Clay Harmon Blog

Extended gamut profiles in Gum printing

Tue, 30 Dec 2025 00:00:00 +0000

Why do CMYK+ conversions?

I have found that I get my best color gum prints when I do five or more color separations. This involves using extended gamut color conversions of my 3-color RGB digital files. Partly this is due to the type of color images that I like to make - darker earth-tones are a favorite of mine, and getting balanced color with just a CMYK conversion is challenging given the inherent variability in a hand coated process like gum bichromate. Small changes in coating thickness of a given layer can have a pronounced effect on the color balance of the finished image. With just three primary colors to use, I can often end up chasing my tail with many adjustment layers to get the colors the way I want them to appear.

Getting extended gamut profiles

Calvin Grier of “The Wet Print” is a color wizard in the alternative process community, and he recommends using an additional layer of iron oxide to help balance colors more easily. This approach is particularly useful for images that contain skin tones which are notoriously difficult to get right in gum printing.

Calvin has generously offered free downloads for some of the custom CMYK and extended gamut color profiles he has created for his Printmaker’s Friend non-toxic gum substitute. These color profiles can also be used for traditional gum bichromate. They can be accessed here and then selecting the ‘Imagesetter Negatives’ tab and scrolling to the bottom. Pay careful attention to the target LAB values he has for each profile. These values indicate the target maximum density on each color separation. Your combination of negative, pigment concentration, and exposure should yield these values if you want to get the best match with his profiles.

Modifying channel names in extended gamut profiles

One issue with his CMYK+Iron Oxide profile is that the process Calvin goes through to create the profile necessitates that the Iron Oxide channel is used as a Yellow pigment and the regular Yellow is put into an ‘Extended gamut Yellow’ channel. These names are embedded in the profile and may confuse you when you begin your color separation process. To reduce confusion with the channel names, I have created a little Photoshop script utility that allows you to manually rename both the channels labels (called colorants in the script) and the profile description before you copy it to the correct destination folder.

This script can be accessed on here. To use this script, unzip the download file, and copy the ICC-editor.jsxbin file into the Applications/Adobe Photoshop 20XX/Presets/Scripts folder. Restart Photoshop if it is running, and the script will now be accessible by clicking the File->Scripts selection.

It is best to modify the channel names and description before copying it to the final location for color profiles. The desktop is a good spot to drop the file to be modified.

Start the script by clicking File->Scripts->ICC-editor with your mouse and the following window will present itself:

Choose the metadata in the profile you wish to modify. If Edit Description is chosen, then this window will appear:

Edit the description as you wish. The description will be the new filename for the modified ICC profile, so concision and brevity for this text field is advisable. Click ‘OK’ to finish editing this metadata field.

Next you may want to modify the channel (colorant) names. In the case of the Iron Oxide .icc profile, the Iron Oxide channel is actually labeled as ‘Yellow’ in the colorant table, and the actual Yellow channel is labeled as ‘Extended Gamut Yellow’. Click ‘Edit Colorant names’ and a dialog will open with all of the channels that the profile creates:

The current names are in grayed-out text in this dialog. To modify a channel (colorant) name, click on the colorant to be modified and type in a new name. Change any others you need to change. The edited colors will be in full contrast. Click OK to save these changes to the profile metadata.

You will now be back at the main dialog, and the edited metadata fields will have an asterisk after then on the button indicating they have been modified:

Clicking ‘Save’ in this dialog will suggest saving the modified profile to a folder named ‘output’ at the same location as the original ICC profile is located. You can use the file picker dialog to choose another location if you want. You will be presented with a final file save confirmation dialog after ‘Save’ is clicked:

Copy the modified profile to a folder where Photoshop can find them. On macOS, this location is Library/Application Support/Adobe/Color/Profiles.

Open the image file that you are working with in Photoshop, and then select Edit->Convert to Profile.... Because we are using an extended gamut profile, the ‘Advanced’ button must be clicked on the upper right-hand side of the dialog, and then an option for ‘Multichannel’ will appear. Click the selection box arrow and select the profile you want to use for the conversion.

For an extended gamut conversion such as RGB to CMYK+Iron Oxide (Cyan, Magenta, Yellow, Black, Iron Oxide), select the Multichannel option and then select the target output .icc profile:

One thing to be aware of is that the Multichannel options create separate channels that are essentially spot colors for each of the selected colors. Photoshop’s color engine cannot accurately convert these spot channels to the computer display’s color profile, so the results on screen will look weird. Here is an example of an input RGB color file that is to be converted to a multi-channel output:

Once the file has been converted to the 5-color CMYK+IO profile, the image looks like this:

This looks weird, but if the separations are done using the extended gamut profile, and then printed, the results should look like the original image.

This may seem to be a lot of trouble for a color gum print. The need for extended gamut profile is certainly image dependent. But if you have images that contain earth tones and skin tones, this extra step is certainly worth it to be able to get the final print looking like you want it to.

Mangling Color: separation neg best practices

Fri, 14 Nov 2025 00:00:00 +0000

Why do CMYK and CMYK+ conversions?

I have found that I get my best color gum prints when I do four or more color separations. This involves using CMYK or extended gamut color conversions of my 3-color RGB digital files. At first glance, this would seem straightforward: just use the ACE (Adobe Color Engine) to convert a file to a target .icc or .icm color format. The devil is in the details, of course, and when converting noisy, underexposed, and heavily edited low-light images, it is easy to run headlong into significant pattern and banding noise on some of the output color channels when using a conversion workflow that does not take into account the non-linearity inherent in any conversion that is not in the native file’s 3-color space.

Evil lurks in the shadows

I had an image that I had severely underexposed and required a 3-stop exposure increase in Lightroom, and when I exported it using AdobeRGB as the output color space, I noticed significant banding on several of the channels. After investigation, I discovered the reason for this is that AdobeRGB->CMYK conversions involve some compression of values in the dark areas, and quantization errors get magnified when converting from a 3-color space to a 4-color space such as CMYK. Maintaining the perceptually linear ProPhoto RGB color profile through the entire editing pipeline prior to conversion to CMYK solved the problem.

Do this

If your attention span has been hopelessly Instagrammed, here is the short version: Always import your RAW camera files into Lightroom using a perceptually linear color space conversion. Your best option is to use ProPhoto RGB (the default colorspace in Adobe Lightroom), or perhaps ACEScg. When exporting to Photoshop for the color separation step, always export the file in this same perceptually linear color space at 16 bits! This will greatly reduce, although probably not completely eliminate, the pattern noise in the darkest areas of the image. The improvement will be very noticeable in the K (black) channel which seems to accrete all the sensor’s pattern noise into a steaming pile of…

One very important lesson to learn from this is that you should always be shooting in a RAW format, which gives you flexibility in this file conversion pipeline. Most phone cameras now have a RAW format, and you should be using this for any low-light image that you eventually may want to print in color gum.

Another interesting discovery I have made that may ultimately be an Adobe bug: I get slightly different results when I export a file from Lightroom than one that I export from Adobe Bridge when using the same AdobeRGB output!

I have gone down a rabbit hole on this whole color conversion process over the last two days and my biggest takeaway is that from now on I am changing my default editing space in Photoshop to ProPhoto RGB. This colorspace is the least problematic and has given me the cleanest separations.

Any time that your image file is converted from one color space to another there is potential for digital mischief to occur. The problems tend to snowball if the image processing sequence stretches or squeezes the information in the image file in any way.

Recommended workflow for CMYK and CMYK+ separations

Convert raw files into a perceptually linear color space for editing. ProPhoto RGB is the easily available recommendation for this. Maintain ProPhoto RGB through the entire process until ready for conversion to your destination color space. Change color settings in Photoshop to default to ProPhoto RGB for all editing. Never export a file as 8 bits!
Do all noise reduction and editing in ProPhoto RGB. When ready to convert to CMYK or CMYK+, Select Image->Convert to Profile, and select the output profile you want to create for your separations. The image will probably change slightly. Use curve adjustments on each channel to get the image colors exactly the way you want them to appear.
Run the CMYK Separation Negs script on the converted file.

CMYK and extended gamut ICC profiles

Calvin Grier is one of the most technically accomplished color alt-process printers around, and he has generously offered free downloads for some of the custom CMYK and extended gamut color profiles he has created. They can be accessed here and then selecting the ‘Imagesetter Negatives’ tab and scrolling to the bottom. Pay careful attention to the target LAB values he has for each profile. These values indicate the target maximum density on each color separation. Your combination of negative, pigment concentration, and exposure should yield these values if you want to get the best match with his profiles.

To use any of these profiles, download, unzip, and then copy the files to the folder where Photoshop can find them. On macOS, this is Library/ColorSync/Profiles. Open the image file that you are working with in Photoshop, and then select Edit->Convert to Profile...

For a straight 3-color to 4-color conversion (RGB->CMYK), you should select the radio button for CMYK conversion on the left side of the dialog, and then select the CMYK .icc profile you want to use for the conversion:

For an extended gamut conversion such as RGB to CMYKOGV (Cyan, Magenta, Yellow, Black, Orange, Green, Violet), select the Multichannel option and then select the target output .icc profile:

One thing to be aware of is that the Multichannel options create separate channels that are essentially spot colors for each of the selected colors. Chances are these extended gamut options will muck around with the appearance of your color image. The reason this occurs is that the Photoshop color engine cannot accurately convert multichannel colors back into the RGB display space of the computer monitor. The results will look weird on the screen, but should print accurately once the separations are made and printed in the color process you are using.

Duotone Photoshop action

Sun, 27 Jul 2025 00:00:00 +0000

Duotone proofing station action

Usage

This simple action will create what Calvin Grier calls a ‘duotone proofing station’ from a Photoshop document. The action creates layer groups from the original image, one that will be the highlight color, and a second that will be the shadow color.

First, make sure the image mode is set to RGB, even if it starts out as a complete grayscale image. Then run the action from the Action pallet. The Action creates 3 major groups.

At the top of the stack is a copy of the original image. Visibility of this layer can be flipped on and off with the ‘eye’ activation button next to the layer. This is used to compare the tonal balance of the colorized image with the original.

The next two layers down the stack are the highlight and shadow layer groups.

In each of these layer groups is a layer at the very top of the stack that is used to determine the color for that particular part of the image. It has no color in it after the action is run, but to add a highlight color, use the color picker to create a color for this layer. Keep in mind that you will not have to reduce the color intensity for the highlight layer - this is handled automatically in the partitioning process.

Use the paint bucket tool to pour color into both the highlight and shadow color layer. Make sure the layer is active so that the color is applied to only that part of the image.

Flip the visibility of the ‘original’ layer off to preview what the duotone color split will look like.

The addition of the color will probably change the overall tonal appearance of the image slightly. You can adjust the points in the curve adjustment layer for each layer group to modify how the duotone will look.

Output

If you are using this action for creating two different plates for polymer photogravure, you will want to create two output files - one for the highlights and one for the shadows.

To create the image to use for printing the highlights, turn off visibility for the ‘Shadows’ group and the original. Also turn off visibility of the color layer in the highlights layer. Flatten the image and save this as ‘highlights’ or some other memorable name you can concoct. Use the history pallete after saving this and undo the flattening.

Next export the shadow image file as a separate image file in a similar way by turning off visibility of the highlights, the shadow color, and the original layer. As before, flatten the image and save it with a memorable name.

The same process can be used if you want to output the two layers as negatives, but you should invert them before saving if you are not using the negative printing switch in QTR Print Tool.

Installation

Download the necessary files here.

To install this action, open the Actions pallet and click the hamburger menu on the upper right, and select ‘Load Actions’ from the menu and then navigate to the location where you have unzipped this file and select the photogravure-actions.atn file and load the action set. Note: this action set also includes a few other simple actions for creating image files that can be printed as negatives to make relief plates. These just save a few menu clicks if you want to convert an image to a bitmap.

Next, you need to copy the duotone-highlights.acv and the duotone-shadows.acv files to the /Applications/Adobe Photoshop 2025/Presets/Curves folder. Finally, copy the duotone-highlights.alv file to your user library in the ~/Library/Application Support/Adobe/Adobe Photoshop 2025/Presets/Levels folder. You will probably have to supply an administrator password to do some of these copy operations. These instructions work for Macintosh computers. I don’t use Photoshop on Windows, but hopefully if you are advanced enough to want to create duotones you can figure out the comparable file locations on a Windows box.

Once all these are copied, you can run this action from the Actions pallet.

A Happy path for setting up the P900 for DTP

Wed, 26 Feb 2025 00:00:00 +0000

As I write this post, the Epson P900 is the least expensive option for a printer that will work with DTP techniques without issues with the printer putting roller marks on the plate. The P700 has the same print head, and in theory should work. However, the thick media front-loading mechanism is different than the P900 and I have never gotten a satisfactory result.

The P900 is the devil

I have assisted several people in getting their P900 printer set up, and there are a load of land mines that must be avoided if you hope to use the printer to its full potential.

#1 Don’t use the Apple printer driver

It seems like you should be able to just open System Settings and add the printer through the Apple dialogs and prompts. Unfortunately, this installs the printer as an AirPrint printer, and one of the critical Epson programs for handling media will simply not work.

#2 Use the Epson driver downloaded from the Epson site for your region

Download the installation .dmg file from Epson. Double click to install the printer. Make sure that your printer is on, and that you have already used the hardware panel on the front of the printer to connect it to your local network. I have had iffy results using the direct-USB connection. In one case, we were able create a USB-connected printer and make the Media Installer work. In another, we failed after multiple attempts. The WiFi option worked every time.

#3 Get a beverage, because it can be slow to install

There were points in the Epson driver installation process that I thought it had failed, but just waiting a few more minutes allowed the installation to finish successfully.

#4 Test if the Epson Media Installer works

The test for whether your installation went smoothly is to navigate to /Applications/Epson Software and double click to open the Epson Media Installer app. There will most likely be dire warnings from Apple. Ignore them and start the app. You should see a screen like this:

Click the import button and download the DTP media file from here, and try to import this media settings file. If that works, download the profile to the printer, and go to step #5, otherwise, go to #4A!

#4A If the media file fails to install…

Move to plan B. Click the Media Download tab in the Medial Installer app, and download a profile from Epson for posterboard:

Update August 2025

Thanks to the persistence of Robert Hartung in dealing with ‘extended wait times’ at Epson USA, it is apparently the case that a Posterboard media type is no longer available from their download site. Epson recommend choosing Ultra Smooth Fine Art Paper as the media type, and then downloading that .emx file, and editing as recommended right below this paragraph.

This should install properly, and now click the original Media Management tab. Highlight the poster board media type and click ‘Copy Media’ to create a copy of this. Edit the settings and change the name of this profile, and then download it to your printer. These are the settings you need:

#5 Change some hardware settings on the P900 printer panel

Finally, to keep the printer from driving you crazy as it rejects your carrier, change a few things on the printer hardware panel.

Navigate to the house icon, and choose ‘Printer Settings’. First select ‘Error Notice’ and turn off both ‘Paper Mismatch’ and ‘Detect Paper Meandering’.

Next, select ‘Thick Paper’ and turn it to ‘Off’

#6 You just thought you were done

The final thing is to check if the printer installed as a Bonjour printer when you did the initial setup. You need the printer to be a Bonjour printer so QuadToneRIP can recognize it over the network. In System Settings, open the ‘Printers and Scanners’ page and see if the Epson you installed is listed as a Bonjour printer. If it is, you are done. But if it is not, you need to install a second copy of the Epson driver as a Bonjour printer (Stupid! I know). If the Epson driver shows up as ‘EPSON TCIPIP’, you need to click ‘Add Printer’ and choose Bonjour printer as your printer type and use the same driver that is already installed. You can see evidence of this madness on my printer dialog box showing two different Epson P800s! :

In the Weeds with DTP

Tue, 03 Sep 2024 00:00:00 +0000

The basics of the direct-to-plate (DTP) process are simple: print directly on a photopolymer plate with an inkjet printer using only high-UV-blocking matte black to compose the image. The microscopic matte black ink dots block UV light, while any spaces between the dots are hardened. When the plate is washed out (etched) in water, the unhardened polymer underneath each black dot washes away and leaves the hardened polymer behind. This results in a topography of little polymer needles throughout the plate that hold the ink for making intaglio prints. The final print reproduces imagery in a continuous range of tones.

But with a little more investigation, there are several factors that can determine the amount of maximum image black that can be produced in a print made from the etched plate. The characteristics of the printer head, the dithering routine used in the printer driver, and the physics of the exposing light source all have an effect on the final resulting plate.

Overview

With Jet plates no longer available after their manufacturing facility shut down, we had to locate another source for photopolymer plates that would work well for doing polymer photogravure. I used the Jet plates for years and had very good luck with them.

I was reluctant to go back to the Toyobo KM product line because my past experience with those plates was not good. The Toyobo product has an unfortunate issue with sporadic and unpredictable mottling that I could never find a good way to rectify. The Toyobo plates were also were very picky about processing: they had to be dried thoroughly and quickly after washing out to avoid waterdrop marks. They also have a very noticeable ‘chemical’ stink when you are washing them out. My failure rate with those plates was unacceptably high.

We finally landed on the Fuji Torelief WS73 HX4 (aka HY3 ⊕The Fuji Torelief plate material catalog is a confusing inscrutable mess with multiple names for the same material ) plates as being similar to the Jet material. The polymer compound (polyvinyl alcohol)in the Torelief plates is the same as the Jet product, which results in no odor, and no fussiness in processing and drying. The only knock on them is that exposures were slightly longer, and the plates seemed to have a little more plate tone ⊕Plate tone is a term printmakers use to describe the very thin amount of ink that is left on a plate in the very lightest parts of the image. Some printmakers wax lyrical over a subtle amount of plate tone, while others consider it the balm of Satan. than the Jet plates. The plate tone can be dealt with through ink selection and good wiping technique using a synthetic wiping fabric.

My prints made with Torelief plates looked the same as those made with Jet plates. I use an Epson P800 printer, and a densely packed 365nm UV-LED light box for exposing my DTP plates.

The Puzzle

Then, a friend brought in an Epson 3880 printer to calibrate for DTP. I used the same set of calibration steps I used for the P800, but I got some very strange results. My first test is always a simple one: I make a small plate using my calibration image that has 51 steps and print it using a linear ⊕a linear profile means that a 10% black in the image file directs the printer to lay down a 10% ink on the plate, and a 70% black in the image file directs the printer to lay down 70% ink on the plate, and so forth. matte black ink QuadToneRIP .quad file.

Typically this test should reveal a progression of gradually darker steps as more and more printer ink is laid down on the plate until a maximum black level is reached. Then, as more printer ink is added to the plate, so much light is blocked that it results in open bite ⊕Open bite means that while the plate has been etched deeply, there are not enough fine polymer needles sitting in the etched-out area to allow the ink to stay in place as the plate is wiped. The result is an unappealing gray instead of a true black as the polymer is completely etched out.

I made a print with this plate, and here is what I got:

WTF! That light area from 28%-62% was really weird. My exposure times went from 10 minutes to 30 minutes - plenty of time in my high output box.

I tried both less exposure and more exposure, changed the print resolution, even changed the dithering algorithm in QuadToneRIP, but I never could get a satisfactory black level. You can see in this test that starting at 28-30% matte black ink level the print densities start to decrease and go away completely until they start to creep back in at 64%. This just made no sense to me.

Of course, I could have picked 28% as my maximum matte black ink level for the plate, but that would result in a print with a Dmax of 1.1! Completely unacceptable!

A clue is found

My friend David Marsh had recently gotten a P900 printer to calibrate for DTP and I was sort of popping in and out of an email thread helping him get set up. He ran into a number of issues, but I remember one thing he tried that really helped him out was that instead of using 100% matte black ink, he found that a blend of matte black and photo black gave him the strongest blacks in his plates (or more accurately, the prints made from those plates).

This blending of matte and photo black inks is not possible with the P800 and 3880, but I know from my digital negative experience that the Yellow ink is almost as effective at blocking UV as photo black. So I created a series of linear ink blends with different proportions of matte black and yellow ink. I printed the same linear test with each blend, and then evaluated the results:

As you can see, any amount of yellow ink improved the situation, but 75/25 mK/Y seemed to be the best combination without even a hint of that light band in the middle. I used a densitometer to pick the deepest black in the series of steps, which was a 1.5 at the 70% total ink level.

I used that total printer ink value to go through the linearization routine that I typically use. I won’t cover it here, but I will add another post in the future outlining that process. I got a perfect profile.

What is going on?

My theory on why this is happening is that the slightly less effective UV blocking yellow ink is allowing just enough additional light to penetrate around the dot-free areas, and enables the little aquatint needles to anchor to the steel substrate.

What I think is going on is that the middle lighter area in that original test is where the aquatint needles are present, but don’t have a robust enough base (or shoulder) next to the steel and they are getting knocked off during washout. This could be a similar to the ‘blow-down’ phenomena in carbon printing with imagesetter negatives.

The darker areas beyond that are where very little UV light is penetrating, but enough is penetrating that the surface contains some rugged ‘proto-needles’ that are able to hold the ink.

Or I could be wrong about all of that. All I know is that the yellow ink trick really works.

But wait, there’s more

I shared this information with my friend Matt Magruder, and he mentioned that he never has seen this phenomena with these plates, but that he exposes all of them with a metal halide Nuarc unit at Flatbed Press in Austin.

Crap. Another rabbit trail to start down. I ran a quick test comparing my UV-LED lights against my Amergraph metal halide exposure unit. And here it is:

The Amergraph’s light is more collimated than the UV-LED light. The emission angle on the LED diodes is fairly narrow, but not as narrow as the light from a metal halide exposure unit. As you can see, the Amergraph gets some really beefy black print values with low values of printer ink. Carrying this logic even further, a fluorescent tube exposure unit should be the worst light to use for DTP purposes. I’ll leave that investigation to a curious reader.

Conclusion

Don’t get caught up in the idea that DTP works with only matte black ink. A helper ink that is not quite as UV-opaque can be a useful addition when constructing a profile
The intensity of the light is only part of the equation when it comes to exposure for DTP photogravure. A box with double the space between strips will probably not be just half as fast, but more likely a lot less fast than that because the angle of emission in the overlapping illumination areas will be higher than a box where all the strips are spaced as closely as possible.

Resources

Color Muse 2: new version review

Tue, 16 Jan 2024 00:00:00 +0000

The Color Muse 2 is a new version of an inexpensive handheld colorimeter that allows the photographer making digital negatives and positives the ability to measure LAB values from step tests to be used in programs such as QuadToneProfiler QuickCurveDN.

Overview

The original Color Muse colorimeter proved to be a fantastic inexpensive way to measure LAB values for use in applications that use the information to make digital negatives, digital positives, and polymer intaglio plates.

LAB is a three-factor system for measuring the color of a reflective surface. This system measures Luminance, and an ‘A’(green-magenta axis) and ‘B’(blue-yellow axis) value to uniquely specify the color of a reflective surface. This link is to a Wikipedia article on CIELAB for the terminally curious.

For digital negative creation, we are primarily interested in the L* or Luminance value, which indicates the reflectance of a surface on a scale from 0 to 100, where 0 is a perfectly absorbing black surface that reflects no light and 100 is a non-specular surface the reflects all the light that strikes it.

Unfortunately, I lost my original Color Muse during a workshop in another state, and I needed to replace it. When I went back to the Color Muse site, I noticed that they have a new version called the Color Muse 2 that is slightly more expensive, but according to their specifications should be more accurate than the original Color Muse. The cost is $30 more, but its accuracy is 0.2$\nabla$E2000 versus 0.5$\nabla$E2000 for the original. This accuracy is the same as their much more expensive spectrophotometers.

For more information about using the Color Muse, please look at my earlier article on the original Color Muse I wrote last year. This article will only review the differences in the device and the new software interface.

Review

Size

This device is physically larger than the original Color Muse. It has a wider diameter, which can make it a little more challenging to perfectly hit the target square when taking measurements.

Charging

The Color Muse 2 charges the same way with a mini-USB charging cable connected to a USB-A male connector. I suspect this now-anachronistic USB-A port is meant to limit the power delivery to the device because the newer USB-C standard can deliver far more power than the device can probably handle.

Smartphone app

Color Muse has updated its iOS app and the process for taking a measurement is a little more involved than on the previous app.

Download the app for Color Muse from the appropriate App Store for your device. Follow the instructions to pair the device via Bluetooth after opening the app. Then click the ‘hamburger’ menu in the upper left corner of the app and select ‘Settings’ and set the Color Format to ‘CIE Lab’.

Use the device calibration procedure to calibrate the Color Muse once it has connected to your phone. The shutter on the bottom of device is closed for this operation. Once the calibration is complete, open the shutter and center the device over the patch you want to read. Press the ‘Scan Now’ button to make the reading.

You will see the color of the square at the upper left part of the app screen change to the color of the patch that was just read. To see the LAB values of that patch, you must then tap the patch. This is an extra step from the older version of the app and makes this a little clunky to use.

These values must now be entered into your linearization program.

Accuracy

The big improvement that is gained for the additional $30 is the much higher read accuracy of this device.

The original Color Muse was precise, but not accurate. The repeatability and linearity of the readings was very good, but the actual values were consistently about 6 L units too high. This did not matter too much because the linearity is the main thing we are interested in for digital negatives. The indicated dMax values it would give were always too low however:

To assess the accuracy of the Color Muse 2, I measured an actual Q-13 target with both the Color Muse 2 and the Xrite i1 Pro2 spectrophotometer. This chart compares the $100 Color Muse 2 with the $1700 Xrite i1 Pro 2 Basic version(the i1 Pro3 basic version is now $1800).

These are outstanding results, and there is virtually no difference between the readings for the two devices. If one were to average a set of 3-5 readings, I would expect the small differences seen in this chart to disappear. This new version of the devices is now both precise and accurate!

Conclusion

The Color Muse 2 is a big improvement in terms of accuracy over the original Color Muse. The largest selling point of this device is, of course, its very low cost. At $100, it is quite a bargain. That low cost does come at the expense of some convenience. It is a little clunky to use because of the two steps required to make a reading, and all the measurements must be recorded manually in either a spreadsheet or the linearization app being used.

That said, this is a fantastic device for the alt-process practitioner who only occasionally needs to calibrate a process. It would be entirely inappropriate for someone trying to do precise color-managed printing workflows and making ICC profile. The idea of making 500-1000 measurements with this device is crazy and the automated patch scanning available in the Xrite i1 Profiler software is the way to go for that purpose.

Group order for UV-LED enclosures

Tue, 26 Jul 2022 00:00:00 +0000

The history

I have built four UV-LED boxes over the last three years. The last three have been designed to use aluminum enclosures fabricated by a machine shop. The benefit of the aluminum is that the box itself serves as a heat sink for the LEDs and this is assisted by using fans to cool the lights during exposure. As a one-off endeavour, these boxes can be pricey since the setup cost is almost as much as the actual fabrication.

The plan

I have had several inquiries about doing a group order for a standard-sized box to reduce the cost since the setup cost is spread among a larger quantity of units. I have received and assembled a prototype box that has external dimensions of 20” x 30” x 6”, with cut-outs for a Borg relay timer, inlet module for the plug and either a 350W or 500W Meanwell transformer. The design will allow the illumination grid of LEDs to be 18” x 24”, or approximately Euro A2.

Since that time, I have assembled a so-called ‘4X’ box with the LED strips crammed as close as possible (3mm spacing). This amounts to 45 strips in this box, and it is super fast. It exposes palladium prints in 70 seconds and salt prints in 15-20 minutes! This unit requires a 500W transformer and is a little more difficult to build because of the dense spacing of the LED strips.

All of these boxes can be built without any soldering necessary, although I did do some soldered wire splices to save space in all my boxes. But soldering is not required and the three-way splices can be accomplished with 3-pole terminal blocks instead. The strips can be attached to the transformer via snap on LED connectors and 12-pole terminal blocks.

We have had a very positive reception for these boxes, and are offering them as a product at Mountain Intaglio. We will accumulate orders and ship them in batches three times a year. The ordering deadlines are March 31, July 15 and Sept 30. To get the cost of fabrication down, we need to order multiple units at one time, and this is why we try to combine as many orders into one fabrication run as possible.

You will have to supply the UV-LEDs in whatever wavelength you choose from any supplier you choose. Right now, the two main contenders are the Inkjet Mall Type 2 UV-LEDs and the Waveform Lighting RealUV LEDs. Three 5-meter reels of the Inkjet Mall LEDs on an 11mm spacing or four 5-meter reels of the Waveform Lighting LEDs on a 5mm spacing will produce almost identical UV illumination on a 18” x 24” grid.

You can find the boxes at Mountain Intaglio.

This is a sturdy unit that weighs in at under 20 pounds and will be of interest to anyone who wants a portable and sturdy UV exposure unit and has a day or two of time to assemble the unit.

A video thing:

Color Muse: an inexpensive colorimeter to use for digital negatives

Mon, 20 Jun 2022 00:00:00 +0000

Color Muse is an inexpensive handheld colorimeter that allows the photographer making digital negatives and positives the ability to measure LAB values from step tests to be used in programs such as QuadToneProfiler QuickCurveDN.

Overview

This tiny little colorimeter is a great tool for photographers looking for an inexpensive device that will generate LAB values. It has a downloadable app that is installed on your smartphone, and the phone pairs with device via Bluetooth.

The device is charged using a USB cable with a micro-USB connector. One charge seems to last quite a while, but it will discharge over time and will probably need to be recharged if it has been sitting on your shelf for a few months.

Advantages

Inexpensive relative to a professional colorimeter
Portable and easy to carry
Charges with any USB source with a micro-USB connector

Disadvantages

Not intended for photography - LAB values must be read from the phone app and then transcribed to a text file which is then processed using QTR-Data-Tool. This can seem tedious, but in practice is only done occasionally when calibrating a new process. I have created a little web tool that you can use for data entry and it copies the formatted data to the clipboard for subsequent pasting into a plain text file.
Measuring the exact target can be difficult. The targets need larger squares than the downloadable target image files typically provide. It can be difficult to know if you are measuring the exact square for a series of targets. Below, I will have a download link to a 40-step target with squares that are easy to measure with the Color Muse.

Using the Color Muse for digital negative and positive linearization

The device itself is very small. It is a tube that is about 1-1/2 inches tall and 3/4 inch in diameter. It has a built in calibration plaque in the cover for the sensor. Here is what it looks like:

Download the Color Muse app from whatever App Store your smartphone requires. After the Color Muse has been charged, open the app on your smartphone and pair the Color Muse with your smartphone using the pairing function on the app.

Recording the data from Color Muse for linearization

Prepare to record the values from Color Muse using TextEdit

The following instructions assume you are using a Macintosh computer since that is the only system that QuadToneProfiler QuickCurveDN runs on at this time.

Open up the TextEdit app. By default, TextEdit is set to write files using the .rtf file format. This is a rich-text file format that contains information that can format text with different fonts and other visual attributes. We do not want to create a file in this format, so the first order of business is to set the Preferences in TextEdit so that the default format is plain text without any additional formatting codes embedded in the file.

Click the TextEdit->Preferences… menu item with your mouse. A Preferences dialog box will open. This box has two tabs. In the first tab titled ‘New Document’, change the Format selection button to Plain text instead of ‘Rich text’.:

Now click on the ‘Open and Save’ tab in this dialog box, and in the When Saving a File section, click the button that is titled ‘Add “.txt” extension to plain text files’:

Print a test target that the Color Muse can easily read

Many of the default test targets used with devices like the X-Rite i1 Pro spectrophotometer have tiny little patches that are difficult to read with the Color Muse. To get around this problem, I have created two step targets with larger patches that print in a roughly 4x6 inch (~Euro A6) area. The 40-step target looks like this:

Both of these untagged .tif files can be downloaded in a zip file by clicking this link.

Do what you do normally do with this target. Go through your workflow to create a negative (or positive) and print it using your desired print process. Dry the resulting print thoroughly, and prepare to make the measurements with one of the following methods:

Record the measurements - Option 1 using my web form

I have created a javascript web-based page that will allow you to enter the data into a data grid on the web page, and then after it has all been entered, allows you to click a button and copy all the data in the proper format to the clipboard on your mac. Just open a new blank plain text document in TextEdit, and paste it in and save it. This little app allows you to record 21 or 40 step target information.

You can access the webpage here.

Record the measurements - Option 2 using Data-Tool

Measure each patch and record the values in a plain text file

Using TextEdit, create a new blank text file. Give it an informative name like ‘First-linearization-attempt.txt’.

Open the Color Muse app on your smartphone and pair with your Color Muse device following the instructions provided by Color Muse. You will probably need to click the little hamburger menu in the upper left corner of the app and specify that you want the device to read out your CIE LAB* color.

Calibrate the device using the tile that is built into the screw-off cover of the device. Remove the cover and start making measurements of your target.

Read the first patch (Row 1, column A) and read the LAB values from the smartphone app. Type in these values on the first line of the text file, with spaces separating the values. Once the first reading is recorded, hit return to move the cursor in the TextEdit app to the next line, and then make your next measurement (Row 1, column B). Record those LAB values on line two and press return. Proceed in this way until you have measured and recorded all of the patches in the target.

After all the measurements have been made and recorded into the TextEdit file, save it to a place you can easily find on your computer.

Use Data-Tool to smooth the data save the file

If you have not done so already, download and install Data-Tool from the QuadToneRIP website: QTR-Data-Tool. Open up the Data-Tool app, and click on File->Open… and select the text file you just created and saved in a location you will never forget.

The Data-Tool app will open the file and graph the LAB values. It will look like this:

The only decision you need to make now is the degree of smoothing that the Data-Tool application will apply to your measurements. The red-circled set of buttons in the image above is where you select your smoothing choice. Note that the Data-Tool app will resample your data when it is smoothed and either extrapolate or interpolate the data to a fixed number of samples. If you want to use a larger number of samples, Option 1 is a better choice. Click all of them and see the effect each level has on your data. Choose the one that makes you the happiest, and then click on File->Export Data… to open the export dialog box:

Save the data as a text (.txt) file by choosing the File Format as Text Data File. Modify the file name with something like a ‘-DT’ at the end of the name and save it. The reason you want to save it as a different file is in case you want to go back later and choose a different smoothing level on the original file.

The file you have just created can now be read directly into applications like QuadToneProfiler QuickCurveDN to be used in your linearization steps. It can also be used in other applications that accept a LAB file as a measurement file.

Conclusion

The Color Muse is an inexpensive tool to measure and record LAB measurements to use in digital negative creation workflows. It is much less costly than buying a new X-Rite i1 Pro spectrophotometer.

Step-test measurement options

Sun, 19 Jun 2022 00:00:00 +0000

Every digital negative or digital positive approach needs some method for evaluating how much correction needs to be applied to a digital input file in order to create a final print where the tones in the prints match the tones on the monitor.

Measuring step tests

There are several approaches that can be used for creating a measurement file that programs such as QuadToneProfiler QuickCurveDN can then use to create a linearized output for making digital negatives for traditional and alternative photographic processes.

Traditional reflection densitometers

Anyone who began photography in the age before digital cameras probably remembers a piece of equipment called a densitometer. There are two types that are typically seen: reflection and transmission densitometers. Some densitometers make both types of measurements, but the one of interest to those making digital negatives or positives is the reflection densitometer, which measures the amount of light reflected from a print. Darker black areas of a print reflect little light, whereas lighter areas reflect lots of light. These instruments make measurements that in log_D. This measurement value can be input into a file that some of the digital negative programs can use directly. There are densitometers still being manufactured such as the Heiland TRD 2 that measure both reflective and transmissive log_D values.

The main downside to using a reflection densitometer is that it requires the conversion of log_D to L values. Many programs such as QuadToneProfiler QuickCurveDN do this math for you, but if you are using other approaches, it may necessitate plugging the values into a spreadsheet that converts these values for you.

Scanning a step-test on a flatbed scanner with a Q-13 target

Another approach that is relatively inexpensive for making measurements is scanning a step test side-by-side with a Q-13 reflective target. Particuarly for those using QuadToneRIP who already own a flatbed scanner, this can represent a very inexpensive approach. Q-13 targets can be purchased online at BH Photo Video, Adorama, and Freestyle Photo.

This approach is particularly useful for people using QTR, and is sort of automated by using the QTR-StepWedgeTool, a Photoshop script that can be downloaded here. The process is slightly ‘touchy’, and requires you to outline the Q-13 target first, and then the 21-step test, and finally save the results to a file. The results are good, however, and the only downside is that it only works on a 21-step (5% gray increments) step test. Detailed instructions are included with the download, and there are some YouTube videos that also demonstrate how to use this tool.

Measuring a step-test using an X-Rite i1 Spectrophotometer

This approach is probably the most accurate, but this accuracy comes at some expense. New i1 Spectros cost around $1650 for a new base model at BH Photo Video and getting one fully tricked out can be close to $7000. One benefit of this approach is that the same instrument can be used to calibrate your monitor, which is necessary if repeatable results are needed. There are less expensive monitor-only spectrophotometers available however, and one of these should only be purchased if you intend to use it for additional print profiling and digital negative creation tasks.

Used i1 spectrophotometers can be found on eBay for prices between $150 and $500. When looking at used i1 spectros, make sure that you are buying an i1 Pro2 Rev E or newer model. The program needed to interface with this device is the i1Profiler (i1 Publish) software. Many features of this program must be licensed, but by clicking the ‘advanced’ button, a workflow named ‘Measure Chart’ becomes visible and may be used without needing any additional licensing.

Use an inexpensive handheld colorimeter such as the Color Muse

A final tool that can be used is a handheld colorimeter such as the Color Muse. This tiny little instrument is used by decorators and designers to make onsite measurements of colors. It is used by pairing it via Bluetooth to your smartphone, and using an app as the interface to your colorimeter. When a measurement is taken, the app displays the LAB values along with a matching paint color! Of course, we don’t need a paint color, but the LAB values are very useful for us photographers.

The main downside to this instrument is that it does not have a native software interface that will generate the necessary measurement file. This means that you will have to record each individual measurement in a text file and then run it through QTR-Data-Tool to generate your measurement file, or use the little javascript web form I have created here to enter the values which can then be copied and pasted into a text file that can be read by the BW Mastery family of apps.

A more detailed post about the Color Muse is available here.

Summary

There are four good approaches to making measurements to be used for digital negative and digital positive linearization, and three of them generate LAB values directly.

The main purpose of any technique is to establish linearity of luminance (the L in LAB). The absolute values from each technique are less important. I measured the same printed step test recently using all three LAB techniques, and below is a graph plotting the results they generated. The main thing you will notice is that the i1 has noticeably different LAB values, but that all three techniques generate measurements that show the same slope of values. That is the critical element, since the linearization techniques are only trying to establish an even progression of L values from darkest to lightest.

Mac Studio and SoftRAID 6 compatibility

Thu, 21 Apr 2022 00:00:00 +0000

I have been waiting for years for Apple to come out with a truly powerful mid-range computer for photographers and video editors. The new Mac Studio fits that description well. I purchased on to replace my aging iMac and overall I am very impressed with the computer. However…

After I used Migration Assistant to migrate all of my files and applications from my iMac to my new Mac Studio, the new computer rebooted and immediately blue-screened and crashed. I tried booting again with the same result. I wondered if there was a problem with the external drive and RAID software, so I unplugged the RAID enclosures from the Thunderbolt port on my Mac Studio and booted again. The computer booted up without any problems. I reinstalled the SoftRAID software, and rebooted again. After the boot, I plugged in my Thunderbolt drive enclosures and they fired up and SoftRAID recognized them without any issues. I was able to access all my files.

In two weeks of use, however, I noticed that I would occasionally return to my computer and find that it had crashed and rebooted. I examined the crash logs and it seemed to be an problem with the MacOS system playing well with the SoftRAID driver. Restarting it successfully required the same routine of unplugging the RAID enclosures and booting and then plugging them into the Mac Studio only after a successful restart.

I contacted SoftRAID support, and they responded with the following:

There is a M1 bug that affects a small but significant number (5% perhaps) of users. The system will kernel panic when connecting the volume, or shortly thereafter. It’s some kind of threading issue, we are waiting for Apple to fix. But we discovered two main workarounds if you suffer from this: Disconnect one disk when connecting the volume, then push it back in after 5 minutes or so. Or create a new RAID 5 volume with 64K stripe unit size.

And indeed, their FAQ page has this problem problem at the top of the list.

Fixing the issue

I keep my large multi-terabyte photo files on a set of external OWC Thunderbay 4-disk boxes. I use SoftRAID to configure these enclosures in a RAID-5 configuration, which allows the loss of one disk in the array without any data loss. I keep two of these enclosures running. One is my primary working disk for all of my photo library, and the second is a cloned version of the same library of files, but I only clone it every week, reasoning that if I make some gigantic screwup on my main volume, I have a week to resolve the problem before any problems get cloned to my other enclosure.

My plan was to first back up my primary working data volume onto a single high capacity disk as an insurance policy. I do this routinely and even keep one of these disks in a safe deposit box as insurance against complete destruction of my house! I then opened the SoftRAID application, and selected the volume I plan to recreate using a 64K stripe unit size. I right clicked and removed the safeguard and then right clicked again on that volume and deleted it:

Deleting the volume will destroy all the data on the RAID volume! Test your backup to make sure it is good before doing this step

The next step is to create a new volume. In the left panel, select the drives you just released from the previous RAID-5 volume. Note that there will be a notation on each drive that it is “not used for any volumes”.

‘Creating a new SoftRAID volume’)

Click the Create button, and a dialog box will pop up:

The default stripe size is 16K. This selector button should be clicked and changed to the 64K strip size. Click OK and the new RAID volume is created.

The final step is to copy all the data from the backup disk onto the newly created RAID-5 volume with the 64K stripe size.

This should fix the issue.

Clay Harmon Blog

Extended gamut profiles in Gum printing

Why do CMYK+ conversions?

Getting extended gamut profiles

Modifying channel names in extended gamut profiles

Mangling Color: separation neg best practices

Why do CMYK and CMYK+ conversions?

Evil lurks in the shadows

Do this

Recommended workflow for CMYK and CMYK+ separations

CMYK and extended gamut ICC profiles

Duotone Photoshop action

Duotone proofing station action

Usage

Output

Installation

A Happy path for setting up the P900 for DTP

The P900 is the devil

#1 Don’t use the Apple printer driver

#2 Use the Epson driver downloaded from the Epson site for your region

#3 Get a beverage, because it can be slow to install

#4 Test if the Epson Media Installer works

#4A If the media file fails to install…

#5 Change some hardware settings on the P900 printer panel

#6 You just thought you were done

In the Weeds with DTP

Overview

The Puzzle

A clue is found

What is going on?

But wait, there’s more

Conclusion

Resources

Test target download page

mK and Y blended .quad profiles for P800 and 3880

mK and pK blended .quad profiles for P900 printers

Color Muse 2: new version review

Overview

Review

Size

Charging

Smartphone app

Accuracy

Conclusion

Group order for UV-LED enclosures

The history

The plan

Color Muse: an inexpensive colorimeter to use for digital negatives

Overview

Advantages

Disadvantages

Using the Color Muse for digital negative and positive linearization

Recording the data from Color Muse for linearization

Prepare to record the values from Color Muse using TextEdit

Print a test target that the Color Muse can easily read

Record the measurements - Option 1 using my web form

Record the measurements - Option 2 using Data-Tool

Measure each patch and record the values in a plain text file

Use Data-Tool to smooth the data save the file

Conclusion

Step-test measurement options

Measuring step tests

Traditional reflection densitometers

Scanning a step-test on a flatbed scanner with a Q-13 target

Measuring a step-test using an X-Rite i1 Spectrophotometer

Use an inexpensive handheld colorimeter such as the Color Muse

Summary

Mac Studio and SoftRAID 6 compatibility

Fixing the issue