Trying to understand soft proofing
I asked a local print service to send me some ICC profiles for their paper / printer combo, and they did. Now I am using those ICC files to soft proof my images in C1 before I have them printed by this 3rd party service.
I have an image with a very nice, saturated black:
When I activate the preview of the soft proofing receipe, the image starts to look washed out, with deep blacks resembling more a darker grey:
What is confusing is that C1 tells me that this is an even deeper black (1,1,2,1) than the previous one, but to my eye it is clearly not.
Are we supposed to increase contrast and deepen the blacks here in order to get the print we want? Or do I misunderstand the whole concept of soft proofing?
-
Are we supposed to increase contrast and deepen the blacks here in order to get the print we want?
Well, yes and no. In general yes, the softproof gives you the possibility to adjust your image in order to compensate what the printer/paper/ink combo would print, but only to some amount. Both your readout values 5 and the softproof 1 is already very dark near black, I would not reduce this further.
Prints usually don't have the same dynamic range of your camera sensor and also less than your monitor, but this grey you see on the monitor will usually be, or at least appear to be, a little bit blacker if you have the print in your hands. Other apps like PS have black point compensation as an option for softproofing, C1 doesn't, though I don't know if that would be a big difference.
Take care that in the preferences you have set the Color / Rendering Intent to what your lab requires for a certain paper. It can be different for different paper/print media.
Softproofing also allows you to see the difference for different papers (obviously:-), in case you care for very deep blacks in this image consider using a paper which shows a deeper black when softproofing.
Softproofing is not perfect but it gets you half way there, make hardproofs also (actual prints from your lab) maybe with crops to keep the costs down.
Usually you need to provide the image as a jpg or tif with a standard ICC profile sRGB or AdobeRGB to your lab, don't export using the paper icc profile in your export settings, unless your lab explicitely requires that (I've never seen that yet).
You have a profiled monitor? Even then beware that usually the prints tend to be darker than they appear on screen even with as low as 60cd monitor brightness in my experience. Probably because I don't use the standard light source from a photobooth, when I look at a print, but normally a less bright one.
Experience with actual prints or smaller hardproofs is valuable especially before you print big and costly.
I tend to brigthen the midtones a bit for every print (btw I keep a separate variant for the adjustments I make for images I send for printing).Did I forget something? Probably yes. Trial and error.
One last thing: If you know where the print will be looked at, maybe at a wall, it is beneficial if you know the lighting conditions, brightness and color temperature, so that you can adjust your variant for it. E.g. if my print is being lit by a 2700K LED spot or halogen bulb I adjust the white balance differently than if it is lit by a 4000K LED spot, I actually have both. I don't use there actual numbers but I slightly adjust for it by taste and experience, depending on the subject and colors or greys in the image.
If the main light source is only daylight (and or the print is looked at mainly during the day) then it can be difficult because in winter or summer or miday or morning, where are the windows headed to, etc. can make a difference.
2 -
Softproofing is not perfect but it gets you half way there, make hardproofs also (actual prints from your lab) maybe with crops to keep the costs down.
I think you pointed out some valuable, practical things to consider. Color management is not "done" with buying a calibration tool and using some profiles. This is what I learn now.
What I noticed is that this washed out impression is mainly when proofing matte papers. Glossy soft proofs usually have richer blacks, but also color tints. Just something to consider, maybe.One of the interesting things is the white balance of my eyes (or my brain). I can look at an image and cleary see a yellow tint. Then I "calibrate" my eyes by looking at a white sheet of paper under a daylight bulb. Suddenly, the yellow tint of the image disappears. So yellow and blue tints seem to be reinforced by our perception and some kind of assumption about pure white.
0 -
Yes, mostly matte paper. But it could be interesting to print on matte paper and see how the print actually looks, maybe compare with the softproof and a glossy paper as well.
Indeed the brain is a super computer, e.g. if you look at something which is supposed to be white your brain tells you it is white though it might have a yellowish, greenish or pinkish cast. On the other hand, the eyes are not perfect, if you look at something blue for a while then the blue perception chemicals will reduce and with thischemical imbalance a (really) white paper will look like the opposite color, yellow in this case of blue.
Color management is not "done" with buying a calibration tool and using some profiles.
Yes, but without it it is much harder or next to impossible.
Cheers,
BeO0 -
So what is the workflow regarding all those profiles?
1. I calibrate my monitor and use that profile in C1.
2. I use the lab's profile to soft proof and make some corrections.
3. What profile do I use in my exported file that I send to the lab, just a generic sRGB profile? I am asking because the exported image with the generic sRGB has slightly different colors to the one from the calibration tool.
Edit: I think I found the answer: https://support.captureone.com/hc/en-us/community/posts/4409692293521-Correct-ICC-profile-for-export0 -
1.: Not sure what you mean with "use that profile in C1" but yes, you should set your monitor to a monitor profile which you have created with your calibration or profiling software. I use the Eizo "Color Navigator" app (which support Eizo hardware calibration) and only set my desired profile in that app, no other setting in Windows or C1 needed, but your profiling application might require something different, I just don't know. C1 should pick up the set profile automatically.
2. Yes.
3.: For exporting to a JPG file I suggest
0 -
Not sure what you mean with "use that profile in C1" but yes, you should set your monitor to a monitor profile which you have created with your calibration or profiling software. I use the Eizo "Color Navigator" app (which support Eizo hardware calibration) and only set my desired profile in that app, no other setting in Windows or C1 needed, but your profiling application might require something different, I just don't know. C1 should pick up the set profile automatically.
Okay, I think what I understand is that I just need to set the calibrated ICC profile in my operating systems color management. Then in C1, I don't need to set this as my main ICC profile anymore, instead I can just use the generic sRGB.
0 -
Okay, I think what I understand is that I just need to set the calibrated ICC profile in my operating systems color management.
Yes.
Then in C1, I don't need to set this as my main ICC profile anymore, instead I can just use the generic sRGB.
You can set the proof profile you want, sRGB, Adobe RGB, printer profile from your lab, or "No proof profile" (see menu View>Proof profile).
You can enable/disable Recipe proofing (either with the eyeglass icon or also via View menu) for quicker change of proof profile in the export recipe tool.
If you set "No proof profile" my understanding is that C1 will take the values from the raw file (which have no color yet) and demosaic these values to RGB pixels using the camera profile set in the Base Characteristics tool, then C1 uses the monitor correction profile (from the operating system color management) to convert these values to values within the monitor gamut and also correcting for color errors of the display. Here is also where the Rendering Intent should be used by C1 (actually for every color space conversion).
Softproofing with a specific profile converts these values into the color space (or printer gamut) as described by the selected profile also correcting for color errors of the display by using the set monitor profile in Windows.
Only for actually exporting the color correction profile of the monitor will not be used.
I am not an expert, but I believe that for a better soft-proofing for prints done in a lab this process should have an additional conversion step. I don't know if I am right because I don't know any software which is doing this.
- If the scenario is that you will export a JPG file and send it over to the lab, then the color values from the camera profile after demosaic the image raw values should be converted to sRGB and then converted to the printer icc profile, then color corrected for color errors of the display.
- If the scenario is that you will export a TIFF file and send it over to the lab, then the color values from the camera profile after demosaic the image raw values should be converted to Adobe RGB and then converted to the printer icc profile, then color corrected for color errors of the display.
The second option would preserve more colors (or convert them differntly), namely the colors which are outside sRGB but inside AdobeRGB, because many printers are able to print colors outside sRGB.
This additional conversion would then simulate an export to a file with color space xxx then print with paper yyy. The current methods don't allow to simulate this export step.Btw, if your are not printing or send for printing but publish in web as JPG then you could also consider to set your monitor profile in the OS to sRGB (calibrated of course). I find it more reliable than the simulation (softproof to sRGB) of Windows + C1, but I can't put my finger on it.
For all matters SFA has brought up I can only say that I fully agree. :-)
0 -
The hard proofs from the lab came back. What I noticed is that they are a bit too dark (as predicted), but also too warm. I assume this could mean that my monitor's calibration is a bit too cold (6500K). Or maybe I am simply going down the rabbit hole of trying to match the colors and tones of my local images with the equipment of a third party lab?
0 -
Yes, 6500K seems a bit too cold.
I wouldn't go down that hole (or at least not yet). Ask your lab or just use 5000K which is kind of a standard setting for printing and compare them with the hard proofs you have.
0 -
I am curious, did you make more experience you want to share?
0 -
Hi BeO,
Well thanks for asking. I am a bit stuck with the complexity of the whole post processing/printing process at the moment.
I bought some monitor calibration hardware (Datacolor Spider) and used that to calibrate my monitor. Then I used the printer profile from a well known lab in my city to get a soft proof of my images in C1. It seemed to show a strong magenta tint in all my images, so I did some color corrections. My expectation was that with those corrections, the images would be printed in the way that I had envisioned them on my screen at home.
Then I personally went to the lab with my corrected images. To my surprise, all the corrections I had done to get rid of the magenta tint were completely useless.We corrected the brightness and colors of all my images on their monitors, using the inital, pre-soft-proofed versions of my images as a starting base.
A few days later they contacted me again and I went to look at the first prints. They all seemed to have a blue tint. At that point, the entire process become so opaque (and expensive) to me that I stopped pursuing it for the time being.
Best regards0 -
Color management is complex, and remote diagnosis of a complex matter even more. But a few things come to my mind:
Make sure your image file is in a standard space like sRGB (export recipe).
Open the image in C1 or better in another app which supports color readout on a bigger area than 1 pixel (e.g. in Photoshop? or in darktable, it is for free, maybe overkill for that purpose but it has that option), then select an area which is supposed to be white but shows a strong magenta tint and check the RGB numbers.
When I started with color calibration I read reviews about monitor calibration hardware and chose an x-Rite. I have read that other devices use organic color filter material which can degrade over time, I think Datacolor Spider was amongst them. But, this is quite a few years back and might have changed/bettered.
But even with the x-Rite I bought I noticed a light pinkish color cast on my monitor white, I sent it back and ordered a new one, this time a very light greenish tint, but barely noticable and I kept it. What I want to say is that no technical device is perfect, but if the Spyder is responsible for a strong magenta cast (but I am not saying it is) then get another device.Send your files to another respectable printing service. Maybe your city lab has something wrong. Make small prints (resize or crop) to keep costs down, maybe compare prints from both labs on the same paper made from the same files.
I understand that you make a pause, it can be frustrating if you cannot nail down the issue, but don't give up printing completely. If you get it right it is more rewarding than just on the computer screen.
1
Post is closed for comments.
Comments
12 comments