Re: [Lcms-user] sRGB with unintended D50 whitepoint
Have you tried the ICC verification tool on those profiles? It can detect any deviation from standard... http://color.org/profdump.xalter I've tried it and the profiles are just checked ok. Then it is just a matter of guessing; you could modify lcms source code and do some trial and error. But it is not easy to fix in the CMM side, since apparently there is no issue to fix at all. Let me know if there is anything I could do. Regards Marti El 25/02/2013 19:09, Pascal de Bruijn escribió: On Tue, Oct 16, 2012 at 1:43 AM, Graeme Gill gra...@argyllcms.com wrote: Pascal de Bruijn wrote: Another peculiar fact is that Argyll's iccdump reports this: tag 0: sig 'desc' type 'desc' offset 288 size 140 Unable to read: 1, icmTextDescription_read: ScriptCode string too long You get that if the Mac ScriptCode string is 67 bytes. From the ICC V2.4 spec, section 6.5.17: The localized Macintosh profile description contains 67 bytes of data, of which at most count bytes contain a ScriptCode string, including a null terminator. The count cannot be greater than 67. Thanks. We're still getting reports of print services (saal digital) ignoring our embedded profiles. While we haven't gotten any concrete information as to why this may be happening, one difference are these ScriptCode strings. So I'm wondering if we can force lcms2 to not embed these strings, so we can ask a user to give it a try? Regards, Pascal de Bruijn -- Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user -- Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
On Tue, Feb 26, 2013 at 5:01 PM, Marti Maria marti.ma...@littlecms.com wrote: Have you tried the ICC verification tool on those profiles? It can detect any deviation from standard... http://color.org/profdump.xalter Oh right, I didn't know there was one freely available. Silly me. So I checked, and had a bit of a *headdesk* moment... We decided not copyright our embedded profiles, so we left out the cprt tag... which apparently is required. DOH! So I added cprt, now I have to wait for the user to test of course :) I've tried it and the profiles are just checked ok. Then it is just a matter of guessing; you could modify lcms source code and do some trial and error. But it is not easy to fix in the CMM side, since apparently there is no issue to fix at all. Possibly. The problem with troubleshooting an issue with a users vendor is the long communication line, assuming the vendor supplies any useful information at all. And even if the profile is 100% spec compliant, there may still be a case of weird/quirky soft/hardware. Thanks for your time once again. Regards, Pascal de Bruijn PS: I'd like to extend my apologies for the recent incident. -- Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
On Tue, Oct 16, 2012 at 1:43 AM, Graeme Gill gra...@argyllcms.com wrote: Pascal de Bruijn wrote: Another peculiar fact is that Argyll's iccdump reports this: tag 0: sig 'desc' type 'desc' offset 288 size 140 Unable to read: 1, icmTextDescription_read: ScriptCode string too long You get that if the Mac ScriptCode string is 67 bytes. From the ICC V2.4 spec, section 6.5.17: The localized Macintosh profile description contains 67 bytes of data, of which at most count bytes contain a ScriptCode string, including a null terminator. The count cannot be greater than 67. Thanks. We're still getting reports of print services (saal digital) ignoring our embedded profiles. While we haven't gotten any concrete information as to why this may be happening, one difference are these ScriptCode strings. So I'm wondering if we can force lcms2 to not embed these strings, so we can ask a user to give it a try? Regards, Pascal de Bruijn -- Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
Am 25.02.2013 19:09, schrieb Pascal de Bruijn: On Tue, Oct 16, 2012 at 1:43 AM, Graeme Gill gra...@argyllcms.com wrote: Pascal de Bruijn wrote: Another peculiar fact is that Argyll's iccdump reports this: tag 0: sig 'desc' type 'desc' offset 288 size 140 Unable to read: 1, icmTextDescription_read: ScriptCode string too long You get that if the Mac ScriptCode string is 67 bytes. From the ICC V2.4 spec, section 6.5.17: The localized Macintosh profile description contains 67 bytes of data, of which at most count bytes contain a ScriptCode string, including a null terminator. The count cannot be greater than 67. Thanks. We're still getting reports of print services (saal digital) ignoring our embedded profiles. While we haven't gotten any concrete information as to why this may be happening, one difference are these ScriptCode strings. So I'm wondering if we can force lcms2 to not embed these strings, so we can ask a user to give it a try? It is usual to use low level tools by people who know what they do for non standard editings. Tried IccXML? kind regards Kai-Uwe -- Everyone hates slow websites. So do we. Make your web apps faster with AppDynamics Download AppDynamics Lite for free today: http://p.sf.net/sfu/appdyn_d2d_feb ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
On Wed, Sep 12, 2012 at 11:00 PM, Marti Maria marti.ma...@littlecms.com wrote: Ideally, I'd love to match the canonical Microsoft HP profile. Would that be possible using LCMS2? Sure, just create a plain sRGB and overwrite the media white point with D65. I'm not entirely sure how to go about that. As the whitepoint is set at profile creation, I'm only switching to v2 after the whitepoint has been set. Would you have any code snippets? Also, that won't get rid of the chad tag will it? But as said in my other mail that is not going to recover the old functionality. Right. Our point is to remain compatible with old and antiquated CMS's, as the new style profiles seem to break Windows XP's Windows Photo Viewer application. And you'd be surprised how commonly used that still is. And we've had reports from users who sent for prints to saal digital, which rejected our tagged files. Another peculiar fact is that Argyll's iccdump reports this: tag 0: sig 'desc' type 'desc' offset 288 size 140 Unable to read: 1, icmTextDescription_read: ScriptCode string too long I've attached the ICC we're generating, I'd appreciated any help/hints to resolve this issue. Regards, Pascal de Bruijn dt_srgb.icc Description: Binary data -- Don't let slow site performance ruin your business. Deploy New Relic APM Deploy New Relic app performance management and know exactly what is happening inside your Ruby, Python, PHP, Java, and .NET app Try New Relic at no cost today and get our sweet Data Nerd shirt too! http://p.sf.net/sfu/newrelic-dev2dev___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
On Wed, Sep 12, 2012 at 2:26 AM, Graeme Gill gra...@argyllcms.com wrote: Marti Maria wrote: Not according ICC, see http://www.color.org/srgbprofiles.xalter#v2 They require D50 for media white point to avoid the wrong Von-Kries issue. Hi, They may do it that way with their profiles, but nothing in the V2 specs requires it, the canonical Microsoft HP sRGB profile doesn't do it that way, and doing it that way wrecks absolute colorimetric and proofing uses. Ideally, I'd love to match the canonical Microsoft HP profile. Would that be possible using LCMS2? Regards, Pascal de Bruijn -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
On Tue, Sep 11, 2012 at 11:51 PM, Pascal de Bruijn pmjdebru...@gmail.com wrote: On Mon, Sep 10, 2012 at 10:28 PM, Marti Maria marti.ma...@littlecms.com wrote: Hi, Apparently Firefox only handles v2 by default (v4 can be enabled in settings) Unfortunately Firefox only handles a very limited set, matrix-shaper only. Either V2 and specially V4 are largely unsupported by Firefox. See this test page on ICC site: http://www.color.org/browsertest.xalter Try with IE9 and then with Firefox Well, with v4 enabled, that works just fine... but v4 is disabled by default at this time. Anyway, creating a V2 sRGB with lcms2 is easy: hProfile = cmsCreate_sRGBProfile() cmsSetProfileVersion(hProfile, 2.1); cmsSaveProfileToFile(hProfile, srgb_v2.icc); cmsCloseProfile(hProfile); Yes, this works :-) Ah, thank you. Hmmr... It seems that creates a v2 profile, but still with a D50 whitepoint encoded (instead of the traditional D65, for v2 sRGB). I know for v4 the white point is adapted using a different tag, but that tag shouldn't end up in v2 profiles right? Regards, Pascal de Bruijn -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
Unfortunately Firefox only handles a very limited set, matrix-shaper only. Either V2 and specially V4 are largely unsupported by Firefox. See this test page on ICC site: http://www.color.org/browsertest.xalter Try with IE9 and then with Firefox Well, with v4 enabled, that works just fine... but v4 is disabled by default at this time. That would be great news... but tried 15.0.1 and still it doesn't work. Next release maybe? Anyway this has nothing to do with V4, is just a plain and common V2 LUT profile. Regards Marti -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
Hmmr... It seems that creates a v2 profile, but still with a D50 whitepoint encoded (instead of the traditional D65, for v2 sRGB). I know for v4 the white point is adapted using a different tag, but that tag shouldn't end up in v2 profiles right? Hi, Not according ICC, see http://www.color.org/srgbprofiles.xalter#v2 They require D50 for media white point to avoid the wrong Von-Kries issue. But anyway, the only difference is in the absolute colorimetric intent, and the old way was giving a huge blue cast that I doubt anybody would want. Regards Marti -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
Marti Maria wrote: Not according ICC, see http://www.color.org/srgbprofiles.xalter#v2 They require D50 for media white point to avoid the wrong Von-Kries issue. Hi, They may do it that way with their profiles, but nothing in the V2 specs requires it, the canonical Microsoft HP sRGB profile doesn't do it that way, and doing it that way wrecks absolute colorimetric and proofing uses. Graeme Gill. -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
On Mon, Jun 25, 2012 at 4:18 AM, Graeme Gill gra...@argyllcms.com wrote: marti.ma...@littlecms.com wrote: Hi Marti, Yes, as Kai-Uwe says, this is the expected behavior for a V4 CMM, and that is what lcms2 is. Please note most if not all modern CMMs behaves in such way, for example all Adobe products. Right, but V2 profiles should be processed according to the V2 specifications. Right, and there are some real-world practical issues with using V4 to being with. Apparently Firefox only handles v2 by default (v4 can be enabled in settings). Is there some sample code, to generate a v2 sRGB profile with LCMS2? Regards, Pascal de Bruijn -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
Am 23.06.12, 20:51 +0200 schrieb Pascal de Bruijn: We've recently discovered that Darktable's (a photography application) internal sRGB profile seems to encode a D50 whitepoint even though we are passing a D65 whitepoint. The current theory is that this changed when we migrated from lcms1 to lcms2. The whitepoint goes into the chromatic adaptation tag. D50 is defined for v4 profiles in the spec. Look for some examples on the official ICC website: http://www.color.org/srgbprofiles.xalter Is there any way we could stick with ICCv2 (for greater compatibility)? The compatibility argument applies to what specific CMM? Most CMMs support v4 profiles after all these years. Picking a on disc ICC profile is no option for darktable? kind regards Kai-Uwe -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
Hi Pascal, We've recently discovered that Darktable's (a photography application) internal sRGB profile seems to encode a D50 whitepoint even though we are passing a D65 whitepoint. The current theory is that this changed when we migrated from lcms1 to lcms2. Yes, as Kai-Uwe says, this is the expected behavior for a V4 CMM, and that is what lcms2 is. Please note most if not all modern CMMs behaves in such way, for example all Adobe products. The difference is only in the absolute colorimetric intent. It is fully described in this paper from ICC: http://color.org/ICC_white_paper_6_v2_and_v4_display_profile_differences.pdf If anyway, you prefer the old behavior despite that would made your application inconsistent, you could use cmsSetAdaptationState(0) at the init time. Also, we noticed lcms2 seems to default to ICCv4: Yes, v4 is from 2001 (11 years right now!) Actually most vendors are using modern versions. Is there any way we could stick with ICCv2 (for greater compatibility)? Well, lcms2 can read and use v2 profiles, no problem on that. If you want to save profiles as v2, this is as easy as calling cmsSetProfileVersion(hICC, 3.4) before saving the profile. On depending on the version number the tags are saved by using different structures. But again, Kai-Uwe is right. Check the applications you want to be compatible. Photoshop for example, which is regarded as the industry standard works in the V4 way. Best regards Marti -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
Re: [Lcms-user] sRGB with unintended D50 whitepoint
marti.ma...@littlecms.com wrote: Hi Marti, Yes, as Kai-Uwe says, this is the expected behavior for a V4 CMM, and that is what lcms2 is. Please note most if not all modern CMMs behaves in such way, for example all Adobe products. Right, but V2 profiles should be processed according to the V2 specifications. Graeme Gill. -- Live Security Virtual Conference Exclusive live event will cover all the ways today's security and threat landscape has changed and how IT managers can respond. Discussions will include endpoint security, mobile security and the latest in malware threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/ ___ Lcms-user mailing list Lcms-user@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/lcms-user
[Lcms-user] sRGB with unintended D50 whitepoint
Hi,
We've recently discovered that Darktable's (a photography application)
internal sRGB profile seems to encode a D50 whitepoint even though we
are passing a D65 whitepoint. The current theory is that this changed
when we migrated from lcms1 to lcms2.
This is our code:
https://github.com/darktable-org/darktable/blob/master/src/common/colorspaces.c#L189
So the big question is why lcms2 seems to be ignoring the whitepoint
we pass along? Are we doing something wrong?
Could this bit of code in cmsvirt.c explain it:
if (WhitePoint) {
if (!cmsWriteTag(hICC, cmsSigMediaWhitePointTag,
cmsD50_XYZ())) goto Error;
cmsxyY2XYZ(WhitePointXYZ, WhitePoint);
_cmsAdaptationMatrix(CHAD, NULL, WhitePointXYZ, cmsD50_XYZ());
// This is a V4 tag, but many CMM does read and understand it
no matter which version
if (!cmsWriteTag(hICC, cmsSigChromaticAdaptationTag, (void*)
CHAD)) goto Error;
}
Is D65 encoded into a different tag, and the old tag is defaulted to D50?
Also, we noticed lcms2 seems to default to ICCv4:
cmsSetProfileVersion(hICC, 4.3);
Is there any way we could stick with ICCv2 (for greater compatibility)?
Regards,
Pascal de Bruijn
dt_srgb.icc
Description: Binary data
--
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and
threat landscape has changed and how IT managers can respond. Discussions
will include endpoint security, mobile security and the latest in malware
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/___
Lcms-user mailing list
Lcms-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/lcms-user
