Thread: [Lcms-user] Re: convert ASCII to devicelink-profile
An ICC-based CMM for color management
Brought to you by:
mm2
From: ignacio r. de c. <ign...@bp...> - 2005-06-16 08:58:27
|
Jan-Peter: Others in the list have addressed your first question below. Regarding the second one (how to convert imagepixel to ASCII) I have two different answers: 1) You can always use Gretag's freeware tool ColorLab to open an image and save it as text. 2) I wrote two tools, TIFFtoIT8 and IT8toTIFF, that allow the conversions between pixels and ASCII and viceversa. They make use of Marti's CGATS parser and TIFFLIB. Although they work for me, they are limited in scope (address only my needs: RGB, LAB, CMYK, Separated TIFFs, image offsets, pixel replication) Should Marti consider a Contrib folder that would gather tools and (acceptable) pieces of code writen with lcms? I wouldn't mind sharing mine... Regards, Ignacio >My question: >- Is there code for converting ASCII to a devicelink-profile ? >- Is there code for converting imagepixel to ASCII ? >- can we integrate this in lcms ? > >greetings from the colorspace >:-) Jan-Peter > >-- > --=20 ______________________________________________________________ Ignacio Ruiz de Conejo Tel. +34 93.5821342 Color & Imaging Group, R&D Fax +34 93.5824816=20 Hewlett-Packard Espa=F1ola, ICD e-mail: ign...@bp... |
From: Hal V E. <hv...@as...> - 2005-06-16 20:02:50
|
On Thursday 16 June 2005 03:57 am, Jan-Peter Homann wrote: > (If a printer-driver supports devicelink-profiles, user could e.g.print > a testimage, apply the coloradjustments, they want, and print later > every job with this coloradjustment) > > Use your imaging application as devicelink-profile editor > ------------ > > Often an ICC-conversion delivers not exactly the result, which users > expect. But with some coloradjustment, it would be nearly perfect. > > A way to solve this: > 1. Link the ICC-profiles together to an devicelink-profile > 2. Transform an IT-8 like testimage with the devicelink-profile > 3. Edit the transformed testimage in your imaging application > 4. Link the edited values with the original IT-8 values to a new > devicelink-profile. > 5. Apply this devicelink-profile instead your normal ICC-transformation. > > Imagine e.g. the next version of gutenprint comes with lcms-support and > some standard-icc profiles. The results for your printer are good, but > not optimal. > The described way is easy to go, without the need to buy a > spectrophotmeter and a profiling software. > > So this technology makes sense for all color-aware users of opensource > applications and not only for graphic arts color geeks. I am not too sure that just having a profile editor and some canned profiles is enough. I use non-standard archival inks in my printers and in cases like this canned profiles for the printer and standard inks will not be close enough to give a good starting point for the profile editor. In Windows I have been using ProfilePrism. This software uses a scanner as it's "spectrophotometer". Of course we all know that scanners are not particularly good at this. The software is setup to scan both the test printer target and an IT8.7 target at the same time so this allows it to "calibrate" the scanned result. In addition newer versions have an editor that allows users to manipulate profiles. This editor is designed to use an iterative approach to do the modifications to the profile. The editor presents the user with two views of the test image and in the editor the user will adjust one of the images to look like the printout of the test image. After the user thinks the image is corrected the software will generate a new profile with the changes needed to correct it in the direction indicated by the user. Then the test image is printed with the new profile and used to make additional adjustment to the image to fine tune the result. My experience is that the earlier versions of the software did not give particularly good profiles and that it works better with some scanners then it does with others. With the current version and using a scanner it likes (an inexpensive Canon LIDe 20) I am able to get very good printer profiles without even using the profile editor. But it is very nice to have the editor available to make minor adjustments to my profiles if needed. In the long run I am hoping that something like this will become available as an open source project. I believe that argyllcms is setup to do printer profiles but currently needs a spectrophotometer and is not setup to use a scanner. Nor does it have any type of profile editor. And of course it is a command line only tool at this time. CUPS 1.2 beta is now available. This version of CUPS was supposed to be CM aware (I have not tested it). In addition it appears that the GutenPrint project is also looking into what they need to do in this area. We are getting closer to having open source CM aware printing subsystems. So it is becoming important that users have ways to create good printer profiles. I agree with Jan-Peter that most users other than those that are true color geeks will not have a spectrophotometer on hand to be able to create these profiles. Therefore I think that a printer profiling application that was designed to work with a scanner and has a profile editor is a good compromise for those who want more control than just using a canned profile but who can not justify the cost of a spectrophotometer. Hal |
From: Gerhard F. <nos...@gm...> - 2005-06-16 22:09:06
|
Hal V Engel schrieb: >In the long run I am hoping that something like this will become available as >an open source project. I believe that argyllcms is setup to do printer >profiles but currently needs a spectrophotometer and is not setup to use a >scanner. > Argyll actually does support the use of a scanner as poor man's colorimeter for printer profiling. But due to the well-known reasons, I never got really good results (using an epson scanner with a CCFL lamp, and an IT8.7 target as color reference). Regards Gerhard |
From: Hal V E. <hv...@as...> - 2005-06-17 00:42:30
|
On Thursday 16 June 2005 03:08 pm, Gerhard Fuernkranz wrote: > Hal V Engel schrieb: > >In the long run I am hoping that something like this will become available > > as an open source project. I believe that argyllcms is setup to do > > printer profiles but currently needs a spectrophotometer and is not setup > > to use a scanner. > > Argyll actually does support the use of a scanner as poor man's > colorimeter for printer profiling. But due to the well-known reasons, I > never got really good results (using an epson scanner with a CCFL lamp, > and an IT8.7 target as color reference). > > Regards > Gerhard Cool I will have a look at it. My experience is that scanners with florescent lamps, which is most scanners, will have a significant amount of metamerism with some types of inks. The problem is particularly pronounced with pigment based inks like the archival inks I use but is much less of a factor with dye based inks. The problem is that the ink in the printer target and the IT8.7 target will exhibit different amounts of metamerism which results in the printer target and the IT8.7 target shifting by different amounts and in different directions. The effect varies for different inks, papers, lamps... and I suspect is almost impossible to compensate for with so many variables. The Canon LIDe 20 (and some other scanners) use three light emitting diodes (one each red, green and blue) that results in a light spectrum that is very close to sun light. This minimizes the affects of metamerism. I have never been able to get good results using a scanner with a florescent lamp with the inks that I use. Others on the ProfilePrism forums report getting good results using dye inks and scanners with florescent lamps. |
From: Gerhard F. <nos...@gm...> - 2005-06-17 07:14:27
|
Hal V Engel schrieb: >The Canon LIDe 20 (and some other scanners) use three light emitting diodes >(one each red, green and blue) that results in a light spectrum that is very >close to sun light. > Actually I don't believe that. LEDs are basically know to be narrow-band light sources, so I'm wondering, whether the LIDEs actually capture the full spectrum w/o holes. Anyway - only the result counts ... -Gerhard |
From: Gerhard F. <nos...@gm...> - 2005-06-17 09:57:30
|
Hal, just FYI, see http://www.argyllcms.com/doc/Scenarios.html#PP1 http://www.argyllcms.com/doc/Scenarios.html#PP4 Regards, Gerhard -- Weitersagen: GMX DSL-Flatrates mit Tempo-Garantie! Ab 4,99 Euro/Monat: http://www.gmx.net/de/go/dsl |
From: Bob F. <bfr...@si...> - 2005-06-18 15:27:56
|
On Fri, 17 Jun 2005, Gerhard Fuernkranz wrote: > Hal V Engel schrieb: > >> The Canon LIDe 20 (and some other scanners) use three light emitting >> diodes (one each red, green and blue) that results in a light spectrum >> that is very close to sun light. >> > Actually I don't believe that. LEDs are basically know to be narrow-band > light sources, so I'm wondering, whether the LIDEs actually capture the full > spectrum w/o holes. Anyway - only the result counts ... This is indeed an interesting issue. At first glance it seemed like nonsense to me, but after some more thought it seems that if the color of the LEDs are indeed narrow-band light sources and their colors happen to be located close to standard primary colors, then it should be possible to capture a nice scan. Using primary-color light sources may not solve metamerism though since there is still the sensor and the properties of what is being scanned to consider. Bob ====================================== Bob Friesenhahn bfr...@si..., http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer, http://www.GraphicsMagick.org/ |
From: Jan-Peter H. <ho...@co...> - 2005-06-20 16:50:28
|
Hello list (cc. Wolf Faust) If some people have succes by profilinig their printer via scanners. We should get even better results, if somebody produce an additional IT-8 chart as inkjet-print, measure this with a spectrophotometer and sells this chart incl. characterization-data like a photographic IT-8 chart incl. the measurement-data. A user who buys this chart inkl. data makes a special scanner-profile only for reprofiling his printer. This would make only sense for pigmented ink, because prints of dye-inks have to heavy colordrifts over the time. Wolf, if some people would be interested on such thinks, would you like to distribute also such charts additional to your photographic produced charts ? :-) Jan-Peter Bob Friesenhahn schrieb: > profiling printer with a profiled scanner > This is indeed an interesting issue. At first glance it seemed like > nonsense to me, but after some more thought it seems that if the color > of the LEDs are indeed narrow-band light sources and their colors > happen to be located close to standard primary colors, then it should be > possible to capture a nice scan. > > Using primary-color light sources may not solve metamerism though since > there is still the sensor and the properties of what is being scanned to > consider. -- -- homann colormanagement ------ fon/fax +49 30 611 075 18 Jan-Peter Homann ------------- mobile +49 171 54 70 358 Kastanienallee 71 ------- http://www.colormanagement.de 10435 Berlin --------- mailto:ho...@co... |
From: Graeme G. <gr...@ar...> - 2005-06-20 23:57:10
|
Jan-Peter Homann wrote: > Hello list (cc. Wolf Faust) > If some people have succes by profilinig their printer via scanners. > We should get even better results, if somebody produce an additional > IT-8 chart as inkjet-print, measure this with a spectrophotometer and > sells this chart incl. characterization-data like a photographic IT-8 > chart incl. the measurement-data. My experience was that it did noticeably improve the quality of the profile. If you're going to do this though, I wouldn't use an IT8.7/2 chart. It very poorly samples the colorspace. Generate something with a lot more squares, more evenly spaced. Graeme Gill. |
From: Hal V E. <hv...@as...> - 2005-06-21 01:03:23
|
On Monday 20 June 2005 04:58 pm, Graeme Gill wrote: > Jan-Peter Homann wrote: > > Hello list (cc. Wolf Faust) > > If some people have succes by profilinig their printer via scanners. > > We should get even better results, if somebody produce an additional > > IT-8 chart as inkjet-print, measure this with a spectrophotometer and > > sells this chart incl. characterization-data like a photographic IT-8 > > chart incl. the measurement-data. > > My experience was that it did noticeably improve the quality of the > profile. If you're going to do this though, I wouldn't use an IT8.7/2 > chart. It very poorly samples the colorspace. Generate something with > a lot more squares, more evenly spaced. > > Graeme Gill. You are probably correct about using something better than an IT8.7 target. But I don't know of any calibrated color chart that is widely available. But then again I have not looked since all of the profiling software that I am aware of uses the IT8.7 for calibration. Are there anyother options? |
From: Graeme G. <gr...@ar...> - 2005-06-21 02:22:22
|
Hal V Engel wrote: > You are probably correct about using something better than an IT8.7 target. > But I don't know of any calibrated color chart that is widely available. But > then again I have not looked since all of the profiling software that I am > aware of uses the IT8.7 for calibration. Are there anyother options? The chart doesn't have to be widely available. Make a custom one. Graeme Gill. |
From: Kristian K. <kri...@ki...> - 2005-06-30 19:51:37
|
Hi, I found the define LONGLONG either u_int64_t or unsigned long long in lcms.h. As I know (on windows) LONGLONG should be a signed value, the ULONGLONG is used for unsigned (defined in winnt.h, oledb.h and wtypes.h as __int64). I compile 64bit adress compatible and using the LONGLONG / ULONGLONG for many sizetypes. The lcms define cause trouble on my signed LONGLONG, that this become unsigned (bad if your asking for negative values). Is there a special reason why you are using unsigned values for this define ? kind regards, Kristian Kratzenstein -------------------------------------------- Kristian Kratzenstein Gettorf Kristian.Kratzenstein@KielNET.net -------------------------------------------- |
From: Marti <ma...@li...> - 2005-06-30 20:41:23
|
Hi, >I found the define LONGLONG either u_int64_t or unsigned long long in > lcms.h. That's solved in the CVS. You can either get a snapshot or wait until release 1.15, which is scheduled for september. 1.15 is still missing some features, and I have to upload latest sources to CVS as well. So, be aware sources are going to change. But this has to wait because I will be attending the libresoftware meeting in Dijon next week. Regards, -- Marti Maria The littlecms project. www.littlecms.com ----- Original Message ----- From: "Kristian Kratzenstein" <kri...@ki...> To: <lcm...@li...> Sent: Thursday, June 30, 2005 9:48 PM Subject: [Lcms-user] Question regarding used LONGLONG define > Hi, > > I found the define LONGLONG either u_int64_t or unsigned long long in > lcms.h. > As I know (on windows) LONGLONG should be a signed value, the ULONGLONG is > used for unsigned (defined in winnt.h, oledb.h and wtypes.h as __int64). > I compile 64bit adress compatible and using the LONGLONG / ULONGLONG for > many sizetypes. The lcms define cause trouble on my signed LONGLONG, that > this become unsigned (bad if your asking for negative values). > Is there a special reason why you are using unsigned values for this > define ? > > kind regards, > > Kristian Kratzenstein > > -------------------------------------------- > Kristian Kratzenstein > Gettorf > > Kristian.Kratzenstein@KielNET.net > -------------------------------------------- > > > > > ------------------------------------------------------- > SF.Net email is sponsored by: Discover Easy Linux Migration Strategies > from IBM. Find simple to follow Roadmaps, straightforward articles, > informative Webcasts and more! Get everything you need to get up to > speed, fast. http://ads.osdn.com/?ad_id=7477&alloc_id=16492&op=click > _______________________________________________ > Lcms-user mailing list > Lcm...@li... > https://lists.sourceforge.net/lists/listinfo/lcms-user > > > > -- > No virus found in this incoming message. > Checked by AVG Anti-Virus. > Version: 7.0.323 / Virus Database: 267.8.8/35 - Release Date: 30/06/2005 > > -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.323 / Virus Database: 267.8.8/35 - Release Date: 30/06/2005 |
From: Hal V E. <hv...@as...> - 2005-06-20 23:21:15
|
On Monday 20 June 2005 09:50 am, Jan-Peter Homann wrote: > Hello list (cc. Wolf Faust) > If some people have succes by profilinig their printer via scanners. > We should get even better results, if somebody produce an additional > IT-8 chart as inkjet-print, measure this with a spectrophotometer and > sells this chart incl. characterization-data like a photographic IT-8 > chart incl. the measurement-data. > > A user who buys this chart inkl. data makes a special scanner-profile > only for reprofiling his printer. > > This would make only sense for pigmented ink, because prints of dye-inks > have to heavy colordrifts over the time. > > Wolf, if some people would be interested on such thinks, would you like > to distribute also such charts additional to your photographic produced > charts ? > > :-) Jan-Peter > The problem is that there would have to be IT8.7 charts make for each specific ink being used. Since there are many pigmented inks being used this would result in a large number of possibilities. I think that the resulting charts would be very specialized and expensive. Currently I can get standard 5x7 Wolf Faust charts for $15 each. These have a useful life of about 2 to 3 years. If charts were offered printed in MediaStreet Enhanced Generations gen 6 inks (the ink I currently use) I would expect these to be significantly more expensive - on the order of $75 per chart perhaps more. On the other hand these might have a longer useful life than the current charts since the ink is rated for almost 200 years on some papers. Having ink specific IT8.7 charts would make the creation of printer profiles using a scanner significantly more accurate then when using generic IT8.7 charts since this would eliminate the metamerism problem. I would also expect it to be easier to get good results. This is because the IT8.7target and the printer target would both shift in nearly the same way in response to a given light source (it might also be influenced by the papers used in both targets) and the resulting profiles could be nearly as good as those created with a spectrophotometer. Would I pay $75 for an IT8.7 chart printed with the inks I use? Perhaps since I understand the implications of this. But many using scanner based profiling software might not understand how this would give them better results and many might be reluctant to spend that much more for something that they do not understand. On the other hand the profiles I am now getting using a standard Wolf Faust IT8.7 chart and an LED based scanner give me better output with my 3rd party inks and papers than the canned profiles Epson ships with the printer using the specified Epson inks and papers. The difference is not huge but it is there none the less. I should add though that it has taken me a lot of effort to get to this point and the learning curve is a steep one. Would a ink specific IT8.7 chart be enough better to justify the extra cost? The only way to know for sure would be to give it a try. Since these charts do not currently exist there is no way to test this. I also agree with Jan-Peter that this only makes sense for pigment inks. First because these inks exhibit very stable colors after the inks are dry (say after 24 to 48 hours) and because pigment inks exhibit much higher levels of metamerism than do dye based inks. It also appears that printer manufacturers are moving away from dye inks to pigment inks. For example all of the newer Epson printers are using pigment type inks where as a few years ago most of the Epson printers were using dye inks. > Bob Friesenhahn schrieb: > > profiling printer with a profiled scanner > > > > This is indeed an interesting issue. At first glance it seemed like > > nonsense to me, but after some more thought it seems that if the color > > of the LEDs are indeed narrow-band light sources and their colors > > happen to be located close to standard primary colors, then it should be > > possible to capture a nice scan. > > > > Using primary-color light sources may not solve metamerism though since > > there is still the sensor and the properties of what is being scanned to > > consider. This of course is true. My point was not that this would totally eliminate metamerism but rather that total metamerism effects would be reduced to low enough levels that it becomes a minor rather than a major problem. I have seen this first hand in my own digital darkroom and many of those that post to the ProfilePrism mailing list report the same thing. The author of ProfilePrism also uses LED based scanners to the exclusion of others for doing his own profiles. The real issue is not that the targets color shift in the scanner but rather that the printer target and the IT8.7 target tend to color shift in different directions and by different amounts. If a different scanner light source like LEDs results in smaller shifts or reduces the difference in the shift direction in both targets you have an improvement in the profiles that are created. Which of these are improved by using an LED based scanner I do not know but my gut feeling is that it is both. The point is that if both targets sifted by the same amount and in the same direction you would get nearly perfect profiles even if the amount of color sift was large. That is why Jan-Peter's idea of having ink specific IT8.7 targets is appealing. This would result in both targets responding to the scanner light source in nearly the exact same way. This would make the scanner light source much less critical and open up the possibility for more users to create high quality printer profiles without the need for very expensive equipment. |
From: Bob F. <bfr...@si...> - 2005-06-20 23:35:06
|
On Mon, 20 Jun 2005, Hal V Engel wrote: > Having ink specific IT8.7 charts would make the creation of printer profiles > using a scanner significantly more accurate then when using generic IT8.7 > charts since this would eliminate the metamerism problem. I would also You are saying that your scanner acts like a spectrophotometer? I find it difficult to believe that using LEDs in the scanner causes it to not suffer from metamerism. Bob ====================================== Bob Friesenhahn bfr...@si..., http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer, http://www.GraphicsMagick.org/ |
From: Hal V E. <hv...@as...> - 2005-06-21 00:59:01
|
On Monday 20 June 2005 04:34 pm, Bob Friesenhahn wrote: > On Mon, 20 Jun 2005, Hal V Engel wrote: > > Having ink specific IT8.7 charts would make the creation of printer > > profiles using a scanner significantly more accurate then when using > > generic IT8.7 charts since this would eliminate the metamerism problem. > > I would also > > You are saying that your scanner acts like a spectrophotometer? I > find it difficult to believe that using LEDs in the scanner causes it > to not suffer from metamerism. > > Bob No I am saying that IF the amount and direction of the color shift caused by metamerism are the same for both the IT8.7 calibration chart and the printer target then metamerism in no longer a significant issue. LED scanners seem to reduce the amount of metamerism significantly and this results in better printer profiles because the magnitude of the difference of the color shift between the calibration target and the printer target is (much) smaller. But if the IT8.7 calibration chart and the printer target used the same inks then both will have almost exactly the same amount and direction of color shift due to metamerism. This would make metamerism (close to) a non-issue even for scanners with light sources that have high levels of metamerism with the inks being used in the printer. Again the real issue is not that there is metamerism in the scanned images but rather that the amount and direction of the color shift from metamerism is different for the calibration chart and the printer target and that the magnitude and direction of the shifts for each are not known. The above assumes that the profiling software is like ProfilePrism. With ProfilePrism you scan both the printer target and the IT8.7 target at the same time using the same settings. That image is pulled into the software and used to calibrate the scanned image as part of the process of creating the profile. When I tried to use ProfilePrism with a non-LED scanner I was never able to get good results when trying to profile with pigment inks. Others using the same software versions reported good results when using dye inks. Dye inks have low levels of metamerism. When I switched to a LED scanner my results improved dramatically with the same version of the software. One user on the ProfilePrism list has profiled the same printer using ProfilePrism with a LED scanner and some other software with a spectrophotometer and reported that the profiles were very close to the same with the spectrophotometer profile being a tiny bit better. So no I am not saying that my LED scanner is the same as a spectrophotometer or even that I can get results that are as good. But rather that if I get everything just right (this requires a significant amount of experience with the software) I can get results that are almost as good but at a lot lower cost. As I pointed out I am currently getting results that are somewhat better than the canned Epson profiles (the ones that need to be hand installed that are for specific printers, papers and resolutions). So I must be getting very close to the best results obtainable from my setup. Could I do better if I had a spectrophotometer? Yes I am sure that I could but I am also fairly sure that the difference would not be a large one. I would really like to have a spectrophotometer but I currently can not afford one so I have to make do with what I have. I suspect that there are many other users that would be more than a little reluctant to spend what it costs to buy a spectrophotometer but who would like to be able to get good results using less expensive equipment such as a scanner. |
From: Bob F. <bfr...@si...> - 2005-06-21 01:06:16
|
On Mon, 20 Jun 2005, Hal V Engel wrote: >> >> You are saying that your scanner acts like a spectrophotometer? I >> find it difficult to believe that using LEDs in the scanner causes it >> to not suffer from metamerism. > > No I am saying that IF the amount and direction of the color shift caused by > metamerism are the same for both the IT8.7 calibration chart and the printer > target then metamerism in no longer a significant issue. LED scanners seem Ahhh, but metamerism still exists! Metamerism due to different ink will be elminated but there may still be metamerism which causes the scanner to think that two colors are very similar when they are actually not. Bob ====================================== Bob Friesenhahn bfr...@si..., http://www.simplesystems.org/users/bfriesen/ GraphicsMagick Maintainer, http://www.GraphicsMagick.org/ |
From: Graeme G. <gr...@ar...> - 2005-06-21 02:20:18
|
Bob Friesenhahn wrote: > On Mon, 20 Jun 2005, Hal V Engel wrote: >> No I am saying that IF the amount and direction of the color shift >> caused by metamerism are the same for both the IT8.7 calibration >> chart and the printer target then metamerism in no longer a >> significant issue. LED scanners seem > > Ahhh, but metamerism still exists! Metamerism due to different ink will > be elminated but there may still be metamerism which causes the scanner > to think that two colors are very similar when they are actually not. This can be one issue. Another more certain one when profiling CMYK devices is that there are spectrally different combinations that have the same same appearance due to the amount of K used. Using a printed scanner calibration chart can't compensate for this (in fact you have to choose the level of GCR to print the scanner test chart). You would have to print several scanner test charts with a range of GCRs, and then use the appropriate scanner profile for measuring printer test patches with the matching GCR to hope to compensate for this effect. Graeme Gill. |
From: Gerhard F. <nos...@gm...> - 2005-06-21 19:52:43
|
Graeme Gill schrieb: >> Ahhh, but metamerism still exists! Metamerism due to different ink >> will be elminated but there may still be metamerism which causes the >> scanner to think that two colors are very similar when they are >> actually not. > > This can be one issue. Another more certain one when profiling CMYK > devices is that there are spectrally different combinations that have > the same same appearance due to the amount of K used. Recently, I've actually evaluated an example for metamerism regarding black generation. For the worst case I encountered, I computed a color difference of 7-8 dE94 units under F10 illuminant, for two patches, printed with different black generation, which have approximately the same color under D50. That's pretty much and by far not neglectible! The evaluation was based on measurement from a professional inkjet LFP. > Using a printed scanner calibration chart can't compensate for this > (in fact you have to choose the level of GCR to print the scanner test > chart). You would have to print several scanner test charts with a > range of GCRs, and then use the appropriate scanner profile for > measuring printer test patches with the matching GCR to hope to > compensate for this effect. ... or probably restrict to RGB printer profiling (with all its disadvatages, but under Windows it's in many cases the only option anyway). Btw, I'm actually wondering whether well-known spectral imaging approaches might also work for turning a scanner into a more colorimetric device: The idea is to scan the same image a) without, and b) with M different color filter foils between the glass and the paper. So we finally end up not only with three captured RGB channels, but with M+1 scans, yielding a total of N=(M+1)*3 (for instance 12 or 15) captured color channels. Now it should be possible to characterize either the N-channel to spectral mappping (e.g. based on the first few (e.g. 10-15) principal components of typical printed reflectance spectra), or a N-channel to XYZ mapping, which should suffer less from metamerism, than the RGB to XYZ mapping does. Of course there are several issues to consider: - the optimal choice of the filters - spatial uniformity of the filters - general and cheap availability of suitable filter foils - spatial image registration for the M+1 images is necessary - I'm not sure, if it works as well with narrow-band scanners - What kind of target to use for characterization - is an IT8.7 target with spectral reference data sufficient? - other issues I missed so far ??? Any comments or objections? Regards, Gerhard |
From: Graeme G. <gr...@ar...> - 2005-06-22 03:13:47
|
Gerhard Fuernkranz wrote: > Btw, I'm actually wondering whether well-known spectral imaging > approaches might also work for turning a scanner into a more > colorimetric device: > Any comments or objections? Some comments spring to mind: This probably won't work for narrow band scanner sensitivities. How do you calibrate your spectral scanner ? What maths is needed to extract spectral info from RGB+filter output ? Is it all worth it ? For that much effort, you may be better of buying a spectrometer, if your aim is to read test patches, rather than do a spectral scan of artwork etc. There is a bunch of literature about this sort of thing, mostly from the art preservation/reproduction folks, who are most active in pursuing spectral reproduction. Graeme Gill. |
From: Gerhard F. <nos...@gm...> - 2005-06-22 23:43:21
|
Graeme Gill schrieb: > Some comments spring to mind: > > This probably won't work for narrow band scanner sensitivities. Thats' what I suspect as well. > How do you calibrate your spectral scanner ? > What maths is needed to extract spectral info from RGB+filter output ? I was basically thinking of methods like here http://www.cis.rit.edu/mcsl/research/PDFs/Camera02.pdf But as you say, there's a bunch of literature, describing several possible methods. Likely we do not even need to reconstruct the reflectance spectra, but a mapping from the N device channels of the "modified scanner" to XYZ (under D50) is sufficient. Basically this could be established like a RGB to XYZ mapping. However, a CLUT with 9, 12, or even 15 input dimensions is no longer practicable. With a large number of device channels, maybe even a Nx3 matrix matrix is sufficient? The more device channels (with linearly independent CMFs) are captured, the closer the device to XYZ mapping can be described by a matrix (assuming linear raw digital counts). http://www.cis.rit.edu/mcsl/research/PDFs/Camera05a.pdf reports pretty good numbers with only 6 or 9 device channels and a linear mapping. But they don't use a scanner, but a camera which likely behaves a priori more colorimetric. So if a Nx3 matrix does'n suffice, then maybe a matrix using 2nd order mutivariate polynomial terms of the N device values? Regarding calibration, I've no evidence whether the spectra found on an IT8 target are sufficient to establish the mapping. If not, then maybe additional targets with inkjet spectra might be required. > Is it all worth it? Right, that's probably the main question. > For that much effort, you may be better of buying > a spectrometer, if your aim is to read test patches, rather than > do a spectral scan of artwork etc. I do have one anyway, so that't not the issue. But obviously many people are seeking rather for a $100 and not a $1000 solution. Thogh the art reproduction folks have proven that these methods basically do work, I have no evidence, whehter it may or may not work eventually in practice with a consumer scanner, cheap filter foils, etc. (due to various possible limitations I can imagine). But granted that it would work reasonably, wouldn't it be worth a few dollars for the end user to buy some filter foils, and the effort to make e.g. four or five scans instead of one? Regards, Gerhard > There is a bunch of literature about this sort of thing, mostly > from the art preservation/reproduction folks, who are most active > in pursuing spectral reproduction. > > Graeme Gill. |
From: Graeme G. <gr...@ar...> - 2005-06-23 02:29:12
|
Gerhard Fuernkranz wrote: > Regarding calibration, I've no evidence whether the spectra found on an > IT8 target are sufficient to establish the mapping. If not, then maybe > additional targets with inkjet spectra might be required. A reference that only has three independent channels (ie. Photographic film) would likely not have enough independent channels (ie. metamers) to give you enough information to establish your matrix values. You'd probably need one or several charts with known spectra (or XYZ) values, composed of individual colorants (such as paints), or charts with several different base colorants. Graeme Gill. |
From: Hal V E. <hv...@as...> - 2005-06-21 02:21:19
|
On Monday 20 June 2005 06:05 pm, Bob Friesenhahn wrote: > On Mon, 20 Jun 2005, Hal V Engel wrote: > >> You are saying that your scanner acts like a spectrophotometer? =A0I > >> find it difficult to believe that using LEDs in the scanner causes it > >> to not suffer from metamerism. > > > > No I am saying that IF the amount and direction of the color shift caus= ed > > by metamerism are the same for both the IT8.7 calibration chart and the > > printer target then metamerism in no longer a significant issue. =A0LED > > scanners seem > > Ahhh, but metamerism still exists! =A0Metamerism due to different ink > will be elminated but there may still be metamerism which causes the > scanner to think that two colors are very similar when they are > actually not. > > Bob Yes when the calibration chart and the printer target have different inks o= r=20 colorants then the only question is the magnitude of the difference. If it= =20 is very small (like with an LED scanner) then the profile can be very good.= =20 But if it is large than the profile is very bad. =20 If both the calibration chart and the printer target use the same inks then= =20 both will have almost exactly the same metamerism characteristics since in= =20 the scanner both will be lighted by the same light source and imaged using= =20 the same sensor and if scanned in the same pass, like they are in=20 ProfilePrism, scanned with the same scanning parameters. Under those=20 conditions if two colors on the two charts appear to the scanner to be the= =20 same color then they are extremely close to the same color. That was the=20 whole point behind having both use the same inks. It almost totally=20 eliminates one vary large variable. =20 Of course it can not totally eliminate it since the calibration chart and t= he=20 printer target will very likely be on different papers. But the light=20 source, sensor and ink are no longer variables and have instead been=20 transformed into constants. The only remaining variable that can influence= =20 metamerism is, in your words, "the properties of what is being scanned" whi= ch=20 in this case is the media that the calibration chart and printer target are= =20 printed on. If the media are different (highly likely) then there will sti= ll=20 be a very slight difference in the metamerism characteristics of the=20 calibration chart and the printer target. But my experience is that the sa= me=20 ink printed on different media will have almost exactly the same metamerism= =20 characteristics - the ink formulation accounts for at least 90% of the=20 metamerism characteristic. In addition my experience is that the differenc= e=20 between the same inks on different media is less if the media are similar. = =20 That is the metamerism characteristics of a given ink will be almost=20 identical if both media are glossy or both are mat. The ideal setup would = be=20 to have the calibration chart printed on the exact same media as will be us= ed=20 in the printer and if possible on the same type of printer. This would=20 result in the calibration chart having exactly the same metamerism=20 characteristics as the printer target.=20 Again the real issue is how to minimize the affects of metamerism on a scan= ner=20 based printer profiling process. If the calbration chart and the printer=20 target have nearly the same metamerism characteristics, no matter how much= =20 metamerism actually takes place, metamerism's impact on the profiling proce= ss=20 has been minimized.=20 I should add that although I think that ink specific calibration charts wou= ld=20 be helpful to those using scanners to profile printers I do not think that= =20 there is anything other than an extremely low probability that these will=20 ever become available. So how useful these may be is likely a mote point.= =20 Hal |
From: Jan-Peter H. <ho...@co...> - 2005-06-16 10:57:15
|
Hello Ignacio, hello list I think, it would be a big help for developers, if your tools (or parts) would be part of the lcms contribution. This would make it easier for developers to realize things like I describes in mails before: saving and transporting coloradjustments as devicelink or abstract profile ------ (If a printer-driver supports devicelink-profiles, user could e.g.print a testimage, apply the coloradjustments, they want, and print later every job with this coloradjustment) Use your imaging application as devicelink-profile editor ------------ Often an ICC-conversion delivers not exactly the result, which users expect. But with some coloradjustment, it would be nearly perfect. A way to solve this: 1. Link the ICC-profiles together to an devicelink-profile 2. Transform an IT-8 like testimage with the devicelink-profile 3. Edit the transformed testimage in your imaging application 4. Link the edited values with the original IT-8 values to a new devicelink-profile. 5. Apply this devicelink-profile instead your normal ICC-transformation. Imagine e.g. the next version of gutenprint comes with lcms-support and some standard-icc profiles. The results for your printer are good, but not optimal. The described way is easy to go, without the need to buy a spectrophotmeter and a profiling software. So this technology makes sense for all color-aware users of opensource applications and not only for graphic arts color geeks. -- An alternative to the devicelink-way would the use of abstract profiles for storing color-adjustments and apply them additionally in the ICC-chain. So please make it easy to the developers to integrate such functionality in their imaging applications, by deliviring tools with lcms. :-) Jan-Peter ignacio ruiz de conejo schrieb: > Jan-Peter: > > Others in the list have addressed your first question below. > > Regarding the second one (how to convert imagepixel to ASCII) > I have two different answers: > > 1) You can always use Gretag's freeware tool ColorLab to open > an image and save it as text. > > 2) I wrote two tools, TIFFtoIT8 and IT8toTIFF, that allow > the conversions between pixels and ASCII and viceversa. > They make use of Marti's CGATS parser and TIFFLIB. > Although they work for me, they are limited in scope > (address only my needs: RGB, LAB, CMYK, Separated > TIFFs, image offsets, pixel replication) > > Should Marti consider a Contrib folder that would gather tools > and (acceptable) pieces of code writen with lcms? > I wouldn't mind sharing mine... > > Regards, > Ignacio > >> My question: >> - Is there code for converting ASCII to a devicelink-profile ? >> - Is there code for converting imagepixel to ASCII ? >> - can we integrate this in lcms ? >> >> greetings from the colorspace >> :-) Jan-Peter >> >> -- >> > > -- -- homann colormanagement ------ fon/fax +49 30 611 075 18 Jan-Peter Homann ------------- mobile +49 171 54 70 358 Kastanienallee 71 ------- http://www.colormanagement.de 10435 Berlin --------- mailto:ho...@co... |
From: Gerhard F. <nos...@gm...> - 2005-06-16 12:37:11
|
> --- Ursprüngliche Nachricht --- > Von: ignacio ruiz de conejo <ign...@bp...> > 2) I wrote two tools, TIFFtoIT8 and IT8toTIFF, that allow > the conversions between pixels and ASCII and viceversa. > They make use of Marti's CGATS parser and TIFFLIB. > Although they work for me, they are limited in scope > (address only my needs: RGB, LAB, CMYK, Separated > TIFFs, image offsets, pixel replication) Ignacio, sounds interesting. - does it support both, 8-bit and 16-bit images? - does it support the conversion of two images with the same dimensions (original image, and color-transformed version of the image) into a single CGATS file? Thanks, Gerhard -- Geschenkt: 3 Monate GMX ProMail gratis + 3 Ausgaben stern gratis ++ Jetzt anmelden & testen ++ http://www.gmx.net/de/go/promail ++ |