What's new
Photoshop Gurus Forum

Welcome to Photoshop Gurus forum. Register a free account today to become a member! It's completely free. Once signed in, you'll enjoy an ad-free experience and be able to participate on this site by adding your own topics and posts, as well as connect with other members through your own private inbox!

jpeg color does not match Photoshop color


The images in dropbox are comparisons of PS and firefox and chrome from today.

The monitors were calibrated on Monday but when saving the calibration I don't remember which name I used. Usually "laptpo". It usually defaults to one and I use that but sometimes I am unable to and then add a 1 or 2 afterwards. I don't believe I did that on Monday.

The display 1 states Generic PnP Monitor - can I assume that is my laptop?
If so then I need to change display 2.
However I don't know how to change it.
 
Hi Puppychew

As a side topic. If Tom, myself, or other Forum members cannot help you with root cause on your color calibration problems, there is a paid service support options with Datacolor and your Spyder product at: http://spyder.datacolor.com/consumer-support/#jtabs-2 At some point time is money and you will further ahead with paid resources to solve your problem.

a) I suggest redoing your hardware calibration since that does not take too long and be sure type in a name for the profile that is distinctive and includes the date so you don't have any confusion. Also, note, as I understand it, the final page of the Spyder 4 calibration besides given you a naming option, also shows the exact folder path where it is going to store the Monitor Profile. We want to make sure we don't introduce more areas on confusion as we have enough of those already.

b) Verify that in the Windows Color Management options that under the Devices tab you can locate and assign that profile to your monitor.

c) I think we already covered this yet to identify your monitor in a multi-monitor setup, when in the Devices tab of the Windows Color Management, you click the identify monitors button and a "BIG" number should appear on the screen. Below is an iphone snapshot of my screen for that interface. Notice that
- When clicking the Identify Monitors a big number "1" shows up on my screen
- That I have the checkbox clicked to use my settings for this device
- That the dropdown I have it set to show Display: 1..... to match the big #1 that popped up
- And that the associated ICC profile in the box below is displayed (which is where your Spyder 4 profile should show up and I would remove all other profiles so there is not confusion what is associated with your monitor)
- Please confirm that you were able to complete the above for your monitor (most likely #1 or #2) and use a screenshot to show the dated Spyder profile is there and assigned

iphone screen shot.jpg

d) You did not mention yet have your restarted Firefox at some point so the change settings are in place (just trying to make sure something does not fall between the cracks)

e) assuming all of the above steps, please put adobe and firefox side by side and see how they look. If not we will move on to another step.
 
I think it is best to do one thing at a time. I will contact Spyder for assistance on Device 1 vs 2. Their office is closed right now (new Jersey time) and I will contact them first thing tomorrow. I really appreciate your patience!
 
Your welcome. BTW, if a second monitor is not connected, it will not show the "1" and "2" per what I have read. When you are just using your laptop, what I have read is that a number "1" will now show and that info for a second monitor will not show either. More expert help on the Windows front is advisable since I primarily use my Mac. Let us know how it goes.

ADDED EDIT: It is also possible that what displays for me on Windows 7 might be different as I am running a Virtual Machine (through that is pretty rare)
 
Last edited:
Ok - I contacted Spyder and the Display 1 and 2 are correct. I had thought it should be different names but no. The large numbers show up on the correct monitors -1 being my laptop. They did instruct me on doing a more intense calibration which I did. In dropbox I have screen shots for both displays and named the calibration july ___. I also redid the PS vs firefos and PS vs chrome images labeling them new in dropbox.

I will be going away for a long weekend, returning Monday. Please let me know the next step.
Thanks!
 
Last edited:
Hi Puppychew
Great that you got some help with Datacolor folks to verify your settings and also go for a more advanced calibration.

The bad news is we are still seeing a saturation difference.

Here are the next steps I suggest

- Post your ICC profile to your dropbox so it can be examined. Most likely it is just fine yet wanted to cover the bases.

- One of the potential issues would be that your profile is created in version 4 ICC yet Firefox is not set to accept Version 4. This does not always cause problems yet through this would be good to check first before trying to insert the ICC Path directly in FireFox

Go to this website and follow its instructions: http://www.color.org/version4html.xalter

If Firefox is only version 2 compliant and not version 4 compliant then it will look per the website like:

Supports_V2_Only.jpg

The image will look totally fine if both V2 and V4 are supported.

This only matters if you ICC profile is V4 yet a possibility.

If you get the image above, got to about:config (inserted in URL space) and in the search bar type "color_ " (yes with the underscore yet no quote marks)

In the image below click or doubleclick on the configuration setting so that V4 is enabled (mine shows as false right now).

Screen Shot 2015-07-02 at 12.43.35 PM.png

Restart Firefox and show PS and Firefox side by side again and see if it makes a difference.

------------------------

There is more after this step yet lets get this one out of the way. Your are really hanging in there. Reminds me of this scene from Monty Python where the Black Knight just never gives up:
https://www.google.com/url?sa=t&rct...rcn360WWm2ua4xFqwR06LRA&bvm=bv.96952980,d.aWw

Enjoy your holiday weekend.
 

Attachments

  • Screen Shot 2015-07-02 at 12.42.14 PM.png
    Screen Shot 2015-07-02 at 12.42.14 PM.png
    113.8 KB · Views: 3
Last edited:
Cool video!
Icc profiles are in drop box. July=laptop July monitor=monitor
When opening the image in firefox, it looks exactly the same.
In config sys it was already set to "true"

Talk to you Monday!
 
I am going to go out on a limb because I think I know what is causing the problem. I could be totally red-faced if I am wrong yet I will even venture that this is solved.

Yet first, a required step is to open up this link:

http://tinyurl.com/n9lyxnk

Turns out I believe that this is caused by two factors:

1) Firefox does not support two monitor profiles and by default is using the primary monitor profile which is the laptop profile
2) The laptop profile is acutally quite smaller then the sRGB test image and the monitor profile is a little larger than sRGB (I examined the profiles in ColorThink Pro). So the sRGB image is converted into the laptop color space (smaller gamut) and those data numbers are being applied into a considerably larger monitor color space. Moving data in smaller color space and using that color data directly in a wider gamut space without converion will increase saturation.


The solution to this problem was actually the next step we were going to take on Monday which is to go into about:config and set the path for the monitor color profile and force the use of that monitor profile. This will of course make colors not look right on the laptop yet it will look right on the monitor.

This would normally not show up as a big difference if both monitors where close to the same color space such as sRGB. The rub being is that the laptop profile is not close to sRGB and the two monitors are quite different in color space differences. Upon inspection in ColorThink Pro, the gamut volume of the monitor color space is a full 50% larger than the laptop color space.

I simulated the Color Management conversions that were going on and what came out was an image that matched what puppychew was seeing. More images and details a little later on yet I "think" this is it.

John Wheeler
 
Last edited:
A few more details on the color mismatch between PS and browsers.

Turns out that in dual monitor systems, PS will automatically adjust to using the correct monitor profile if you moved the window from one monitor to another. Firefox and Chrome do not do this. They access the primary or #1 monitor and use that profile by default. I found this article that goes into more details about this behavior particularly in step 3d: http://www.metalvortex.com/blog/2012/03/16/831.html

This dual monitor color issue does not typically raise its head both monitors have a color profile close to sRGB. However that is not the case in puppychew's case. Here is a 3D gamut plot in Lab space for the two color profiles. The solid color is the profile for the laptop display and the wireframe is the profile for the monitor. Notice that the larger gamut is considerable in the yellows and greens. I also included the 2D project below the 3D object. The black line represents the sRGB color space. The laptop gamut is considerably smaller than sRGB and the monitor is somewhat larger gamut than sRGB:

Screen Shot 2015-07-04 at 12.46.13 AM.png

At this point I thought that this was most likely the problem. I followed this up by following the color conversion process starting with Tom Mann's test image in sRGB (steps below). After following the suspected Color Management path in puppychew's system starting with the original image, I ended up with the same over-saturated image as reported by puppychew. The difference between puppychew's image and the one I created had less that a ∆E of 1 for 10 sample points on the image. That made me pretty convinced that this was strong evidence of the root issue and the solution being to hard fix the monitor profile in the about:config page of Firefox (not sure how to do the same in Chrome). I compared puppychew's and my created image side by side below. The left is the original over-saturated image puppychew was seeing and the right is the recreated match following the suspected color management data path in puppychew's system. The images are visually identical.

recreate-oversaturation.jpg

EDIT: Clarification added to steps below

Here were the steps for recreating the over saturatoin
1) Start with sRGB test image
2) Convert to the Laptop Color Profile with no Black Compensation (simulates the conversion to the laptop screen profile which is the default behavior for Firefox and Chrome in a dual monitor system)
3) Color data is now "Assigned" to the wider gamut monitor color profile since the color data was viewed on the wider gamut monitor, yet no conversion of the color numbers to the monitor profile was done.
This matches the data captured by the screen cut and paste my puppychew
4) Final step was to Convert to Profile back to sRGB so this could be posted and properly seen with any color managed software including web browsers.

It will be interesting to see that once the monitor color profile is hard wired into Firefox is the color mismatch goes away.

Hope this incremental information is helpful.
 
Last edited:
I will be going away for a long weekend, returning Monday. Please let me know the next step.
Thanks!

HI Puppychew
Hoped you enjoyed your long weekend. If you get the chance, let us know if you problem is solved by the suggestions in the last couple of posts. Just curious
 
Very interesting.
I added the monitor profile to FF and redid the samples. They are in dropbox.
I can see the difference!
Does this mean that I am all fixed?
 
Last edited:
I can see the difference!
Does this mean that I am all fixed?

Well Puppychew, if you want to be all "fixed", that would mean a trip to the vet :bustagut:

Ok seriously. Just to look at how close they are visually. I made sure they were all in the proper color space and compared the flowers and face side by side below:

Flower-Compare.jpg

Face-Compare.jpg

Considering some of the limits in the workflow, these are not a perfect match yet pretty darn good.

Note that the ps vs chrome one did not look fixed and I don't know how to configure Chrome to update its targeted profile.

As far as being done, one thing you have to realize is that in a dual monitor system, anytime you update your monitor profile, you will also need to update Firefox. Also, note that if you are using you narrow gamut laptop monitor standalone, Firefox will not have the proper profile for it and all the images colors will be off there as well.

Other then that, I think you are "fixed" :)
 
WOW - this has really been something! I've definitely learned a lot.

My color sample chart was created using individualColor-Sample-Chart---Historic-Colors-srgb.jpg
images from adobe 1998 and moving them into this one image and saving as adobe 1998.
The color used was the rgb number and sometimes I tweaked it by eye to have it be more similar to the paper color chart.

I have now converted it to SRGB. The colors are still off compared to my paper chart. My next job should be re-doing the samples in srgb.

I greatly appreciate all your help - both of you. I started a new business after losing my job and thanks to your help I can do things 100 times better now.
Thank you so much!
 
....The colors are still off compared to my paper chart. My next job should be re-doing the samples in sRGB ...
At this point, because you haven't yet profiled your printer (with a particular type of paper and inkset), you should have absolutely ZERO expectation that the image on paper will match what you see on screen. Fix one thing at a time, and when you know you have fixed something, don't go back and tweak it to try to make it match something that hasn't yet been fixed.

Cheers,

Tom M
 
You're welcome Puppychew

Glad this part of the Color Management workflow is much more under control and that it will help your productivity.

A couple things. Not sure about your original image yet the one you uploaded to this Forum did not have an embedded color profile. Hard discipline to adopt sometimes to include that color space: :rolleyes:

As Tom already started bringing up, the Color Managed Workflow does not end at the monitor:

- When printing, using an ICC profile that matches the printer/paper/ink combination is quite critical. Also the settings in PS and your print driver are pretty critical as well for good colors

- It is also very important that the color temperature with which you create your profiles is the same for the expected viewing environment

- And when comparing the print to the screen (which is really never done side by side yet glance at screen and then glace to the image under the proper illumination to see how close a match you have.

------------------------

Otherwise, you can run into similar issues that you did as per this lengthy thread (just in another part of the workflow).

So I second Tom's heads up that we have only looked at part of the problem/Color Workflow.

Glad we have made a good step forward on your issue (and I also learned some things along the way too)
 
One more thing. According to the linked article I gave in an earlier post of this thread, Safari browser apparently does adjust the color profile depending which monitor in which Safari is being displayed the same as Photoshop. FYI
 

Back
Top