Re: filmscanners: VueScan 7.3.5 color space bug?
I disagree ... what we end up with in Photoshop should never go thru the gamut funnel you describe above. What we end up in Photoshop should be the result of as few profile conversions as possible, and only one should be needed ... i.e.: sorry, didnt i say that? device space = scanner space . scanner data is corrected in scan program (which is done in some device independent internal space) image is represented in scan program through monitor icm (but which doesnt do anything with the image data) image is saved through whatever space you choose to. conversion (corrected output) from scanner space to working space: scanner space scan program correction saved to working space X or no conversion: scanner space vuescan scans only uncorrected output in scanner space to pshop jan
RE: filmscanners: VueScan 7.3.5 color space bug?
Julian writes ... Thanks, Ed. This seems like a good interim solution, until VueScan gets real color management. But I think there may be a bug in there. I set up 7.3.5 to display images in Apple RGB, and to save files in Adobe RGB. The file images, when opened up in Photoshop, are significantly darker than VueScan's preview and scan displays. Just so we're all on the same page ... what is your display calibrated to? Are we talking about the Mac version of Vuescan? I have to ask because what you describe is exactly what I'd expect to see. That is, my display is D65 gamma=2.2 ... since Vuescan doesn't compensate for my display, if I put a preview up in D50 1.8, adjusted the color gamma just right, and than asked to deliver D65 2.2, I'd certainly expect the difference. I believe the key to the problem is Vuescan doesn't compensate for your display, and you should not preview in D50 1.8 unless your display is configured for it (granted ... it may be and I'm all wet ...g...) cheerios ... shAf :o) Avalon Peninsula, Newfoundland
RE: filmscanners: VueScan 7.3.5 color space bug?
Julian writes ... Thanks, Ed. This seems like a good interim solution, until VueScan gets real color management. ... My idea of real color management would be Vuescan's ability to accommodate a scanner's device profile ... but I understand the complexities of Vuescan accommodating LUT profiles. On the other hand ... Ed's implimentation of monitor color space is a mystery. The option implies a device space but instead makes only working spaces available, and what does device RGB imply in this context? sRGB should be my preference ... it is most like my D65 2.2 monitor ... but is not my monitor's color space! While monitor spaces are generally 2D matrix profiles, Vuescan should be able to work with them ... and Vuescan, in this context only, should allow us to choose our specific display profile. my $0.02 cheerios ... shAf :o) Avalon Peninsula, Newfoundland
Re: filmscanners: VueScan 7.3.5 color space bug?
In a message dated 12/17/2001 11:46:30 PM EST, [EMAIL PROTECTED] writes: I have some sample images which demonstrate that 7.3.5 is actually worse than 7.3.4, at least for the above combination of color spaces. http://julianv.home.mindspring.com/color_test/color_test_1.html This is just too complicated for me - there are lots of different color space conversions, screen captures, Photoshop conversions, etc. The simplest thing to do is to preview an image with Color|Monitor color space set to Adobe RGB (gamma 1.8) then preview it immediately again with Apple RGB (gamma 2.2). The gamma 1.8 preview is darker than the gamma 2.2 preview. That's how it's supposed to work. A separate issue is to determine why the Photoshop window looks different from the VueScan Scan window. This is controlled by how you have your monitor profiled, whether you've given the right monitor profile to Mac OS / Photoshop, and whether Photoshop is set up properly. It would be useful if someone could do this same test using Photoshop on Windows. The Scan window in VueScan should look very similar to the image in Photoshop. Regards, Ed Hamrick
RE: filmscanners: VueScan 7.3.5 color space bug?
Ed writes ... ... This is just too complicated for me - there are lots of different color space conversions, screen captures, Photoshop conversions, etc. It really isn't difficult. With regard to the color space you embed to the 'cropped' image, keep doing it like before. With regard to Vuescan's preview and scan presentations, you should associate Vuescan color with whatever we select as monitor space ... and leave it to us to select the color space which best matches our monitor (... but I do wish you'd allow us to select something other than sRGB ... i.e., our actual display profile). All else you'd need consider is no double conversions ... that is, what is in Vuescan's color space (monitor space) is independent of the file color space we choose to embed. I.E., [Vuescan space] == 'device RGB' = 'monitor space' [embedded space] == 'device RGB' = 'file color space' NOT [embedded space] == 'device RGB' = 'monitor space' = 'file color space' ... else, we shall all take up a contribution and buy you an installation of Photoshop ...*smile*... cheerios ... shAf :o) Avalon Peninsula, Newfoundland
Re: filmscanners: VueScan 7.3.5 color space bug?
It would be useful if someone could do this same test using Photoshop on Windows. The Scan window in VueScan should look very similar to the image in Photoshop. Regards, Ed Hamrick I did more or less, as told in a previous post.: i replace here part of my reply in your 7.3.5. anouncement thread -However i do get correct matching between pshop and vuescan when i set both monitor and file space to adobe RGB. That is nice but i dont think it is correct. The custom monitor profile should be selected in the display list which, when the same allover the system, will display all equal. -Or is your Monitor color space in reality a monitor translation of working spaces= working space preview? In order to preview one space and scan to a different space? The scheme should be scannerscanner icmvuescan interface space monitor icm monitor and vuescan interface space file icm file and then file file icm pshop interface space (lab) monitor icm monitor These icm things keeps us confused
Re: filmscanners: VueScan 7.3.5 color space bug?
In a message dated 12/18/2001 11:51:45 AM EST, [EMAIL PROTECTED] writes: However i do get correct matching between pshop and vuescan when i set both monitor and file space to adobe RGB. A more interesting question is whether you get correcting matching between VueScan and Photoshop when you use each of the different Color|File color space settings (except for Device RGB) while leaving Color|Monitor color space set to Adobe RGB. Is your monitor set for gamma 1.8? The scheme should be ... That very well may be. However, it's probably more useful to diagnose how it currently works before discussing how to change it. Regards, Ed Hamrick
Re: filmscanners: VueScan 7.3.5 color space bug?
There are no difference in my MAC between the scan window in Vuescan and later on in photoshop. Mikael Risedal -- From: [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: filmscanners: VueScan 7.3.5 color space bug? Date: Tue, 18 Dec 2001 09:47:32 EST In a message dated 12/17/2001 11:46:30 PM EST, [EMAIL PROTECTED] writes: I have some sample images which demonstrate that 7.3.5 is actually worse than 7.3.4, at least for the above combination of color spaces. http://julianv.home.mindspring.com/color_test/color_test_1.html This is just too complicated for me - there are lots of different color space conversions, screen captures, Photoshop conversions, etc. The simplest thing to do is to preview an image with Color|Monitor color space set to Adobe RGB (gamma 1.8) then preview it immediately again with Apple RGB (gamma 2.2). The gamma 1.8 preview is darker than the gamma 2.2 preview. That's how it's supposed to work. A separate issue is to determine why the Photoshop window looks different from the VueScan Scan window. This is controlled by how you have your monitor profiled, whether you've given the right monitor profile to Mac OS / Photoshop, and whether Photoshop is set up properly. It would be useful if someone could do this same test using Photoshop on Windows. The Scan window in VueScan should look very similar to the image in Photoshop. Regards, Ed Hamrick _ MSN Photos är det enklaste sättet att dela ut och skriva ut foton: http://photos.msn.se/Support/WorldWide.aspx
RE: filmscanners: VueScan 7.3.5 color space bug?
Ed writes ... In a message dated 12/18/2001 11:51:45 AM EST, [EMAIL PROTECTED] writes: ... The scheme should be ... That very well may be. However, it's probably more useful to diagnose how it currently works before discussing how to change it. ... On my Win2k computer, display configured for D65, 2.2, I see and get what I expect if I configure Vuescan monitor space for 'sRGB' and ask file space to be 'EktaspaceRGB' ... HOWEVER, I say that with a caveat: I cannot inspect the resulting RGB in Phoptoshop (with monitor color) and discern that my EktaspaceRGB did NOT go thru the gamut funnel I mentioned in the previous post. That being said, it looks ok on my computer ... certainly not what sirius describes. Just please tell me that EktaspaceRGB came direct from device RGB in a single conversion(?) shAf :o)
RE: filmscanners: VueScan 7.3.5 color space bug?
sirius writes ... ... The scheme should be scannerscanner icmvuescan interface space monitor icm monitor and vuescan interface space file icm file and then file file icm pshop interface space (lab) monitor icm monitor ... I disagree ... what we end up with in Photoshop should never go thru the gamut funnel you describe above. What we end up in Photoshop should be the result of as few profile conversions as possible, and only one should be needed ... i.e.: device RGB = file space RGB that is: large gamut=small gamut=large gamut is the same as small gamut!!! shAf :o)
Re: filmscanners: VueScan 7.3.5 color space bug?
In a message dated 12/18/2001 1:53:44 PM EST, [EMAIL PROTECTED] writes: Just please tell me that EktaspaceRGB came direct from device RGB in a single conversion(?) That's correct. Regards, Ed Hamrick
filmscanners: VueScan 7.3.5 color space bug?
On 12/17/01 3:47 AM, [EMAIL PROTECTED] [EMAIL PROTECTED], wrote: I just released VueScan 7.3.5 for Windows, Mac OS 8/9/X and Linux. ... * Added separate color spaces for Monitor and Files Thanks, Ed. This seems like a good interim solution, until VueScan gets real color management. But I think there may be a bug in there. I set up 7.3.5 to display images in Apple RGB, and to save files in Adobe RGB. The file images, when opened up in Photoshop, are significantly darker than VueScan's preview and scan displays. The difference was much less when I was using version 7.3.4. Recall that with 7.3.4 (and its single Color space setting) I was using Apple RGB and gamma 1.8 while adjusting color tab settings, and switching to Adobe RGB and gamma 2.2 just before doing the final scan to file. I have some sample images which demonstrate that 7.3.5 is actually worse than 7.3.4, at least for the above combination of color spaces. http://julianv.home.mindspring.com/color_test/color_test_1.html and http://julianv.home.mindspring.com/color_test/color_test_2.html The behavior in 7.3.5 might be explained if you are (incorrectly) setting gamma 2.2 when Apple RGB is selected for Monitor color space. -- Julian Vrieslander mailto:[EMAIL PROTECTED]