Printing greyscale images and removing unwanted colour tints
Printing greyscale images – unwanted colour tints
Some subtle changes with lighting, and removing colour tints from monochrome prints
Unwanted colour in my prints
The other day I was printing a greyscale image on our Canon iPF8300 printer. I was using the monochrome printing settings in the driver (the Epson equivalent is the ABW mode)
After the print was completed, I happened to look at it under a relatively dim (60 watt) light bulb and noticed that it had a very slight magenta tint. Very slight, but once I’d noticed it I wanted to know more.
Was this specific to the lighting, the paper, the inks or whatever?
Was it something I should worry about?
Lighting and prints
It’s important to realise that a lot of black and white printing involves compromises and assumptions about paper, ink and lighting.
I’m not a big fan of tinting and toning of monochrome images, so I usually want prints as neutral as possible.
Papers vary considerably in their colour, depending on the various substances that have gone into the base and surface coatings.
One component of this is the presence of any optical brightening agents (OBAs)
The role of the OBAs is to absorb Ultraviolet (UV) and near UV (violet/deep blue) and re-emit it in the blue part of the spectrum, to compensate for the natural yellowish tint of most papers.
This has the effect of making the paper seem a brighter white, when there is some shorter wavelength light around. You get this with daylight and some fluorescent lighting, but not with incandescent lighting (tungsten light bulbs).
The presence of OBAs can be a problem when creating ICC printer profiles for some papers, and indeed, my i1 iSis scanning spectrophotometer has a set of printed grey patches on a chart you can use to incorporate OBA correction when building profiles.
- For further info on this, I’ve a review of the iSis OBC (Optical Brightener Compensation) kit. The diagram is from the article, where the bump in the green line shows a peak in blue emission from the OBA in a paper.
One of the problems with the iSis kit is that you do need to take prints and test sheets along to where the viewing light is.
Suffice to say, I’ve not often had print jobs important enough to warrant this, but it does work.
Custom ICC profiles
If I’m making a normal colour ICC profile, then when I create the profile, I can specify the precise viewing lighting.
Normally, profiles are created for D50 lighting – this is similar to a colour temperature of 5000K that you might set for a white balance on the camera or when processing a RAW file. It’s not quite the same, but equating the two is OK for many purposes.
It so happens I can just take along my i1 spectrophotometer and laptop, and take a suitable lighting measurement – once again, it has to be an important job/exhibition with odd lighting for me to think of doing this.
My profiling software allows me to choose from several ‘standard’ lighting types so I can, for example, create profiles optimised for TL84 fluorescent lighting (common in UK retail environments) or good old tungsten lighting.
If I add in methods of refining my ICC profiles for greyscale performance, I can create greyscale prints that optimally match the colour of the paper and the viewing lighting.
Great, but that’s a lot of trouble to go to, assumes I know where the print will be seen and makes use of some not inexpensive measuring equipment and software.
I’ve written elsewhere about using QuadToneRIP (QTR) to produce linearising profiles for B/W print modes, however these only cover linearity of response from maximum black to maximum (paper) white.
What about differences in my perceived colour of the print?
The Canon driver offers a tint correction option.
- (note that I’m showing Apple Mac screen shots, but windows interfaces are similar)
The Epson ABW mode has a similar option for tonal adjustment.
Although both allow for numeric X/Y values, you can see that the axes are different colours and there is no concept of what the numbers actually mean in a quantitative manner.
- The ways the interface works suggest that the adjustments are of ‘a’ and ‘b’ values in the Lab colourspace.
Note that the image shown in the ABW interface does not show the real time effect of changing settings, so don’t rely on it to make alterations.
In the past I’ve not really paid much attention to this option, and as a result I have a whole lot of prints from Epson and Canon printers that I’m very happy with.
Of course, the problem is that once I notice something that might not look right, I need to find out more :-)
It’s important to realise however, that until I pointed out the effect with some test prints, I was unable to get anyone else to notice my magenta tint issue under the tungsten lighting.
- I do actually produce prints as part of our business, so there does have to come a limit where undue attention to detail goes against the aims of the business to turn a profit. I’ve never believed in the perfectionist approach, I just go for extremely good… ;-)
Some test targets
I printed out three copies of our B/W Test Image on Innova IFA-11 Smooth Cotton Natural White paper on the iPF8300.
This is an OBA free warm white paper that I use for some of our B/W print work.
One print used the Canon monochrome driver at its default tint setting, but with a QTR profile created for this paper for linearisation purposes.
- Note that QuadToneRIP (QTR) is a shareware package – well worth the $50. It is regularly updated and runs on Windows and Mac OSX. It is actually a complete B/W print system that gives very good quality B/W prints from a number of Epson colour printers with OEM (manufacturers) colour ink sets and also supports custom black and white ink sets.
The two other prints used custom colour ICC profiles, optimised for greyscale (lots of patches) and created with Tungsten (Illuminant A) and D50 viewing light options.
Each print had some small lettering in the corner, for me to see which was which. I was reassured that most times I had to look at these marks to identify the prints, since the differences were slight.
I often get asked for suggestions about learning more about the nuts and bolts of Colour Management.
My usual suggestion is Bruce Fraser’s Real World Colour Management. My own copy is well thumbed. It’s my first port of call if I’m asked a question and I feel I don’t quite understand an issue well enough to be absolutely sure of an answer.
Check latest price/availability from Amazon.com or Amazon.co.uk
See some other books Keith has on the shelf, on our Books Page
A second target consisted of grey strips printed on two different papers, but at different settings of the
green/magenta(ish) Y axis (-5 to +25 in steps of 5)
One paper was the IFA-11, whilst the other was a very glossy, very white (OBA) Olmec paper (more info). Both had custom media configurations for the printer.
In the examples below, I’ve taken photos with a Canon 1Ds Mk3 camera of the prints, viewed under different lighting.
The shots also included a very accurate grey card to enable me to set the white balance during RAW image processing in Adobe Camera Raw.
The images are designed to give an impression of what things looked like visually, although the colour response of the camera has not been optimised (profiling etc). These test shots are taken in my home and are most definitely not in lab conditions!
- I’ll leave accurate measurements of such differences in prints as an exercise to the reader – one with considerably more patience and spare time than me :-)
If you move your mouse over many of the shots you can see the effect of massively bumping up the saturation and vibrance, just to exaggerate any colour variations. Do remember though, that you’re seeing web images in a browser on your own monitor – I can unfortunately make little in the way of guarantees of precise colour reproduction.
First of all – inside our print viewing stand (PDV-3e – full review)
This gives a very good D50 representation – you can see the edge of the grey card poking out from behind the lower print.
Outside in the conservatory, it’s a cold overcast day – only 5280K needed for the white balance, but this light has more UV and short wavelength visible light in it.
Now for indoors, but with electronic flash.
Note the higher colour temperature, but different effects – this light has a filter removing most UV
This shot includes the grey patches, printed out and stuck on to a sheet of OBA free paper.
The patches on the left are the Olmec glossy paper, and on the right, the same IFA-11 as the test prints.
This sheet is shown in the viewing stand below.
Note how much more blue the Olmec paper looks in the unenhanced view.
Our light box aims to reproduce D50 lighting quite accurately, and it has a UV element, making the paper look much more blue from its OBA content.
The electronic flash picture earlier has a higher colour temperature, but UV filtering.
OK, so I’ve lots of views of printed greys under different lighting, what does this tell me?
First up, my original problem of noticing a magenta tint under tungsten lighting.
With this paper, the mono print mode seems to be aiming at a colour temperature well above any I’d normally expect to view prints under.
If I’m using the driver’s monochrome print mode I’ll therefore make some adjustments.
If the print (on this particular paper) is going to be displayed in relatively dim conditions lit by tungsten lighting, I’ll print it (on the iPF8300) with a +18 setting. Ordinary halogen spotlights gets +12, whilst daylight fluorescent lighting gets +8. For unknown viewing environments I’ll go for +8, since I’d prefer to err on the side of the print looking a tad magenta in dim light, than green in strong lighting.
- BTW…a reminder again of the fact it was only myself that ever noticed the slight tint!
A print made with the IFA-11 paper after my testing…
Secondly, the problems with OBAs – this seems to make quite a difference to the coolness of the print, particularly if the viewing environment has variations in the amount of UV present (some indirect daylight vs. mostly tungsten in the evenings)
I should note though, that the Olmec paper I tried had the most obvious presence of OBAs in it of any papers I’ve recently looked at and hence had in ‘stock’ in the print room.
What to do differently
It seems the B/W print mode performance is a bit more dependent on paper types than I’d perhaps assumed.
This means that when evaluating papers for our iPF8300 I’ll be looking a bit more carefully at differences between tungsten lighting and what I see on our viewing stand.
I’ll also be looking at how one of our other popular papers IFA-49 (Innova Ultrasmooth gloss), performs under different lighting. This paper does have some OBA’s present.
I’ll be looking more at profiling later this year and have already produced a grey optimised profile for the IFA-49 in several lighting variants for further testing. (The IFA-49 is -very- like the popular Epson Exhibition Fiber – just more robust ;-)
- ‘Problems’ with Epson TFP/ExF and the iPF8300
Specialist print software?
I’m in the process of further testing the True Black and White software on the iPF8300 and will be addressing these issues of lighting effects as part of that.
For Epson users, QuadToneRIP (QTR) is well worth a look
For some printers, there is the option of using custom monochrome ink sets. This is not currently very practical in the UK, since many options are only imported from the US, and I know that I’d rather not rely on the vagaries (and costs) of directly importing ink cartridges.
- Article written some time ago about using QTR and a monochrome ink set.
How can this help -your- printing?
If you’ve a decent photographic grey card, run off some B/W test prints and see if your print setup shows a noticeable colour shift. You may need to photograph the print and card to see for certain.
If it does show a change, decide whether it’s worth bothering about, or whether a different paper choice might be better. A good paper for your colour work may not be the best for monochrome.
If you’re printing with the Canon or Epson B/W print modes, explore whether the colour tint adjustment can produce a better overall look for where your prints are most commonly viewed.
If you are making your own printer profiles, then explore the profile creation options to see if allowances for lighting are included (see the Spyder3Print SR review for some examples)
Whilst well made (and B/W optimised) colour ICC profiles can help, they often need to be created for just one lighting condition for monochrome. Also, at the moment I don’t know of anyone supplying custom profiles that produces B/W optimised ones.
If you’re trying to make monochrome prints on a printer that doesn’t have multiple black inks and were wondering if this was going to help get rid of colour casts then I’m afraid you’re probably out of luck. The fix below -might- help.
Fixing a monochrome tint when printing B/W in colour
An approximate fix in Photoshop can be made with a test print, a grey card and a photo of the print.
Photograph your test print and the grey card in even light of the type you want to view the print in.
Process the raw file taking great care to set the white balance from the grey card
Add a curves adjustment layer to the image
Add a ‘Vibrance’ adjustment layer to the image (you can use hue/saturation, but the vibrance controls are smoother)
Turn up the vibrance/saturation until the tint is clearly visible in your print. If you have set grey balance correctly, the grey card will still look grey – if not, go back to your raw file and try again.
Go to the curves adjustment layer, select the grey point eye-dropper and click on a mid grey part of your test image in the photo.
The test image should now look neutral and the grey card will take on a tint – you may need to click elsewhere on the image to get the best overall correction.
This curve can now be saved and used as a correction curve to apply to a monochrome image (in RGB mode) before printing with colour inks on a printer without a specific monochrome print mode.
This is quite experimental, so expect some variation in results.
It works by using the vibrance layer to ‘amplify’ the colour tint in your print, so as to make the curve setting much easier to see.
When I first noticed that slight tint under dim tungsten lighting, I thought it might just be a setting I’d adjusted wrongly.
It’s turned out to be a lot more complex than that and I’ve learnt more about the details of making B/W prints and potential limitations of different media in the process.
Actually, when I look at some of my earliest experiments in B/W inkjet printing (~2003), such colour casts were a common problem, and one reason that I first investigated the use of specialist ‘Mono’ inks in our old Epson 1160 (Lyson SG inks). It was what also first interested me in colour management, so it’s quite fitting that I return to the issue.
I did wonder whether to write up this article, since I felt it had no strong message or really useful hints and tips to mention. Perhaps it’s because the ‘problem’ isn’t even visible to many, I’ve not got the satisfaction of clearly answering a question?
However I do feel it’s been a useful experience that has had the effect of raising the quality of our own B/W print work by another step, and suggested further avenues to explore.