View Issue Details

IDProjectCategoryView StatusLast Update
0006583ScribusColor Managementpublic2007-12-17 00:29
Reporteroberten Assigned Tofschmid  
PrioritynormalSeverityminorReproducibilityalways
Status closedResolutionfixed 
PlatformLinuxOSUbuntuOS Version7.10
Product Version1.3.5svn 
Fixed in Version1.3.5svn 
Summary0006583: small inconsistency in the colour list
DescriptionThe colour list shows 3 columns: rgb/cmyk, spot/reg, colour name. The first two should be merged into one: spot and registration are neither rgb nor cmyk, and rgb or cmyk colours are neither spot nor registration... By the way, when you define a spot colour with rgb coordinates, it still shows as cmyk.
TagsNo tags attached.
Patch

Activities

christoph_s

2007-12-15 15:56

administrator   ~0018303

Tested, fixed. Thanks.

jghali

2007-12-15 19:50

administrator   ~0018306

Last edited: 2007-12-15 20:08

I have no problem with showing a rgb icon when a spot color is defined with rgb coordinates. I think it's the right thing to do. However i *strongly* disagree with merging spot with the rgb and cmyk icons. Even if spot colors are a specific kind of color, they have coordinates defined in a specific color space, rgb or cmyk currently. This colorspace is important as soon as user choose to convert spot color to process colors as we give him the possibility in pdf export dialog. A rgb spot color can give substancially different result from a cmyk spot color when converted to a cmyk process color. Spot colors are also called named color : they are just a color + a name. Or a color with a specific flag if you prefer. And they are still defined using rgb, cmyk, or even Lab coordinates...

avox

2007-12-15 20:13

administrator   ~0018307

jghali's reasoning sounds fine to me.

So we stay at three columns: colorspace, flags, name?

Flag "spot" means that the colorspace is the alternate colorspace.
Flag "register" should be combined with the CMYK colorspace (never heard of RGB separations). Or we have a "no colorspace".
What about a new flag for our default colors (CMYK black, CMYK white and None)?
Maybe an 'x' as icon?

oberten

2007-12-15 22:13

reporter   ~0018308

While I was reporting this, I was wondering what the point is to show in the list if the colours have been defined as CMYK or RGB (or Lab when it will be available). As they will eventually be converted to the same colour space anyway (except for PDF/X but this isn't a final format), that colour space being meant for screen (various flavours of rgb) or for print (cmyk/spot colours/both/part of it/hexachrome/opaltone/...).

jghali

2007-12-16 00:02

administrator   ~0018310

>>> As they will eventually be converted to the same colour space anyway (except for PDF/X but this isn't a final format)

That's the case for now, won't be the case in the future. At a point the option to export "as is" will be added. So color will retain their original colorspace, and it will be possible to have a mix of rgb/cmyk/lab colors.

oberten

2007-12-16 00:34

reporter   ~0018311

Last edited: 2007-12-16 00:36

Of course, in a work (sla) or exchange(pdf/x) format it must keep the "original" coordinates, just like it works now, that doesn't mean the type of coordinates needs to be shown in the colour list. It's enough that it shows when you create or edit the colour.

>>> At a point the option to export "as is" will be added.

As far as I know, a software like Scribus is meant for 2 medias only: video or print. When I speak about final format, I mean that for video (projection or screen) all colours will be converted to rgb in order to be sent to the display, whatever original coordinates they have. It's the same for print, whatever the colour format was, everything is sent to the machine with its inks model.

So for the final result, we just need to know if the colour will be printed directly or with the combination of different inks.

If you want to get a look at the final result on your screen, you'll have to use colour management. For rgb as well as cmyk. There are a lot of different profiles for cmyk that give very different results so cmyk is not more reliable as rgb for spot conversion... that's why spot colour exist...

[edit] And it's also why spot colours are usually defined with Lab coordinates... [/edit]

jghali

2007-12-16 02:37

administrator   ~0018312

>> that doesn't mean the type of coordinates needs to be shown in the colour list. It's enough that it shows when you create or edit the colour.

When working and importing files it's easy to end up with spots colors which will look similar on screen but use different color model. Having the ability to directly determine the color model is absolutely *crucial* to decide of eventual color values/model correction and avoid costly errors. For example we regularly see cases where a job was initially planned to print with a spot color, and finally get printed on a cmyk press. Needless the say that the layout work is not fully redone. In such situation, it is absolutely vital to avoid having spots being defined using different color models in order to maintain color consistency of the job. And to detect such situation on the layout side, the type of coordinates has to be immediately visible , even for a spot.

>> It's the same for print, whatever the colour format was, everything is sent to the machine with its inks model.

I would better say "everything ends up being sent to the machine with its inks model". In modern worflows, colors are send "as is" and converted only at the printers by the rip.

>> There are a lot of different profiles for cmyk that give very different results so cmyk is not more reliable as rgb for spot conversion... that's why spot colour exist...

Recent PANTONES libraries uses colors defined using calibrated RGB which provide also direct conversion to Lab. Knowing those values, converting a spot color to a specific colorspace be it rgb-based or cmyk-based is quite straightforward knowing the corresponding profile. If the spot color is reproductible in the destination colorspace, the result will match the spot color. I have already done such conversions knowing the spot color and the press profile with quite good results.

Now why spot color exists:
1) because printing using 2 colors (eg: black + spot) is less expensive than printing CMYK
2) because a press using 2 groups is far less expensive than a press using 4 groups
3) better color reproductibility
4) better gamut range
The color reproductibility argument is secondary and could be easily discussed as soon as a spot is used with a percentage < 100%.

>> And it's also why spot colours are usually defined with Lab coordinates...

That's quite recent in fact... That has not been always the case and this is in fact quite rare according to analysis performed on all file we receive (several hundreds / days). In most of the cases, spot colors are re-defined using cmyk values on export. Softwares such a QXP or Indesign keep only Lab values for spots when colors are exported "As Is" or some specific option is enabled...

jghali

2007-12-16 03:07

administrator   ~0018313

>> If you want to get a look at the final result on your screen, you'll have to use colour management. For rgb as well as cmyk. [...] that's why spot colour exist...

I was about to forget a point. Spot colors have always been known to provide quite bad preview of final result on screen:
1) When based on cmyk values, preview accurracy is generally horrible as soon as the real spot color is outside of a standard press gamut
2) When based on rgb values, better, but see bellow
3) When based on lab values, software provide generally no way to specify press dot gain or a specific color profile, in which case accuracy of preview is quite limited as soon as ink coverage is different from 100%

oberten

2007-12-16 09:38

reporter   ~0018315

Thanks for the explanation. I'm certainly too far away from the real world of printing. My thoughts where based on colour concepts rather than reality...

jghali

2007-12-16 19:45

administrator   ~0018323

I reverted the rgb/cmyk/spot icon merging. However spot colour defined with rgb coordinates now show are rgb.

Issue History

Date Modified Username Field Change
2007-12-14 23:08 oberten New Issue
2007-12-15 12:00 fschmid Status new => resolved
2007-12-15 12:00 fschmid Fixed in Version => 1.3.5svn
2007-12-15 12:00 fschmid Resolution open => fixed
2007-12-15 12:00 fschmid Assigned To => fschmid
2007-12-15 15:56 christoph_s Status resolved => closed
2007-12-15 15:56 christoph_s Note Added: 0018303
2007-12-15 19:50 jghali Status closed => feedback
2007-12-15 19:50 jghali Resolution fixed => reopened
2007-12-15 19:50 jghali Note Added: 0018306
2007-12-15 19:54 jghali Note Edited: 0018306
2007-12-15 20:08 jghali Note Edited: 0018306
2007-12-15 20:08 jghali Note Edited: 0018306
2007-12-15 20:13 avox Note Added: 0018307
2007-12-15 22:13 oberten Note Added: 0018308
2007-12-16 00:02 jghali Note Added: 0018310
2007-12-16 00:34 oberten Note Added: 0018311
2007-12-16 00:36 oberten Note Edited: 0018311
2007-12-16 02:37 jghali Note Added: 0018312
2007-12-16 03:07 jghali Note Added: 0018313
2007-12-16 09:38 oberten Note Added: 0018315
2007-12-16 19:45 jghali Status feedback => resolved
2007-12-16 19:45 jghali Resolution reopened => fixed
2007-12-16 19:45 jghali Note Added: 0018323
2007-12-17 00:29 christoph_s Status resolved => closed