The Piezography Heretic: To Convert Or Not To Convert?

This is an article about workflow – the sequence of steps used in making a Piezography print – with particular emphasis on the colour management aspects.  In the first section I describe what I term the “orthodox” workflow, which may assist new Piezography users understand the process, as key aspects of the standard QuadToneRIP (QTR) workflow are not well-explained in the New Piezography Manual (28 August 2013 edition).  However the rest of this article is intended for experienced users as it deals with alternative “heretical” workflows.  I’ve tried to make it as simple and straight-forward as possible, so if you’re a new user and feeling brave, then good luck to you with the rest of this article.   Just keep in mind that some of the adjectives and colourful phrases in this article are used tongue-in-cheek.

The Orthodoxy

I assume you have either created or been given an ICC profile corresponding to the Piezography QTR curve that you’re printing with.  The image you wish to print is supposed to be in Grey Gamma 2.2, so convert it from whichever colour space it’s in into Grey Gamma 2.2, edit it with the aid of a soft-proof to achieve the image you want, and print.  Do not convert the image to the ICC profile in order to print.  In Windows you use QuadToneRIP GUI (QTRGui) to print.  In Mac OS X you print using a companion program to QTR, Print Tool, as you can no longer print direct to QTR from Photoshop, because of changes to colour management since Mac OS X 10.6.8.

Under this workflow, soft-proofing is a little different.  We’re not converting to the ICC profile in order to print.  So when you Select “View | Proof Setup | Custom …” in Photoshop, you must check the option “Preserve Numbers”, which shows you what your print will look like if printed exactly as-is in Grey Gamma 2.2.  It will help to get a good soft-proof if your monitor can be set to a colour temperature of 5000K (D50) and have its brightness and contrast set to match the characteristics of the paper, a process termed calibration for print.  This is the orthodox approach.

The difference between a preserve-numbers soft-proof on and off for an image in Grey Gamma 2.2 will depend on the image, but for an image with deep shadows they will appear lighter in the soft-proof.  This is not true for all images, the difference for high-key images, for example, will be negligible to none.  The difference is much more marked for matte papers than gloss/baryta.  However in an image with substantial shadow tones, and especially deep shadow detail, you most definitely will notice the difference for matte papers.   To see the effect most clearly, it may help to temporarily uncheck “Simulate Black Ink” and “Simulate Paper Colour” in your soft-proof settings when doing this comparison.

New users who don’t realise the importance of editing using this soft-proofing approach often complain about light and/or flat and/or weak prints.  Which is understandable, as this orthodox workflow is not all that well explained in the New Piezography Manual, unfortunately.  Well, Grey Gamma 2.2 is mentioned quite a lot, but not the “Preserve Numbers” soft-proof setting. 

Perceptual Linearity

Now anyone who has printed colour using ICC colour profiles will be wondering why on earth don’t we just convert our images to the ICC for printing, just as we do for colour? 

[Note: For colour printing, this conversion is normally undertaken on-the-fly when selecting a printer/paper profile in the Photoshop print dialog, rather than being done manually.

Well, you can, but you risk being burnt at the stake as a heretic.  Let me explain.  The ICCs generated by the QTR ICC profile creation tools have been designed to provide “perceptual linearity”.  If you want to understand what this means, read this post by QTR creator Roy Harrington.  My short summary is that they result in a better screen-to-print match if you convert to them for printing, although there’s more to it than that.  In order to achieve this, a conversion to the ICC at print time will seriously compress shadow detail.  Will shadow detail get lost?  In an image that has any significant shadow detail, yes.  This is what will get you burned at the stake. 

Printing an image in Grey Gamma 2.2 as-is via QTR without profile conversion will result in more open shadows and more shadow detail – more than what you see on the screen (without  a soft-proof).  Converting to the ICC profile will have the opposite effect.  There was a fairly heated debate about this some years ago – see here and here if you’re interested, but you’ll need to read the entirety of these two threads.

To get some idea of the effect of converting to an ICC or not, here is a comparison of linearisation plots from the two approaches.  For those not familiar with this concept, to create a linearisation plot, print a standard 21×4 test chart using the QTR curve for the paper in question, which looks like this:

21x4 random

then measure the printed chart using an i1 Photo or similar, process the results through QTR-Linearize-Data, and plot the luminosity column from the output in a spreadsheet (there’s a little more detail on this process in the Re-linearising A Piezography Curve article).  The ideal result looks like this:

EEM without ICC conversion

In the above chart, the vertical axis measures luminosity and the the horizontal axis measures the density of the patch, with 0 being white and the 100 being black.  The pink line plots the ideal numbers and the black the measurement results.  In this case, the luminosity values are very close to the ideal straight line, which is why Piezography is described by its creator as “linear in Gamma 2.2”, which I interpret to really means “linear in luminosity” and printed in Grey Gamma 2.2.  It’s not just Piezography that has this property – any QTR curve that is linear will fit this description if the image is printed in Grey Gamma 2.2.

To demonstrate the effect of converting to an ICC, I assigned the untagged 21×4 TIFF file to Grey Gamma 2.2,  converted it to the ICC (something you’d otherwise never do to the 21×4), printed and measured.  Here is the linearisation plot, and you can see the effect on the shadows:

EEM after ICC conversion

It’s linear for much of the range, but not in the shadows, and overall it’s a fair bit darker.  The rationale is that this provides a better screen-to-print match.  The counter view is “but what about the shadow detail?”

The Heresy

So what’s a person to do?  Well I have two ways of dealing with this.  In images which don’t have a lot of shadows and shadow detail, I sometimes still convert the image from Grey Gamma 2.2 to the ICC profile when printing.  But whatever you do, don’t tell anyone.  I find that I get better local contrast in the shadows in such situations.  But when I do choose to convert, I carefully compare versions before and after the conversion to ensure that any important detail is not lost.  I also often apply a Photoshop curve to boost the shadows a little to offset the effect of the conversion.  But I only do this rarely nowadays.  Mostly I now print using one of the hybrid or pragmatic approaches below.

[A note of caution: – images that you’re printing for measurement purposes, such as the 21×4 set of grey patches, should never be converted to an ICC or greyspace.  If you do anything with them at all, they should be assigned to Grey Gamma 2.2.]

Being Pragmatic – Assigning

A more pragmatic approach is first to convert the image to Grey Gamma 2.2, assign the image to the ICC profile, edit the image to suit, and print in Grey Gamma 2.2.  Assigning to the ICC doesn’t change the numbers in the file, so you can defend yourself against charges of heresy by pointing out that the file really is still in Grey Gamma 2.2, despite what embedded profile appears to be.  I find it easier to edit and save images assigned like this, you just need to be able to track that you’ve done so, for later reference.

[Note: you’d need to assign the image to a greyscale profile.  If you want to assign the image to an RGB profile, then you’ll need to first convert the image to AdobeRGB, which for a monochrome image is just the RGB version of Grey Gamma 2.2.]

What assigning does is tell Photoshop to interpret and display the image on the monitor differently.  There’s a different interpretation for the numbers in the file.  By assigning to the ICC like this, in effect you’re getting most of a soft-proof.  If you also want the “Simulate Black Ink” and “Simulate Paper Colour” options, then you’ll have to enable soft-proofing.

An additional advantage is that if you want a JPG that matches (more or less) the print, then you can simply convert a copy of the image with the assigned profile to sRGB, for example for web display.  If the Piezography inkset that you’re using is other than neutral, e.g. Special Edition, then the sRGB image should be a good approximation of the toning in the print.

I have to confess that I don’t fully understand the implications of how this would work on a OS X, printing from Print Tool.  On Windows, QTR ignores any embedded profile.  You’d need to be certain that you had Print Tool configured to do the same.

Being Pragmatic – A Partial Conversion

There are some people who want to convert to an ICC, but haven’t got one because they haven’t got a measurement device to create one.  They want the better screen-to-print match and apparently don’t care about or are unaware of the problems that this causes in the shadows.  This technique of Paul Roark’s uses a Photoshop curve to approximate the ICC conversion process.  Doing this is the exact opposite of the orthodox approach and is pretty much guaranteed to get you burnt at the stake!

Actually, on a more serious note, Paul’s Photoshop curve is an excellent way of using a compromise workflow, somewhere between the two alternatives discussed above.  If you find the Grey Gamma 2.2 workflow print too light and the ICC conversion print too dark, you can put his Photoshop curve on a layer and have the opacity in the 40-60% range and use it to edit down the shadows up or down to where you want them by adjusting layer opacity, with the preserve-numbers soft-proof as your guide.  It allows me to dial in an intermediate rendering.  So it’s a useful tool no matter which workflow you use.

Last: Resizing and Resampling Next: Re-linearising A Piezography Curve


Leave a Reply

Your email address will not be published. Required fields are marked *