We’ve all seen the ads for washing powders promising whiter whites. One of the ways detergent companies achieve brighter whites is by putting chemicals in their products that fluoresce slightly under ultraviolet light. These chemicals are often referred to as an optical brightener. We can’t see most of the fluorescence but some of it extends into the visible spectrum and makes white clothing appear a little whiter. Paper manufacturers often put the same chemicals in their papers for the same reason – to make the paper look whiter. To our eyes the effect is subtle but the spectrophotometers that we use to create printer profiles are more sensitive to the UV end of the spectrum than we are and so the effect of an optical brightener on them is stronger and so may bias the way they see colour on a particular paper.
To counter the influence of optical brighteners many spectrophotometers have filters that cut out UV light or multiple light sources with or without UV. Simply cutting out UV light is crude but can be effective. However, a crucial element to how we see the paper and the optical brighteners is not just the amount of the chemicals in the paper but also how much UV is in the light that the print is under. That’s why X-Rite have a special procedure for optical brightener correction that allows you to customise the amount of correction based on the lighting that is being used.
So, in this article I’ll go through the options and techniques for coping with optical brighteners and getting the best profile for an inkjet media with high levels of optical brighteners, and also look at a few common light sources and how they might interact with optical brighteners.
UV cut or UV included
Most modern spectrophotometers include multiple light sources. My X-Rite i1iSis has a M0 (Type A) light that includes an element of UV and also a M2 light source that doesn’t have any UV. The i1iSis 2 and i1 Pro 2 also have an M1 (D50) light source that includes UV. With the i1 Profiler software you can choose to scan with just the one light source or perform multiple scans to capture data with them all. For my tests I used my i1iSis to measure a glossy inkjet paper that I suspected would have high levels of optical brighteners and also a specialist proofing paper that I knew would not. It is usually glossy and semi glossy photo inkjet papers that contain the most optical brighteners but they can also be present in brighter matte papers as well.
A recent upgrade to i1 Profiler included a very welcome extra feature. You can double click on any measured patch and be given a wide variety of spectral and colorimetric data. If you’ve done a multiple scan under different light sources you can compare the results. I printed out a profiling chart on each of the two papers and measured them with the i1iSis using the dual scan function to capture both M0 and M2, so with and without UV.
As you can see in the screenshot above the glossy paper clearly had a lot of optical brightener in it whereas the proofing paper had very little. I then saved both sets of measurement data (UV included and UV excluded) for each paper and used the Data Analysis function in i1 Profiler to compare them. The glossy paper showed large differences between the two sets, up to a maximum of delta E of over 10 and an average of over 4. The maximum difference for the proofing paper was less than 0.5 delta E and the average only 0.06. So for the glossy paper there was a very large difference between the two sets of data. This means that any profiles made from each data set would give very different prints. On the other hand, for a paper with no optical brighteners there was no significant difference in the data and so you could expect very similar results made from profiles from each data set.
When I printed my test image on each paper with both the UV included and excluded profiles the results were pretty much as I expected when viewed under daylight. You could clearly see a large difference between UV included and UV excluded on the glossy paper. The UV included result was closer to the image on my EIZO ColorEdge monitor. The UV excluded print was yellower, because the paper was seen as less blue by the spectrophotometer. Even with the proofing paper where the difference between the prints was very hard to spot the UV included result was just a little cooler and closer to the image on screen. So, optical brighteners can’t just be ignored. The effect they have on the appearance of the print is important and whilst spectrophotometers do see more UV than we do if you cut out UV all together then they don’t see the paper in the same way that we do. However, as I said earlier, a lot does depend on the light source and the ideal way to correct for optical brighteners would be to take account of the actual viewing light used.
Optical Brightener Correction
Realising that simply omitting or including UV in the measuring light source was too crude X-Rite developed their Optical Brightener Correction (OBC) technology. In i1 Profiler there is a special workflow that creates an ICC profile for a paper using a multiple scan measurement and then prints out a special test chart of variations of grey tones. That test chart can then be examined under the viewing light and compared to a reference chart. You can either use a specific set of grey references or use the ColorChecker Proof card. You can then tell the software which patches most closely match the reference and a customised optical brightener correction is made to the profile that compensates for the paper, lighting and your perception.
I ran through the OBC procedure for the glossy paper, using daylight, and ended up with a print that was very close to the UV included one, but possibly a little closer to the image on screen. The one problem with the procedure is that it is very subjective and it can be hard to choose between the very similar levels of grey. Also it is very much based on whatever light you are using at the time. So, if you know the lighting the print will be viewed under, such as in a gallery, you can customise the print accordingly, but if you don’t know the final viewing light the procedure isn’t quite so useful.
My next step was to view the test prints under a variety of common light sources and also my D50 controlled lighting view booth. I then used SpectraShop 5 to measure each light source to see which, if any, had a large component of UV. The common lights were an old style incandescent light bulb, a fluorescent tube, a cool LED spot light and a LED light bulb. Whilst I had the ambient filter on my i1 Pro 2 I also did a quick measurement of the cloudy, British, November daylight I had been using for the daylight viewing. In each case either the UV included or the OBC profile looked best, but under some of the light sources there wasn’t as much difference between them and the UV excluded result as there was under daylight. Interestingly daylight had far more UV than any of the artificial lighting, with only the D50 view booth displaying much output below 400 nm, the area of the spectrum where UV starts.
The examples above are based on a couple of particular papers under specific lighting environments. Results might vary from paper to paper but whenever you are profiling inkjet media it’s worth doing multiple scans under different measurement conditions to check if a paper has significant levels of optical brighteners. If it does then it could well be worth the time creating and testing profiles made both with and without UV. If you are having a profile made for you by a profiling service then ask if they create profiles with UV included or excluded. I’m quite happy to provide both. If you know the final viewing light then I would recommend going through X-Rite’s optical brightener correction workflow to get the very best print you can.