5 0 ( can e s)

Measur ment Tone r prod ction/r sp nse curve (TRC)

Measur ment in L* v lues an t ole anc s for L* dife enc s between original an ca

-tur d imag e

Refe enc target Gr y scale ( pe tral y neutral) in a tes chart as def ined in Annex A that is idealy

eq aly sp c d in L* v lues

Aim The smale the deviation between the L* of the p t ches in the r fe enc targ et an the

L*v lues r pr sent ed b the digital code v lues the mor ac urat e the t one r prod c

-tion C r e t t onal ca tur is ex e t ed from L* 9 t o ≈ L* 5

Not es Also k own as t one r sp nse

Charact eris ic Gain mod latio

Desc iption The g in is a measur for the slo e of the OECF The g in mod lation desc ibes the

v riation of the slo e in an L* original versus L* output curve If the g in chang es sig

-nificantly it can ha pen that even with the t one r prod ction being within t ole anc s

the slo eis t oo flat t o dife entiat e t onal v lues that ar close t o each othe

Rela ted st an dar ds ISO 2 5 0

Measur ment C lculation of the inc emental g in from thet one r prod ction curve (TRC) is done

for gr y scale st eps with a t onal dife enc of at leas 1 L* v lues F or a gr y scale

with an inc ement betwe n the p t ches of 5 L* v lues eve y othe p t ch of the gr y

scale wi l be compar d This means the g in is calculat ed for p t ch i = 1 an 3, 2

Desc iption Noise is u want ed v riations in the r sp nse of an ima ing sy st em

It is introd c d int o the sy st em b the came a originating from the sensor, the elec

-tronics, or the imag eproc s ing Also r fe r d t o as visual noise

Re lated stand ar ds ISO 1 7 9 (came as)

ISO 2 5 0 ( canne s)

Measur ment Noise is measur d in u iform ar asof an imag ean ex r s ed as s an ard deviation

of the signal in these ar as ISO 1 7 9 desc ibes the method t o measur the visibi ty

of noise in imag es b sed on the h man visual sy st em This method shal be used t o

det ermine the noise for each gr y p t ch in the gr y scale To desc ibe the visibi ty

of noise the viewing con itions for the imag e ne d t o be defined The imag es shal be

viewed on a monit or in a viewing environment that satisfies ISO 3 64 an the g

eomet-ric viewing con itions shal be a 1 0 % viewing (1 pix el in the imag e eq als one pix el

on the sc e n) on a monit or with a r solution of 1 0 p i viewed from 0,5m dis anc

R efe enc targ et The gr y scale p t ches of the target def ined in An ex A ar used t o det ermine the

noise at v rying signal levels

Aim The ma im um noise level sh uld be low eno g so th t noise is not visible in the images

Not es Signal t o Noise Ratio ( SNR), ac ording t o ISO 1 7 9 an ISO 2 5 0, is not measur d

b this document

Charact eris ic Dynamic rang e

Desc iption Dynamic rang e in this case desc ibes the ratio of the bright es pat ch in the original

that is not clp ed an the darkes pat ch that is a o e thenoise ( SNR of 1)

Rela ted st an dar ds ISO 2 5 0, ISO 1 7 9

is the density of the

bright es p t ch w he e the output signal of the OECF a pears t o be u clp ed an d

ma

is the density of the p t ch with a signal t o noise ratio of1 ( e ISO 2 5 0 for detai s)

R efe enc targ et Gr y scale ( pe traly neutral) in a t es chart as defined in Annex A that is idealy

equaly sp c d in L* v lues with a contras exce ding that of the imag e ca tur devic

Aim d shal be as larg e as p s ible but at leas ex ce d the contras rang e of ty ical o je t

that ar digitiz d with the imag e ca tur devic

Not es

Charact eris ic Ban ing

Desc iption Unwant ed s ripes or b n s that oc ur in a digital imag e

Oft en times these s ructur s oc ur as a r gular p tt ern in an imag e The origin can

be a charact eris ic of the sensor, the p we sup ly or othe ele tronic influenc s on

the imag e ca tur Mis akes in the imag e proc s ing can also be a sour e for Ban ing

Rela ted st an dar ds u known

Measur ment A q antization of Banding is diff icult but a e a ing the columns of multiple l nes in a

u iform ar a an displa ing the digital v lues of the a e ag e l ne helps det ermining

p t ential Ban s A fr quency analy sis pe formed on that lne identifies r gular s ruc

-tur s b pro iding a peak at the fr q ency of the b n ing s ructur

R efe enc targ et To be measured on uniform grey, white, an bla ck s ripes in ch r as defined in An ex A

Aim The mean v alue for the columns of each of the s ripes shal be in the rang e of ex e t ed

fix ed p tt ern noise Ma and min v lues of the signal deviation for the averag ed lnes

should be def ined A ma v lue for a single fr q ency peak should be defined

Not es Ve tical an horizontal but ometimes also diag onal

Ave a ing seve al l nes or columns pe pen icular t o measur d dir ction

Charact eris ic Defect pix els

Desc iption Pix el or sub ix el that o e at es in a wa othe than the one in w hich it is driven

(ISO 9 41-3 2)

Rela ted st an dar ds ISO 9 41-3 2

Measur ment Due to the dife ent characte is ics of defe t pix els the measur ment ne ds t o be

pe formed on a black imag e, a gr y imag e, an a w hit e imag e F or the black imag e

the sensor doesnot g et ex osed t o lght d ring the imag e ca tur This can be done

b closing the lens ca or for scan e s b swit ching of the lght F r the gr y an

the w hit e imag e the sensor ne ds t o u iformly be ex osed t o lght so that the digital

output level r aches a v lue close t o the c ntr v lue of the digital output v luerang e

(e.g 1 7for an 8 bit imag e) for gr y an close t o the ma imum output level for w hit e

A rang e arou d the ty ical noise v lues ne ds t o be def ined for those pix els that ar

ok an al othe pix els out ide the rang e (outle s) ar then spe if ied as defe t

R efe enc targ et Uniform lght our eor targ et lke an int egrating sphe e or a u iformly i uminat ed

w hit e targ et

Aim Identification of single pixels or lust ers th t do not r pr sent the original r spe tively

ar out of the ex e t ed v lue rang e The smale the amou t ofthese pix els the bett er

Not es This ma not be r quir d for eve y ima ing sy st em ‘s qualty che k be a use it is mos ly

taken car of b the man factur r of the ca tur devic Only ifpixels oc ur th t sh w

pro lems this ma be ev luat ed Det ermining defe t pix els is not p rt of the r q ir d

measur ment for ima ing sy st emsq alty analy sis

Charact eris ic White balance

Desc iption A djus ment of ele tronic s i l pictur colour chan el g ins or imag e proc s ing so that

radiation with r lative spe tral p we dis ribution equal t o that of the sc ne i

umina-tion sour e is r n e ed asvisualy neutral ( e ISO 145 4 )

This ensur s that neutral gr y s in the original ar visualy neutral in theimag e

max

,ref ,r ef

The imag e ne ds t o be conve t ed from RGB t o L* a* b* using the colour encoding

de-sc iption of the colour encoding the imag e isin (lke sRG , Ado e RGB, … )

Refe enc targ et Gr y scale p t ches in the targ et def ined in Annex A

Aim The smale the w hit e b lanc the mor ac urat e the r prod ction Tole anc s t o be

def ined in ΔC*

Not es The CIE C* a pro ch has be n sele t ed o e an RGB a pro ch be a use CIE C* bett er

desc ibes the visibi ity of colour cas s

Charact eris ic Colo r repro uction

Desc iption The dife enc betwe n sele ted p ysicaly measur d inp t colo rs an their inten ed

output r n e ing for a given colour sp c

This ex r s es how wel an ima ing sy st em ca tur s an encodes colours In digital

came as colours ar ca tur d an encoded as combinations of R ed (R), Gr en (G) an

Blue (B) The colour encoding sele t ed desc ibes how these RGB v lues ar r lat ed

t o CIEL b v lues

Sinc none of the came as or scan ing sy stems mat ches the colour matching functions

of the h man visual sy st em, the colour r prod ction can s rictly speaking only be

o timal for a spe ific set ofcolo rs for which the system h s be n calbrated (profiled)

Cur ently the e ar only two set of commonly used colours t o profile scanne s an

came as One is the IT8 Targ et ac ording t o ISO 1 641 and the othe one is the C

l-ourChe ke SG ( Semi Glos surfac ) targ et Howeve the e ar othe targ et that ar

a ai a le an ma be used

Aft er profilng the sy st em b using these targ et the targ et ar scan ed a ain an

the colour dife enc is measur d using CIE ΔE 2 00 (CIE 0 5:2 04 ) a pro ch with

S = 1 (ΔE 2 00 SL ) w hich gives eq al weight t o the entir gr y scale The mean an

ma ΔE 2 0 v lues for al p t chesshal be det ermined

Rela ted st an dar ds CIE 1

Measur ment CIELAB colour dife enc measur (ΔE) ac ording CIE 1 spe s

R efe enc targ et C lour pat ches of Targ et def ined in An ex A

Aim The smale t he colo r diferenc s ΔE (L*a* b) th e m ore a ccura te th e colo r reprod uction

Not es This measur ment only pro ides useful information ifthe sy st em h s be n calbrat ed

an profied cor e tly for colo r reprod uction It does not pro ide a colo r reproduction

q alty es imation for othe than the chart colours an ma show inac urat e colour

r prod ction for c rtain ty es of originals

F or monit oring the sy st em with a t es ch rt othe th n the a o e-mentioned profilng

chart a r fe enc scan can be made using the chart desc ibed in An ex A right aft er

the ve if ication Each monit oring scan can then be comp r d t o ther fe enc scan

Desc iption C lour-t o-colour sp tial dislocation of othe wise sp tialy coincident colour featur s

of an imag ed o je t

Rela ted st an dar ds ISO 1 2 3 (r solution measur ment)

ISO 1 084 (chromatic displac ment)

Measur ment Det ermination of edg e location on a pe channel b sis for al slant ed edg es over the

ima ing f ield

R efe enc targ et Slant ed edg es in the t es chart def ined in Annex A

Aim Geometric dis anc betwe n the edg e location in the dife ent colour channels shal

be as smal asp s ible

Not es If a dislocation of the edg es in dife ent colour channels is u iform o e the field a

s an ard colour mis-r gis ration is pr sent If it v ries o e the f ield e.g from c ntr

t o corne s a chromatic a e ration is pr sent F or a lne scan e it ma be visible in jus

one dir ction F or a l ne scan ing sy st em the e ma also be a dislocation that comes

an g oes d e do inconsist ent motion

6.4 Detai s

Desc iption Samplng rat e in pix els pe u it of spac det ermined from ima ing a t es chart with

known g eometric s ructur s

Rela ted st an dar ds

Measur ment Det erminethen mbe of pix els for a block with a given g eometric siz in theimag e

of a t es chart an convert t o n mbe of pix elspe inch

R efe enc targ et Che kerb ard s ructur on Targ et def ined in Annex A

Aim The o tained samplng rat e should be as close as p s ible t o the claimed sampl ng

rat e pro ided in the metadata of the imag e file

Not es Also k own as samplng fr q ency

Samplng rat e should not be confused with l miting r solution

The samplng rat e of a digital r prod ction can be used t o calculat e the siz of the

phy sical r cord if st or d in an u compr s ed format

The sampl ng rat e lmit the ma imum p s ible r solution of an ima ing sy st em A c

-cording t o the Nyq is theor m, it is ne es ary t o ha e at leas two det ecting point

(pix els) on a cycle of a harmonic signal t o be a le t o r prod c the signal In othe

words, t o scan a black-an -w hit e t es s ructur , y u should ha e at leas one pix el for

the w hit e part an one pix el for the black p rt t o bea le t o r prod c the s ructur

Desc iption Measur of the a i ty of a came a sy stem, or a component of a came a syst em, t o

depict pictur detai ( e ISO 1 2 3)

Rela ted st an dar ds ISO 1 2 3, ISO 1 0 7-1

Measur ment Analy sis of the edg e spr ad fu ction in a slant ed edg e targ et Use the samplng fr

-q ency at 1 % mod lation thr shold for l miting r solution

R efe enc targ et The slant ed edg e s ructur s in the targ et defined in Annex A ar designed for SFR

( p tial fr q ency r sp nse)measur ment

Aim Rea ching a fr q ency as hig as p s ible b t not hig e th n Ny uis (to a oid alasing)

for the 1 % mod lation thr shold (l miting r solution) Depen ing on the samplng

rat e the ma r solution th t can be r ached v aries The r solution sh uld be cons ant

o e the f ield of ima ing an the dife enc in dife ent dir ctions (horizontal an

ve tical r spe tively slow scan an fas scan dir ction) should be as smal as p s ible

Not es Also k own as true o tical r solution

Not e that r solution measur ment b sed on slant ed edg eanaly sis r q ir s u

com-pr s ed and unsharpened data

The ratio between the lmiting r solution an the theor tical Nyq is l mit is based

on the o tained sampl ng rat e

Desc iption Ampl f ication of the sp tial fr q ency r sp nse b means of ima e proc s ing to

achieve sharpe a pearing imag es Also, a clas of imag e proc s ing o e ations that

enh nc s the contras of sele tive spatial fr q encies, usu ly visu ly imp rtant ones

Rela ted st an dar ds ISO 1 2 3, ISO 1 0 7-1

Measur ment Analy sis of the edg e spr ad function in a slant ed edg e targ et F or an imag e without

sharpening the SFR should at no fr q ency signif icantly ex ce d the v lue of 1

R efe enc targ et The slant ed edg e s ructur s in the Targ et def ined in se tion 6 ar designed for SFR

( p tial fr q ency r sp nse)measur ment

Aim The SFR should not significantly ex ce d the v lue of 1

Not es

Desc iption Measur of fr q ency b sed on the SFR measur ment w he e a 5 % contras level is

r ached ( e ISO 1 2 3) This is an in icat or for the sharpnes of an imag e

Rela ted st an dar ds ISO 1 2 3, ISO 1 06 -1

Measur ment Analy sis of the edg e spr ad fu ction in a slant ed edg e targ et Use the samplng fr

-q ency at 5 % mod lation thr shold as a sharpnes in icat or

R efe enc targ et The slant ed edg e s ructur s in the targ et def ined in An ex A ar designed for SFR

( p tial fr q ency r sp nse)measur ment

Aim Reaching a fr q ency as hig as p s ible b t not hig e th n Ny uis (to a oid alasing)

for the 5 % mod lation thr shold

Not es

Charact eris ic A cutance

Desc iption The visual pe c ption of sharpnes that desc ibes the q al ty of being c isp or of

containing detai

Rel ated s a nd a rd s

Measur ment The acutanc measur is the SFR weight ed with the contras sensitivity function

(CSF) of the h man eye for a given viewing condition

The used viewing condition in this spe ification shal be the 1 0 % viewing (e.g on

a comput er monit or)with a 1 0 p i pix el pit ch viewed from 0,5m dis anc

R efe enc targ et The slant ed edg e s ructur s in the targ et def ined in An ex A ar designed for SFR

( p tial fr q ency r sponse) measur ment

Aim The acutanc shal be as close asp s ible t o the ma v lue of 1

Not es The CSF shal be calculat ed an a pled as desc ibed in ISO 1 7 9 with the viewing

condition being a 1 0 % viewing on a monitor with 1 0 Ppi viewed f om a 0,5 m dis anc

Desc iption The i luminanc non-uniformity consis s oftwo comp nent :it depen s on h w evenly

the sc ne t o be ca tur d is i luminat ed, an on the degr e of shading introd c d b

the ima ing sy st em

Re lated st an dar ds ISO 1 9 7 ( hading measur ment)

Measur ment C ptur of a uniform w hit e original an analy se for luminanc shading ac ording t o

ISO 1 9 7an det ermine the L* v lue at at leas 1 2 0 point eq aly spr ad o e the

f ield of view an r p rt the ΔL* betwe n thema an the min L* v lue

R efe enc targ et Uniform w hit e pa e or de ived from che kerbo rd p tt ern in the targ et def ined in

Annex A on w hich only the pat ches with the same r fle tanc (e.g w hit e p t ches) ar

analy sed

Aim The i uminanc should be as u iform as p s ible

Not es Make sur no p rt in the imag e is clp ed Sur ou ding l ght should be a oided t o not

influenc the measur ment

Desc iption Depen ing on the ty e of devic , the infrar d cut-of filt er, sensor g eometry an angle

of incidenc the e can be a colour shif in the imag e that g et extr mely visible in case

of a uniform w hit e or gr y original

Rela ted st an dar ds ISO 1 9 7( hading measur ment)

Measur ment C ptur of a uniform white original an analyse for colour shading ac ording to

ISO 1 9 7 an det ermine the chrominanc deviation from the a e ag e for each block

at at leas 1 2 0 p int eq aly spr ad o e the field of view R ep rt the ma imum

v lue of the chrominanc deviation as the chrominanc non-u iformity D

C

R efe enc targ et Uniform w hit e p pe or de ived from che ke b ard patt ern in the targ et defined in

Annex A on which only the p t ches with the same r fle tanc (e.g w hit e patches)

ar analy sed

Aim The chrominanc non-uniformity should be as smal as pos ible

Not es Make sur no p rt in the imag e is clp ed Sur ou ding l ght should beavoided t o not

influenc the measur ment

Desc iption Displac ment from the ideal sha e of a subje t (lying on a plane paralel t o the imag e

plane) in the r corded ima e The dis ortion measur d can originate from thr e

dife ent aspe t : lens g eometric dist ortion, scanne motion dist ortion, imag e

pro-c s ing dist ortion

Rela ted st an dar ds ISO 1 8 0

Measur ment F or a r gular grid of dot , c os es or lne int erse tions the locations ar ev luat ed on

a sub pix el ac uracy b sis These locations ar comp r d t o a r gular grid g ene at ed

from the c ntral ar a of the imag e

R efe enc targ et The chart can be a dedicat ed dot or c os chart or the chart as spe if ied in Annex A

Aim The dist ortion shal be as smal as p s ible

Not es Dist ortion can usual y be charact eriz d an compensat ed Ev luation ma be done

with activ t ed compensation if sy st em alow s for such compensation

Desc iption Describes h w far a given geomet ic dis anc in the original is represented in the image

Rela ted st an dar ds

Measur ment Is r fle t ed in the claimed ve sus o tained sampl ng rat e

R efe enc targ et De ived from che kerbo rd p tt ern in the targ et def ined in An ex A

Aim Should be as close as pos ible t o the original Tole anc s ar proje t depen ent

Not es The r prod ction scale r lat es t o dimensional ac uracy

7 Reporting results

Tog ethe with the ls of r sult the folowing information shal be r p rt ed:

— Came a scanne man factur r, model an se ial n mbe

— Lens man factur r, model an se ial n mbe ( if a plca le)

— Lighting ( if a plca le)

— Dat e, Time, L cation of the measur ment

— Used imag e proc s ing sof war (Name an ve sion)

— Came a set ings that imp ct the ima ing q al ty

— Inclu ing a e tur , sh tt er spe d, ISO ( ensitivity spe d)

— Important imag e data

— inclu ing imag e width an imag e height, claimed samplng rat e, bit pe sample (bit depth),

colour encoding, colour prof ile

7.2 Example r eport for tone reproduction r esults

Based on the t ole anc rang e for the in ivid al a plcation, the L* v lues extract ed from the imag e of

the t es chart shal be r port ed a ains the L* v lues of the original In case the r sult ar pr sent ed

as a ta le the dife enc s betwe n the original and the imag e shal be s at ed in a third column an the

minimum an ma imum dife enc s at the b tt om of the ta le show if the v lues ar within or out of

the t ole anc s Anothe o tion for using the ma an min v lues is the colour d in ication of v lues

that ar out of the def ined t ole anc s, se Ta le 1

Table 1 — Max and min values ne d to be in the given rang e based o the tolerances for the

In case of the pr sentation in a gra h the t ole anc s shal clearly be in icat ed b sep rat e l nes or

marking the ar a of v lues within t ole anc

Figure 2 — Graphical representatio for the tone reproductio

7.3 Gain modulation

Ta le 2 with the r sult for g in mod lation shal contain eithe the p ir of the p t ches for w hich the

g in mod lation was calculat ed or the a e ag e L* v lue ofthe two comp r d p t ches in the f irs column

an the calculat ed g in mod lation in the se on column

Table 2 — Gain modulatio

(ma an min v lues ne d t o be in the spe ified t ole anc level)

Pat ch Gain Mod lation

In case of a gra hical pr sentation the in ication of the up e an lowe t ole anc level is helpful

Figure 3 — Graphical representation of g in mo ulatio

7.4 Dynamic rang e

Dynamic rang e is r port ed as a single v lue in densities lke 3,0 densities

Alt ernatively thedynamic rang e can be ex r s ed in:

— a contras rang e lke e.g 1 0 0:1 (1

The noise v lue is r p rt ed as visual noise v lue ac ording t o ISO 1 7 9 for the viewing con itions

desc ibed in 6.1 o e the L* v lue of the in ivid al gr y p t ches

Table 2 (c ontin ed)

Table 3 — V isual noise value in table form

( he ma v lue should be below the spe if ied t ole anc level)

7.6 Banding

F or Ban ing the gra hical r pr sentation can be the digital v lues o e pix els

Figure 5 — Digital values ove Pix els Regular pe ks in this structure indicate the presence of

banding in imag es