Luminous Landscape Forum
Raw & Post Processing, Printing => Colour Management => Topic started by: drcahmss on March 22, 2011, 09:33:05 am
-
Anyone out there help? We are scanning negatives using Flextight scanners (X5 and 949) and in the past few months have been experiencing colour mismatch when bringing the scans through into Photoshop CS5 (same thing happened in CS4). We are aware that the colours won't match 100% but for some reason they have become very 'warm'. We have upgraded the software to Flexcolor 4.8.8 but to no avail. We have made sure the colour settings are as they should be i.e. Adobe1998 and all other settings are ok. We have been in discussion with Hasselblad but I think they are running out of ideas too. For info we are pc based (XP Pro) and scan as tiff and not 3f. We have also tried scanning on a Mac and get the same mismatch. If anyone has any ideas we'd love to hear them......
-
I don't know the Flextight software so I can't offer any help on that aspect of it, but I do a lot of colour negative scanning, mainly using a Nikon Super Coolscan 5000ED and sometimes (mainly for testing purposes) a Plustek 7600iAi or an Epson V750 Pro, all with SilverFast Ai Studio and I don't have this problem - the images open in Photoshop looking very close to how they appear in the SilverFast scan preview. I mention this just to make the point that regardless of the fact one is scanning negatives, this does not necessarily have a bearing on the problem you face. I assume you've checked with Hasselblad whether they've changed the negative profiles (which you may have "upgraded" to) some time before you starting seeing this problem.
The most I can do is recommend on how you go about the diagnostics. The place to start would be with an assessment of what changed in respect of colour management in the scanning software, Photoshop software and your computer operating system's colour management modules between the time you last had good results and the time you started to get unsatisfactory results with the same process .
-
So you are saying previews in FlexColor and the resulting scan in Photoshop don’t match right? Off hand, how are you calibrating the display? Could you be using version 4 ICC profiles which might be affecting the preview in FlexColor?
I suspect its FlexColor because Photoshop should be doing the previews correctly. If you open the tagged scan in other ICC aware app’s do they match Photoshop (and mismatch FC?). That would lead me to look at the scanning software. Note that you really need to view both previews at 100% if possible and at the same scale (zoomed to match size).
-
Thanks for your reply digitaldog.
I have never heard of version 4 ICC profiles? I presume CS5 is using version4 profiles, but I dont know? But surely Adobe RGB 1998 will be the same in any version of ICC profiles?
What is the difference between older ICC profiles and the newer version 4?
-
The version 4 profiles I speak of are those for your display. What software are you using to profile it? Check the preferences to make sure its not building V4 profiles. Maybe even try building a Matrix profile if set to LUT (or vise versa), assuming the software gives these options.
-
I am using a EIZO Coloredge CG211, and an eye1 with the Eizo ColourNavigator software for calibration of the monitor.
-
I am using a EIZO Coloredge CG211, and an eye1 with the Eizo ColourNavigator software for calibration of the monitor.
Check the preferences.
-
Andrew, why would (or perhaps better - how could) the display profile affect differences in image appearance, seen on the same display, between a Flextight image and a Photoshop image?
-
Some ICC aware applications barf (that’s the technical term <g>) on V4 profiles. Its only a guess but worth a try. Otherwise, its some preview bug in the Imacon software (assuming its vastly different from Photoshop).
-
Got it, thanks. Would you say these days it's generally safer, and one isn't losing much, sticking with v2 profiles for both display and printer?
-
Got it, thanks. Would you say these days it's generally safer, and one isn't losing much, sticking with v2 profiles for both display and printer?
Without question. V4 profiles are not ready for prime time just yet.
-
Thanks Andrew.
-
Mark may have got it but I don't think I do!
As far as I am aware I am only using v2 profiles. But how can I tell?
-
Please bear in mind that this mismatch is happening on Mac as well as PC, and with the latest version of the Flexcolor software.
I would say that the only common thing between the PC and Mac is that we are viewing the images and scan previews on the same make of monitor.
-
Please bear in mind that this mismatch is happening on Mac as well as PC, and with the latest version of the Flexcolor software.
I would say that the only common thing between the PC and Mac is that we are viewing the images and scan previews on the same make of monitor.
Could you please go back to the second paragraph of reply #1 (mine) and tell us what changed between the time these previews were coherent and the time they stopped being coherent. If anything did change, that may be the vital clue into the cause of the problem. For example, did you up-grade the version of Flextight and nothing else changed? If so, that may be the root of the problem.
-
Hindsight is a wonderful thing.
Unfortunately this change happened about 10 months ago and we cannot remember what changes took place to cause this. We have been working round the problem by making corrections in Photoshop since then. I remember spending the best part of two days trying to sort this out at the time, but with no success. Over the past couple of weeks myself and others have revisited the problem and again spent lots of time trying to sort this out.
As I said we have been in touch with Hasselblad over many months about the problem. They have recently serviced our scanners and cannot replicate the problem or their systems. They have however been very helpful in suggesting various things to try but none have been successful. Hence my post here.
-
If your conversion engine is set to Microsoft (ICM), try setting it to Adobe (ACE).
I observed that the ICM engine may not handle the conversion correctly when Absolute colorimetric was selected under ICM, resulting in very "warmed" image after conversion.
When conversion engine is set to Microsoft (ICM) and with Absolute colorimetric intent
- Assigned profile generated with PM5 or MP, converted to ProPhoto: normal
- Assigned profile generated with PM5 or MP, converted to aRGB or sRGB: warmed & high delta E.
My experience is limited to reflective scan only, with CS4/xp3 and with Kodak Q-60 IT8 reflective target.
-
Thanks jc1
Can you tell me exactly how I would go about changing from Microsoft (ICM) to Adobe (ACE).
I am using XP Pro on a PC.
-
Thanks jc1
Can you tell me exactly how I would go about changing from Microsoft (ICM) to Adobe (ACE).
I am using XP Pro on a PC.
I'm not jc1, but anyhow:
Photoshop>Edit>Color Settings>Conversion Options>Engine. At this point use the drop-down menu to select Adobe (ACE).
-
I have just checked my settings and I am using Adobe (ACE).
-
Without question. V4 profiles are not ready for prime time just yet.
Andrew, further on this, and perhaps a bit OT - but anyhow - could you please explain in what ways V4 profiles are not ready for prime time? This is a matter of obvious concern because we are always given a choice in creating profiles (display or printer) whether we want them as V2 or V4. I believe all of my custom profiles are V4 and I don't seem to have a problem, so either I'm not seeing something, or possibly it affects people with set-ups or requirements different from mine. A bit of elaboration of the currently outstanding issues with V4 would be helpful.
-
Andrew, further on this, and perhaps a bit OT - but anyhow - could you please explain in what ways V4 profiles are not ready for prime time?
Most built today are V2 in structure, there isn’t a defined PRMG for their use in both sets of profiles, many software products barf on them (the old scum dot issue for output profiles, improper preview for display profiles). IOW, few if any app’s take advantage of the V4 structure when the profile is built fully for V4 and worse, many applications produce unacceptable results using them.
-
When you say there isn't a defined PRMG, not clear what you mean, as the ICC appears to me to say there is:http://www.color.org/v4_prmg.xalter (http://www.color.org/v4_prmg.xalter). Or do you mean that certain applications aren't necessarily using it properly? If the latter, would you include Photoshop, Lightroom, BasicColor 4 and our various profiling kits such as MP, PMP, Pulse Elite, ColorMunki as "barfing" at V4?
-
Both profiles have to be written to use the PRMG for one. Even if you found say a V4 ProPhoto RGB profile with the PRMG tag, how about all your output profiles? Neither PMP nor PROFILER have the PRMG tags, but I can’t speak for all app’s that build output profiles.
-
OK, so I guess this means if I hit on a combo of profiles that works, just keep using it. All one can do is play back and forth between V2 and V4 and see which set provides the most reliable screen to print matching in my working environment. Raises an interesting issue with respect to scanner profiles! Given that scanning film has become kind of a niche, one wonders to what extent scanner profiling is V4 aware, or if they are all on V2, and if one is using v2 for the scanner but V4 for the display and the printer, could one be getting disconnects because of this inconsistency?
-
I seem to have started a real discussion here about different profile versions.
Which is very interesting, and I hope to read more and learn more.
But back to my annoying problem of preview and resultant scan not matching.
I have now managed to get the preview and the scan to match by using a Mac laptop.
I have also managed to get the preview and scan to match by using a different PC (but same make and model) with the identical software versions loaded, and using identical icc profiles for colour conversion. i.e. as far as I am aware there is no difference in the software between the two PC's!
Just one will scan correctly the other will not!
-
I have now managed to get the preview and the scan to match by using a Mac laptop.
I have also managed to get the preview and scan to match by using a different PC (but same make and model) with the identical software versions loaded, and using identical icc profiles for colour conversion. i.e. as far as I am aware there is no difference in the software between the two PC's!
Intent (Absolute or Relative Colorimetric) may be set differently.
-
Both are on the same rendering intent. Relative Colorimetric!
-
A gentle reminder, 2 places to be checked. :)
(a) Edit > color setting > fewer options > conversion option
Here is for user to set the system (preferred) default.
(b) Edit > convert to profile > conversion option
Be aware that the conversion option over here may be different from (a), as it defaults to the previous or last selection.
-
To add to above.
Depending on the options (ACE, RelCol) selected , the setting (b) is also affecting the "preview" when assigning profile (Edit > assign profile --> preview). Confusion can be avoided if both (a) and (b) are set correctly.
-
Checked both places and settings and everything seems fine.
-
(a) Edit > color setting > fewer options > conversion option
Here is for user to set the system (preferred) default.
(b) Edit > convert to profile > conversion option
Be aware that the conversion option over here may be different from (a), as it defaults to the previous or last selection.
A is only applied when one uses Image>Mode Change. Its totally independent from Convert to Profile/Assign Profile. If you never use Mode Change, doesn’t matter what you set there (other than Dither, that controls dither in many areas in the app).
-
I can only repeat what I said earlier.
I have checked both places and settings, and everything seems fine.
But still a warm result in photoshop, relative to the preview in Flexcolor!
-
Does anyone have any other suggestions as to the cause of colour change?
-
I am still trying to get to grips with this problem. I have been through all the suggestions offered but I still have the problem! If anyone as any other thoughts then they will be gratefully received.
-
I've been using a Flextight X5 for some while, and find it much easier to forget about the colour balance in the Flextight software and just adjust itin Photoshop.
-
It is possible to correct the colour in photoshop but it is more work!
When you have measured the neutral grey in flexcolor and measured your highlights and shadows so that they are neutral it is very anoying to have to do this again in photoshop. I feel that the hasselblad input profile for the scanner is slightly off. Does anyone else have this problem?
-
I agree with John. I feel safer editing the fff file in CS5 and not in Flexcolor.
It is a bit more work to do but the result,I think, is better, mainly when the scan comes from negative film.
George
-
I'm having the same problem with a 949. Looks beautiful in Flexcolor. Warm Colors in Photoshop. Did you find any solution
-
We also have a 949 which is exhibiting the same problem. I am sory to say that I have not found a solution to the problem as yet! I am of the opinion that I will have, in some way, to edit the input profile for the scanner to correct the problem.
Does anyone know of a way, of taking a colour profile and manually editing it to remove some red and yellow?
Are there any software programes out there that can do this?
Any replies gratefully appreciated.
-
I have duplicated the problem on 4 different computers mac and pc.
Creating a whole new profile is an interesting idea...
Hasselblad support won't return my calls or emails as well which is kind of annoying. Its true their customer support is terrible.
In FLEXCOLOR:
I output in sRGB.
My Flexcolor ICM:
Input: Flextight Input
RGB: sRGB
I choose both the boxes imbed profile and convert.
In PHOTOSHOP:
My working space is sRGB and I'm not proofing any colors.
My conversion engine is Adobe ACE.
I've tried 3 different computers 2 with gretag macbeth eye monitor calibration installed and two without to make sure that wasn't confusing things. All three show slightly warmer images in once opened in photoshop.
I'm viewing on an EIZO monitor.
-
I have to give Hasselblad the credit, as they have worked closely with me on this one for over a year now, but have sadly not been able to offer a solution.
In Flexcolor
I output to Adobe RGB (1998)
My flexcolor ICM
Input Flextight X5,949
RGB Adobe RGB(1998)
CMYK Hasselblad 330SKEL30K75
Grey Def Hasselblad Grey
In photoshop
My working space is Adobe RGB(1998) and I'm not proofing any colours.
My conversion engine is Adobe ACE
I'm viewing on an EIZO monitor.
-
I should say that I also choose both the boxes imbed profile and convert.
I have tried 3 different PC's and 2 Apple Mac's and the problem is exhibited on all of them.
Strangely if I scan as 3F and then use Flexcolor to process the file to a tiff, using the same ICM criteria, then the processed tiff matches the preview in Flexcolor. But it is a very roundabout way to achieve a correct tiff file. This very high end scanner should be capable of producing correct results by just scanning as tiff!
Hasselblad do need to get to grips with this problem!
-
Wow that is good to know about 3f. At least there is a semi solution. I was just beginning to experiment with that last week but ran out of time. Did hasselblad ever admit there is a problem. How did you get ahold of them support@hasselbladusa.com?
It looks like I use almost identical settings to you. Its interesting sometimes If I scan with my gretag macbeth monitor profile as either my input or output it seems to be even slightly closer to matching. Say a 97% match whereas its usually its a 90-95% match. I don't like making the changes in photoshop either because I really feel like the flextight has a way of interpreting the film that is unique compared to other scanners i've used and I want it to look just like I see it. Adjusting in photoshop takes me to 99% but like you said it shouldn't be hard to just have it work right...
-
Best way to solve this problem is to deselect both embed profile and convert. When you open the image in Photoshop you should use the correct scanner profile (Flexcolor input for scans of negative film, Flexcolor X5 & 949 if you scan slides) and then convert to working space. The results should be the same if you choose to do the conversion in Flexcolor, but I get a better result this way and the colours on my monitor is the same in both applications.
-
I have tried converting colour the way stefohl suggests, but the colour is exactly the same as choosing embed profile and convert in flexcolor.
i.e. when opened in photosop the result is still warmer than what is displayed by the flexcolor preview.
-
I have tried converting colour the way stefohl suggests, but the colour is exactly the same as choosing embed profile and convert in flexcolor.
i.e. when opened in photosop the result is still warmer than what is displayed by the flexcolor preview.
Do you know if your monitor profile is a version 2 or version 4 profile?
-
As far as I am aware it is a version 2 profile.
-
Any more suggestions, anyone!
There must be an answer to this.
-
Just curious. When you output from Flexcolor with Adobe RGB profile, are you given option for rendering intent - Absolute or Relative?
If such option is unavailable, or it may be default to Absolute, how about try to output with ProPhoto?
-
Many thanks for your reply jc1 I will give it a try.
Do you mean to output as prophotoRGB from Flexcolor and then convert to Adobe RGB in photoshop, as we use Adobe RGB 1998 as our standard colour space.
-
I have just tried the prophoto profile ( converting to Adobe RGB and not converting ). I have found absolutely no difference between either result. The image still appears warmer in photoshop than in the Flexcolor Preview.
Help! anyone any other suggestions?
-
Just thought I would refresh this thread to see if anyone has had any bright ideas about the problem!
-
Just a thought but: When asked about a similar issue with Phocus - how does one control the output space colour conversion Perceptual, Relative or Absolute; Hasselblad said
'[It] is not possible at the moment. The conversion is done via the operating system's color engine'. (And because I'm using Windows) ..' it will be the Microsoft ICM that is used'.
So maybe if you want to mimic Flexcolor's conversions in P'shop change the colour engine it uses for rendering.
At least with Flexcolor - camera, you can strip a calibration image's output profile, create a custom profile and therefore have a a less messy set of output conversions. With Phocus we are limited to having to always output with some kind of a profile pre-shaping it.
Ps if you are looking at Pro Photo RGB for a working space you might find Pro Star RGB slightly better - (it works like Pro Photo but is following L* lab linear curve rather than the 1.8gamma, so will translate better when outputted into other media - Web / Inkjet / Litho.
Alex
-
Here's my suggestion to find out where could be the source of the problem.
Obtain these files from Flexcolor station.
(a) scanner raw image in tiff (scanner raw: RGB data in tiff before profile being assigned)
(b) converted image with Adobe RGB in tiff, from Flexcolor.
(c) scanner profile for (a)
Test to be conducted with CS5
(1) Import the scanner raw image (a) into CS5, assigned with scanner profile (c), converted to Adobe RGB
Do it for both Adobe ACE (with RelCOl) and Microsoft ICM (with RelCol and AbsCol).
(2) Import the tiff image with Adobe RGB (b) into CS5
Questions
(i) Compare images, are (1) and (2) same?
(ii) If they are not the same, is the neutral in (1) correct or still too warm? Which is closer to the "preview" in Flexcolor?
(iii) Is (c) a the custom profile, scanner generic profile, or a dedicated profile (for the negative) that comes along with Flexcolor?
Alternatively, if the files can be made available for investigation,
drop (a), (b) and (c) here. < CLICK (https://dropbox.yousendit.com/Send-file-to-jc1-DropBox) >
Do not drop (a) and (b) together if huge files are expected.
jc
-
Many thanks for both your suggestions delfalex and jc1. I will work my way through both of your experiments ASAP.
I will try to upload some results to you jc1 as soon as i can.
Again many thanks.
-
Re post by Jc1.
Images (1) and (2) are the same. But the preview in Flexcolor before scanning was cooler than the files that were produced.
It is the difference between what is shown in the preview against what is actually achieved in the final scan, that is the problem.
As delfalex suggested I have tried using the Microsoft ICM but with no difference.
Anyone any other suggestions?
-
I wonder if this is just a common issue. It seems completely ridicilous that it would be but maybe all machines are like this and we are particularly sensitive.
Have you had a chance to try another machine? Or only done test on yours?
-
Just Like you if I create a 3f File, I'm able to get 99.9% color accurate scans from what my preview shows.
-
Using the 3f route is a much longer and more complicated process when we are trying to scan as many as possible and as quickly as possible!
Did you ever get any joy from Hasselblad about this problem, from the email address I gave you?
-
I should also add that this problem was not evident many years back when we had just purchased the scanners. It is a problem that appears to have developed over time.
In case anyone suggests we have tried changing bulbs but with no difference.
-
I reviewed all the posts just recently and I decided to investigate the possibility of problems with display icc profiles.
I have now used a MAC (I usually work with PC) to scan a negative direct to tiff file, and viewing the image on a completely different screen, a LaCie screen (I usually work with an EIZO screen)
The scan direct to tiff is again consistently warmer than the preview in Flexcolor.
While the scan as 3f and convert to tiff from 3f in Flexcolor, is almost identical to the preview in Flexcolor.
Again I have to say, there must be an answer to this problem.
Everyone has been very helpful in suggesting possible solutions but none so far have cured the problem.
There must be an answer.
Anyone got any more suggestions?
-
I am now clutching at straws with this!
I thought that the GPU enhancement (enable open GL drawing) within Photoshop may be the cause of the difference in colour.
I opened a file scanned by the Flextight scanner, in Photoshop with the GPU enhancement switched off and then switched on.
Unfortunately there is no difference in the colour.
This appears to be a similar problem to what is being discussed in this thread (although with a different programme)..
http://www.luminous-landscape.com/forum/index.php?topic=62029.0
I have to admit that just like "ippolitois" I am not smart enough to figure this out, so any help would be appreciated.
-
We have just had a visit from an extremely helpful Hasselblad engineer. He was kind enough to pop in to see us because he was up in our area.
It appears that the colour shift that we have been experiencing is because we are looking at a low-resolution preview that cannot accurately display the colours. If we do a higher resolution preview by using the detail window then the true colours are displayed. The resultant scan will then match the image in the detail window rather than in the low res. preview window. This colour shift appears much more prominent on negatives with a very narrow colour bandwidth. Negatives with a greater bandwidth do not exhibit the same degree of shift from low res. preview to resultant scan.
Of course this means making two previews for every resultant scan (as the detail window requires that part of the neg. to be scanned at high res.).
Great emphasis was made on the benefits of scanning direct to 3f format and making corrections later, as this requires only one scan, which could be quicker.This is something that we will investigate soon to see if the 3f workflow would be a much easier and quicker way to scan negatives.
After struggling with this for nearly 2 years I can only thank the engineer very much for taking the time to come and see the problem at first hand and showing us why the problem is occurring.
-
Search for U-tube videos by Ian Scovell (ianscovell001). The videos you are looking for are called profiling an Imacon Scanner and consists of three parts. Very informative. Basically he turns every thing off in Flexcolor and scans to 3f. He then created a tiff from the 3f and opens it in PS where he applies his custom scanner profile.
Good luck