Prints still too dark after calibration (canon i865)

Wilt wrote in post #12427040
Have a look at this monitor check from Dry Creek photo http://www.drycreekpho​to.com …/monitor_sensit​ivity.html

Yeah, I've seen that one. But I have no monitor calibration or printing issues. (Knock on wood laminate!)

Count me lucky, I guess. I don't have to soft proof or make any adjustments to compensate for the prints my lab makes. WiSIWIG

NEC PA241, BasICColor, Spyder 3 puck. 95 cd/m2 Gamma "L" (perceptually linear response curve).

So, one way to look at all of this is to remember that the print is the "correct" version of the data. If you didn't have a display, but you just had your file and a big button to print that image, you take the display out of the equation and the printer renders the data.

When you make changes to the RGB data so that your print matches your display, you are sort of working in reverse. THe print, viewed under your reference lighting, should be the standard against which you try to match your display. That way, any changes you make to the RGB data on your display are for purposes of enhancing and modifying the "correct" data to make it look better, more awesome, etc. in print, as opposed to modifying the correct data to make up for an incorrect display.

That being said, it is all easier said than done. Controlling for the variable parts of the workflow will go a long way toward getting everything in line, assuming that the equipment you are working with has the capability to conform to your workflow environment.

As an example, here are a few images where I generated a series of display profiles of varying white points, at a luminance that matches my reference lighting. In this case, I use Solux lamps, positioned at a distance from the wall where I hang prints such that a display target of 140 cd/m2 gives me a good match. The Solux lamps are rated at 4700°K - but the paper reflects that light at a potentially different CCT, so I need to measure that and take that variable into account. So, I place my colorimeter about 10 inches away from a white piece of Epson Ultra Premium Glossy paper hanging where I examine my prints and I measure the reflected luminance and CCT of the white point for that light+paper system. Between these measurements and my perception of the targeted values, I will try to get a match between my display and my print.

I printed a test print, using the Epson paper profile for my printer, with Relative Colormetric rendering intent and black point compensation. I then use this print, viewed under my Solux lighting as my Reference. I then adjust my target values in my calibration and profiling software to create profiles that will, hopefully, yield a match.

In this experiment, I fixed the luminance target at 140cd/m2, L* gamma curve, minimum black level. I then made profiles with varying white points at D50, 5250°K, 5500°K and 5750°K. I then photographed the print under the reference lighting as well as the image on the display with each of the 4 display profiles. For photographing, I fixed the exposure in manual mode (ISO, Aperture, Shutter) as well as the white balance (daylight). This way, the exposure of the image will be the same regardless of whether I am imaging the reflected light coming off of the print or the transmitted light coming off of the display. If my target luminance is correct, the exposures of the print and the display should match.

For WB in the raw conversion, I color dropper white balanced on the white (unprinted) margin of the print viewed under reflected Solux light. I applied this white balance to all of the images (ie, the print as well as the display captures). Again, I am establishing the print white of the glossy paper, viewed under the reference lighting, as neutral and any changes in the white point of the display profile will be in comparison to this target white point reference.

Here is an sRGB digital version of the test image:

Nice work, Kirk. Good example of what Andrew Rodney preaches... The only correct value is that which gives you a good match between screen and print.

Thanks Kirkt

Honestly as a beginer thats very confusing to me but like i said earlier thats my lack of knowledge - hopefully one day ill understand what you said


Thanks

Thanks Kirkt

Honestly as a beginer thats very confusing to me but like i said earlier thats my lack of knowledge - hopefully one day ill understand what you said - it went right over my head not because of your indepth explanation but because i didnt undertstand half of it

you obviously either do it for a living or are very passionate about this - i at the moment am just getting started but feel disheartened by the amount of knowledge (and my lack of it ) that has to be learnt to ontain a good print - (they wonder why people buy point & shoot)

kirkT

so i take it from reading your comments i should try to calibrate my monitor from the original print without any modifications - think thats where i might have gone wrong i printed after mods and then tried too match the screen to it

I'd start by taking a deep breath and then figuring out what you can control easiest. You have a printer with the printer-paper profile matched to your paper. Good. Use the link to the test image I provided and print that test image (follow the link to the test image, don't simply copy the reduced sRGB image I posted). THere are others on that web page as well if you want to try a couple of different images. Make sure you take note of your print settings, let photoshop manage the color conversion, use either relative colormetric or perceptual, and make the print. Let's assume that your printer is working correctly and nozzles are not clogged, etc. If there seems to be some weird looking output, check to make sure everything is working on the printer end of things, you aren't double color managing, etc. Let the print dry overnight.

I would consider thinking about your work environment and the lighting you use to view your print. You don't need anything fancy, but even a single Solux bulb, or similar controlled light source, in a low-wattage lamp will suffice. Using "daylight" may work okay, but daylight changes and is not available for at least half of a day. Ha! make sure that when you get your viewing light set up you are not illuminating nearby surfaces that have a strong color that may cause a color cast to reflect onto the print. You want your viewing area to be as neutral as practical. If you are going to hang prints on a wall near your display, you probably want to take down that sweet Dungeons and Dragons poster of the red dragon and relocate it somewhere else.

Once you have your print and your lighting, you are set to start figuring out what targets you want to establish for your display. When you calibrate and profile your display for these targets, remember this is just to match the specific paper you are printing on - for general viewing and use of your computer, you may want to stick with something generic like 120 cd/m2, D65, etc. That is your personal preference. However, this print-matched display profile will be used when you are working on images that you intend to print.

The usual approach is to illuminate a piece of blank paper (say you are printing on glossy photo paper) with your reference lighting set up the way you will be viewing your prints. You can then take reflected light readings off of this blank paper using your colorimeter (spyder, iOne, whatever you use) to measure the white point and luminance of the light being reflected off of the surface of the paper. Typically you hold the colorimeter about 10 inches or so away from the paper, taking care not to cast a shadow on the paper and making sure that the angle at which you aim the colorimeter does not cause the colorimeter to see glare on the paper. Your calibration/profiling software should give you the option to take measurements in this situation and give you a reading of the white point and luminance of the reflected light. This will provide you an idea of what you are trying to match on your display. Note that these values will be starting points, not chiseled in stone values from which you cannot deviate.

You will notice that moving your light source closer to, or further from, the paper will result in changes in the reflected luminance reading, as one should expect from the laws of physics regarding light falloff (the further away you place the light from the paper, the less light reaches the paper, right? Google"Inverse Square Law Light" if you care!). That is good, because you can change the position of the light (and therefore the reflected luminance) so that it falls in a range of luminance you want to set on your display - remember that you don't want to set your display too bright or your eyes will start to hurt, but you don't want it too dim that your are compromising its performance or you are unable to see what you are doing. Find a comfortable brightness and sync that with your reference lighting, noting the measurements your are taking off of the blank paper your are illuminating.

Once you get the luminance of the display set up properly, then you can run through a range of white point values to see which suits you best and matches your printed output best. I would start at D50 and work up to D65, in 250°K increments. I find that even though the colorimeter may be reading 4200 °K off of the paper, that white point is too low (warm) when I try to target that white point for my display. Give the reflected white point reading from your paper a shot though and see how your display reacts. Usually, your calibration and profiling software will give you some results of the accuracy of the color reproduction, etc. of your calibration and profile.

In the end, all of the measurements will take a back seat to your visual perception of what matches. The combination of printer-paper profile, lighting and display calibration that looks like a match is the combination that is correct!

Good luck - this is frustrating stuff. It is best to approach the process methodically, otherwise you will be chasing your tail. Also remember that you may end up hitting the limit of your equipment in terms of how good of a match you can achieve, so you may have to compromise to get close enough and then make some minor adjustments to your file to get the exact output you want. You'd like to avoid this last step, but it may simply be unavoidable given the equipment you have - at least you will have a better understanding of why you need to do this, should your workflow require it.

Kirk

thanks kirkt

sorry but i dont see a link

The link is in my first longwinded post with all of the images. Here it is again - scroll through the various images:

http://www.northlight-images.co.uk …le_pages/test_i​mages.html

Also, there are several good articles on the northlight site.

Kirk

Thanks kirkt

it is very frustrating - as i metioned earlier im a beginner i dont even know how to use the spyder to take a light reading from the paper i know how to put it on the monitor and and calibrate - d50 d65 250k increments i have no idea what that means i appreciate you know alot more than me ( thats an understatement) and appreciate your help but even down to illuminate a piece of paper what from the front , back , side you can now see how little i know and how frustrating this really is

one thing i will say is that tset print - pribnted absolutley beautiful if all my prints were as good as that i would be more than happy

For practical purposes, D65 is equivalent to 6500k. D50 is equivalent to 5000k.

Wow, kirk, that's much more indepth to calibrate than I did. I can appreciate the breaking it down into a methodology to get it done. Never thought about that. And it seems you worked backwards, starting with print/printer/paper, then lighting then calibration/monitor, which is exactly opposite of what I did. May have to try that one.

One question though - if you start with print/printer/paper, that assumes that the print is right to begin with, which may not be the case. How would you work around it if you know the print isn't correct?

When you work with a standard printer test image, you assume that the image file itself is edited correctly. That's the definition of any reference standard.

I take the position that your monitor must first be calibrated to SOME standard before you can start comparing print output to it.

i give up

thanks to everyone who offerd help esp kirkt you have been great, i think its me being a beginner whos finding it hard to grasp

thanks again