Fine tuning black and white photo printing
Optimising B&W print modes with a scanner
When ICC colour profiles just won’t do.
Printer drivers often offer a specialist black and white print mode. This doesn’t work with normal ICC paper profiles, so Keith has been looking at ways of improving print quality and consistency through optimising print linearity.
It’s quite easy to fine tune the linearity of black and white photo printing, when using using monochrome print settings such as Epson ABW or Canon Black and White photo print modes.
It’s also a great way to help find out what media settings work best for your paper/printer setup.
Keith looks at several ways, going right back to a solution first explored in one of his articles from 2004.
Towards a better and more predictable B&W photo print
Over the years I’ve looked at several techniques for refining B&W printing. Printer driver software, as well as inks and printers have all improved significantly over the years. This has made the whole process much less of a dark art. Custom ink sets, an interesting option 20 years ago have all but disappeared. This is very apparent here in the UK where they are just not profitable to stock/distribute/support.
This particular article is aimed at people wanting to use the custom B&W print mode available in the drivers of many better quality printers. I’ll show it using the ABW mode for the Epson P5000 printer, but the principles apply whatever printer you’re using.
One of my key suggestions for anyone who wants to improve their B&W printing, is to start by printing a known good test image. If you can’t make a good print of this, then what hope for your own photos?
Some years ago I created a (free) test image. It’s available on this site and has an article describing its use and purposes.
The only section of the image we will be using in this article is the step wedge. The example shown above is a full 51 step grey scale, but it’s easier here to use the 21 step version. The step wedge is in the form above for measuring with a spectrophotometer. However, I’m going back to using an old flatbed scanner – something a lot more people will have sitting around.
If you’re interested, I have an article describing the use of the grey step wedge with the latest i1Pro3 spectrophotometer and one with the i1iO. You can use those methods for creating the same type of adjustment curves I’m looking at here, just more accurately (and expensively).
Back to basics
In 2004, one of the earliest articles on the site looked at linearising black and white print output with a scanner. Whilst I’ve added update notes to the article, I’ve not looked again at that method of making measurements. It was actually a recent email asking about it that reminded me (and why I’m always happy to answer questions about what’s on the site).
If we think of a black and white print going from the darkest black to paper white in terms of a percentage (of black) from 100% to 0% then it would be nice if 50% mid grey printed out at 50% and there was a one to one relationship between the blackness in our image data and the level of black that printed out.
Note that in this article I’m not going into any colour management depths – I use the Grey Gamma 2.2 colour space for editing my B&W work (16 bit), or if still in RGB format (as needed for some Photoshop plugins) Adobe98. The curves approach here is for testing and tweaking print output, not some precise workflow.
Having a linearised print output means that I’m more likely to avoid things like crunched up shadows. Also, when using third party papers, checking the initial linearity helps me evaluate the best driver media setting for the paper.
Many media settings suggested by 3rd party paper sellers for a particular paper/printer are educated guesses at best
If your initial check of linearity is pretty close, then think twice over whether further work is really needed? I’ve seen many new printers/papers where a correction isn’t going to make much difference.
See some of my paper reviews for more. A particularly good example of a paper not needing any adjustment was Fotospeed legacy gloss 325
Printing the target
I’m printing the test image from Photoshop – mainly because that’s what I use for my photo editing. If you’re just using the test image and scanning to check media settings, then use whatever software you’d use for normal printing. [click to enlarge images]
Remember that if you’re using other measuring tools than a scanner, that the image is NOT a profiling target. Only ever print it as you would a normal B&W photo.
Since I do a lot of testing I have stored printing presets. Here’s one for Epson Premium Luster 260 paper (the media setting) and A3+ paper size. It’s using the Epson ABW print mode. For windows and other print software, settings will be whatever you’d normally use.
A quick check shows me that the default B&W print mode is selected.
You can do the measurements with variety of settings. Just remember to write on the print what they were.
If you’re short of paper, then just crop out the test wedge from the test image and print that.
The paper is HP Advanced Satin-matt photo paper 250gsm, from a huge box of sample packs I was sent by HP years ago. I’ve no idea what settings are needed. I’m just guessing at Epson premium Luster 260 on the P5000.
For any measurements it’s best to leave the paper to dry for a bit. More so with matte art papers, but this one was left for a cup of coffee to dry.
Scanning the print
I’m using my old Epson Perfection 1200U flatbed printer. This scanner is over 20 years old and works just fine with Vuescan software.
Vuescan supports thousands of perfectly good scanners. This includes my Canoscan FS4000 film scanner and even my UMAX Powerlook III transparency scanner (via a SCSI->FireWire adapter). I hate seeing old useful kit thrown away, and Ed Hamrick, the writer of Vuescan deserves all the praise he gets. [Vuescan – free trial]
Vuescan is easy to use – I just happen to have it running with the ‘show all the options’ mode enabled.
I only really need the patches scanning, but I include the bullseye gradient – this shows up print problems really well.
Here’s the scan of the print – the print looks fine. However, do remember that you are seeing potential issues in your (hopefully calibrated?) monitor as well – so always look at the real print.
Here’s a detail – no obvious banding – the centre should be a little more defined though.
Measuring the print
Here’s the 21 step target, cropped out (click to see full size).
Opening a levels adjustment layer shows me a histogram.
There are a couple of issues here.
Firstly, the black and white points need to be adjusted, so that black is at 100% and white at 0%. We’re still in RGB mode at the moment, so you can see a range from 0 to 255 (black to white).
It’s easy to move the sliders to set the range, but the peaks are just not very clear.
Adding a bit of blur to the scan evens out the colour of the patches. The ink dots will merge.
Next, I cut out the parts of the image that are not solid colour.
I now have a target with just relatively solid patches of grey.
This makes it much easier to set the black and white points using the levels adjustment
You can easily see a slight nonlinearity.
I then convert the image to greyscale and measure the values for patches.
You don’t need to do this, but I’ve written down pairs of actual vs. measured values.
What I didn’t notice at first is that the mid grey which should be 50 is at 56, suggesting that the whole image needs a slight overall lightening. There is also the more obvious crunching of blacks.
Fixing with a curve
A simple curve adjustment layer can address this to some extent.
Now I say ‘some extent’ since even working at 16 bit I don’t want extreme curves. The curves in the original article represent an unadjusted print output I simply wouldn’t find acceptable today.
Once you’re happy with the results, you can save the curve and quickly apply it to photos before printing.
That’s if like me you use software with layers, such as Photoshop or Affinity Photo. With other software you may be able to give slight lightening to the image (correcting the midpoint) but curves are probably out.
If you find you need a big curve, then there is something else that needs fixing. it could be your media setting or your printer driver settings (You did remember to print the test image as a normal photo?)
However it could be that the paper you are trying just isn’t much good in the particular printer you are using. This happens, no matter how good someone on a forum said the paper was…
The curve when applied to the image, lightens it. Adding a levels adjustment, just to see the histogram shows a much more even arrangement of the peaks.
After applying the curve, I print the test image again, on the other half of the A3+ sheet.
You may need to look at full size to see the changes, although don’t worry if they are not clear. I didn’t set up lighting specially, and it’s just on the top of my desk.
it’s more obvious is this animation showing the original image and the lightened version that was printed.
Remember, the lighter version is -after- the curve has been applied and should look slightly too light, to counteract the darker than desired original print. Changes should be subtle. If not then it’s time to check things again.
What difference does this all make
The shadows are more open in the lightened version of the print, much more how they are meant to be in the test image. You can see more detail in dark parts of the print.
However, that’s looking at the print in bright light. Move into a dim room and you lose the shadow detail. If you know your B&W prints are going into a less well lit environment, it can help to boost overall brightness anyway.
I appreciate that not so many people will want to or be able to work with curves and layers for their images. If so, then look on the scanning technique as a way of finding whether your combination of paper, printer and driver settings is actually any good before wasting ink on prints that may well disappoint.
The process here does rather assume a linear response from the scanner/driver.
A big assumption, meaning that unless your scanner is calibrated this is at best a qualitative guide to the linearity of your B&W print. It can work well, but don’t get too caught up in the precise numbers unless you know your scanner can be relied upon.
There is another example of using a correction profile in an article I wrote about testing and using an unknown fine-art paper
More B&W print info
Other specific articles that may be of interest:
- Making a black and white photograph – from location to print
- A simple black and white print workflow
- Better photography from printing your work
- 3×3 quick tips for improving your black and white photography
Feel free to comment or drop Keith an email if you’d like to know any more…
Never miss a new article or review - Sign up for our occasional (ad-free) Newsletter
Enjoyed this article?
All the latest articles/reviews and photo news items appear on Keith's Photo blog
We've a whole section of the site devoted to Digital Black and White photography and printing. It covers all of Keith's specialist articles and reviews.
For All about using tilt and shift - articles/reviews about tilt/shift lenses
Articles below by Keith (Google's picks for matching this page)