26 March, 2014

Custom ICC Profiles

Picture of (not the real) PDI Printer Test Image

When you're printing on higher-end printers, the printer driver needs information about how a particular paper will be treated by that printer's ink set. The idea is to make the image on the paper appear, as nearly as possible, as it did on the monitor. Each printing paper manufacturer provides downloadable profiles for each of its papers, matched to each of the popular printers. Often, these "generic" profiles are referred to as "canned," because they're designed to address that elusive "typical" machine.

If you would prefer a profile for your actual printer and paper, then you need a custom ICC profile. (ICC stands for the International Color Consortium—the folks who agreed upon a vendor-neutral, cross-operating system, colour management system. Thanks, Guys.)

I usually work with three papers: Canson Baryta Photographique, Canson Platine Fibre Rag, and Epson Hot Press Natural. I've been happy with the output that I've been getting from all three papers using Canson's and Epson's own profiles. But, as I have mentioned before, "You don't know what you don't know." So, I decided that I would do custom profiles for each of these papers (for my Epson 3880 printer).

Assuming you don't have your own printer profiling equipment, you need to go to people who do. I went to ImageScience in Melbourne, Australia. They were quick, and their instructions were clear. (They had some out-of-date advice about the absence of soft proofing in Lightroom, but no one's perfect.)

There are few steps to get to a custom profile
  1. Calibrate your monitor (Strictly speaking, profiling the paper and printer combination doesn't require a monitor calibration. But, as a part of the whole printing process, it is fundamentally important to know that what you're seeing on your monitor will serve as a predictable rendition of what will come out of your printer with a particular paper. At the end of this process, when you print out a test image, the monitor brightness, contrast and colour renditions should be as close to that final image a s possible.
  2. Test your printer's nozzles. If you're working with a hobbled printer, you will not be able to rely on any of the results that follow.
  3. Print out the profile target using the paper that your profiling. The people with the profiling equipment will provide you with that target. This requires printing with all the adjustment tools switched off. The idea is to see exactly how the printer prints to a particular paper with no other colour management. Adobe has a neat little utility to do this. (Adobe needed to provide this because recent versions of Lightroom and Photoshop are missing a, "no color management" option.)
  4. Once you've printed the target, you send it to the people creating the profile, and they will email you back your new profiles.
  5. Then it's time to see how the new profile performs. To do this, it's best to work with a reference image. (The image above is a reference image reduced in size for this post. You can download the original from the ImageScience page HERE.
So, after all this, were there noticeable and significant improvements?

For the Canson papers, the differences between the canned and the custom profiles were minor, but my custom profiles are better. Some improvement was apparent in the skin tones. I watch the skin tones carefully anyway. So while it's good to start from the best base, the skin tone differences would probably get picked up in post. And, the red robot was just slightly brighter.

If you're not sure about going with custom profiles, print out the reference image mentioned above using your canned version. If you're happy with how your profile renders it, then you're probably okay.

I also have a new custom profile for Epson's Hot Press Natural. But testing it against the old profile requires switching to matte black ink. So, I'm not going to test it until I have some other HPN work to do. The Epson 3880 uses a lot of ink in the switch-over between photo black and matte black. I'll provide another post with the matte results when the opportunity arises.

22 March, 2014

Primes vs Zooms

West End, Brisbane

We're not talking photojournalism, sports or wildlife photography or "happy snaps." This is about enthusiast photography. Photography for its own sake. And the question is: Are Primes better than Zooms?

Primes are usually brighter, by 2 stops or more.They're simpler, so they're more robust. They generally have better resolution, better contrast, less distortion, less chromatic aberration. Primes are smaller and more innocuous.

Zooms often have a greater number of lens groups, and that can make them more susceptible to lens flare. And, because the lens hoods for zooms were designed to accommodate the widest focal length, the hood is only optimal for that widest setting. A properly designed hood for a prime, however, is always optimal. And, zooms often pull substantial amounts of air into the inside of the barrel of the lens when extending.

So, what DO zooms have going for them?

Look, I loaded the dice with that first paragraph. That was purposefully deceptive and pretentious nonsense. Those kinds of photographs are, often enough, exactly the shots that any photographer will want to get. And, if your only way to get those shots is to replace one lens with another, then you're probably going to miss some of them.

So, what do zooms have going for them? Hell, they zoom.

Brightness: most modern sensors resist noise well enough to give you the ISO headroom that a zoom needs. Resolution: most of the current digital cameras have resolution to burn if you're photographing for the web or for printing at 8x10 or smaller. Distortion and chromatic aberrations: most cameras correct for these in-camera; and if the camera doesn't, then you can kick them both to death in post-processing.

Wana test the robustness of lenses? Change them every hour or so. One good drop onto a hard floor.... And every time you change the lens is an opportunity for getting dust on the sensor.

The choice is really about what each photographer is comfortable with.

I think that some of the proclivities for primes are attributable to the rangefinder days (still here for many). Rangefinder mechanisms work best with primes.

I'm a mirrorless guy, but I still like my primes. I feel most comfortable with a wide, a standard, and a moderate telephoto. So, there is no way I'm going to get too many of those wildlife photos. But, I understand completely the photographer who feels encumbered by three lenses, when one will do.

So who wins? The photographers who get what they want.

P.S. I do have a couple of zooms.

18 March, 2014

DNG, Pt 2

I'm just going to dive in and assume that you've read Pt 1. I painted a rosy picture for DNGs last time. So it seemed sensible to make sure that reality doesn't have to take the back seat in this discussion.

Conversion safety

Digital information is a wonderful thing. But the reality is that every digital conversion carries some risk of file corruption. So, if you don't have to convert, why should you? I regard the risk of this to be extremely small, and the validation routines for DNGs seem to balance that concern for me.

Conversion time

If you've got quite a few images to import (more than a few hundred), converting to DNGs takes a bit of time. I was watching a video by Adobe's Julieanne Kost. She mentioned that while she eventually converts her images to DNG, she doesn't convert on import. She likes to do an edit first. Only when she's done a cull does she convert the keepers to DNGs. That way she's not converting files she not going to use anyway. If I was a more prolific photographer, I might see that argument as compelling. As it is, I hardly feel the pinch.

Kost's has mentioned another Lightroom tactic that only works if you don't rename your files on import. Lightroom, of course, records that image's name into metadata on import. If you subsequently change the name, however, Lightroom remembers that original name and keeps it stored in your metadata as well. Knowing that original name makes finding the backup RAW a breeze. I rename on import, so the tactic doesn't work for me. But my backed up RAWs are stored by date, so I can find an original image without much trouble.(Renaming files has a bad name with some Lightroom users. It might be a worthwhile topic for another day.)

One way trip

You can't go back again once you've converted to DNG. If you want to work with the original file in the camera makers software, you'll need to keep the original files. Frankly, I don't keep any camera company's software on my computer. It's a benefit of DNG and Lightroom that I don't feel I have to.

Medium format

As I recall, I read this on the Luminous Landscape: Because the metadata that's otherwise in the a"sidecar" file is incorporated into the DNG file itself; if you have a backup program, it will read any change to the metadata as a change in the DNG—and will have to backup the full file. And for medium format shooters, those are big files. If you stick with the original format, however, any changes only affect the sidecar file, that's tiny in comparison. And the backup software only needs to record that small change.

That seems to be a compelling argument for using the proprietary format if you're a medium format shooter. The only thing that might tip the balance back toward DNG is the "tiling" of data within the latest incarnations of the DNG format. That "tiling" allows multi-core processor to work with those large files more quickly.

15 March, 2014

The Good Ol' Days?

Bushland Park, Brisbane, Australia

For me, the iconic camera store in Chicago will always be Altmans on North Wabash. I could go there and oggle the cameras and equipment that I couldn't afford. And, they usually had the work of a local photographer in their foyer. In those days, the late '60s and '70's, it was still Marshall Field and Company ("Fields") across the street. 

Further South on Wabash was Kroch's and Brentano's Book Store — THE book store. And just a few blocks away, was/is the start of Route 66, at Michigan and Adams. Yes, I know some say it's Jackson Street, but if you're heading west, Adams is a one way street—westbound.

Adams Street looking west, with Wabash and the Elevated ("El") up ahead (a Google streetview)

Wabash, what a street.

It's sad, then, to hear of the loss of another bricks and mortar company—the closing of Calumet Photo. It's just the most recent of a long line of lost businesses.

03 March, 2014

DNG, Pt 1

Even if you're a Lightroom or Adobe Camera Raw (ACR) user, you don't need to use DNG. It's entirely optional. So, I'm going to tell you why I use it—after I tell you what it is.

Explaining DNG seemed pretty straightforward before I started writing. It became apparent, however, that considerable context is required. So, I'll do this in two parts. In this first part I'll explain what RAW and DNG are (as opposed to JPEGs), and then explain the benefits that led me to use DNG. In the second part I'll get into some of the implications and problems.


When you set your camera to record an image as a JPEG, the camera evaluates the incoming image and adjusts it to make it look like what the camera's designers thought that the photographer expected. More goes on than I'll describe here, but I'll try to hit the high points: In making a JPEG, the camera corrects for chromatic aberration and distortion; it sharpens the image, adjusts the exposure and contrast, guesses at the white balance and may do some noise reduction. It squeezes the image into a colour space (usually sRGB), reduces the image into an 8-bit format (256 levels in each of the three colour channels of red, green and blue). And, finally, it compresses the image to make it even smaller by deleting information in a manner designed not to affect the apparent quality of the image any more than necessary. (Because information is lost in that process, it's called "lossy compression.") Cameras usually do a damn good job, and you can use the resulting JPEG immediately, and just about anywhere.

RAW images, on the other hand, skip that in-camera processing. The unadjusted image is simply written to your memory card, usually in the camera company's proprietary format. (It's not as simple as this, however, because many companies can't resist some RAW fiddling, but for the sake of this discussion we won't go into that.) With RAW files the image elements mentioned in the JPEG creation process will be decided in post processing using image manipulation software. Enter, (ta da) Lightroom and ACR. 

When you import your RAW files into Lightroom or ACR you're faced with the option of either keeping the files in the proprietary RAW format of your camera, or converting them into DNGs. 

DNG is an image format created by Adobe as a standard to deal with the alphabet soup of proprietary formats. And, in addition, it's designed to add capabilities that the other formats lack—we'll come to those. The idea is to have a format that can be used as a digital negative. 

Many users don't convert, but I do—here are my reasons:
  • I want to "futureproof" my images.
  • Once an image is a DNG, there's no need for external, associated ("sidecar") files, everything remains in the one DNG "container."
  • There's a small saving in file space.
  • I'm confident that I'm not losing any quality in the translation.
  • I can find, view, and use DNGs in other programs.
  • Adobe has a plug-in for correcting colour casts that only works with DNG files
Future Proofing

DNGs will never be as popular or as ubiquitous as JPEGs, but they're already pretty popular; if only because DNG doesn't have any real competition. And DNG is an open standard. DNGs are used, for example, as an archival format in a project funded by the Library of Congress. And, as Lightroom and ACR become more popular, so does DNG.

One file

In Lightroom, if you choose to import your original RAW file in its proprietary format, any Lightroom actions saved to metadata are held within an Extensible Metadata Platform (XMP) "sidecar" file. That separate file must remain in the same folder as the original file and keep its associated file name. With DNG, all saved metadata is held within the one file. You can even tell Lightroom to embed the original file into the DNG file for possible extraction later. (That does increase the files sizes substantially.)

Smaller files

There a small, but real, size advantage in using DNGs—with no loss of quality. You don't see that every day.

Confidence in its quality

DNG is built upon the tagged image file format for electronic photography (TIFF/EP). TIFF, like DNG is an open standard (i.e. not proprietary). The default DNG is a "lossless" format, so it contains all the usable data from the original file.

Accessible through other programs

In the event that a DNG file is disassociated from Lightroom, it's still accessible. I can see my DNG images using Windows Explorer, for example; or through other programs such as Picasa. And, if I double-click on a DNG, it opens directly into Photoshop.

Using Adobe's "DNG Flat Field" plug in

Colour casts can be a problem at the widest end of the Sony 10-18mm zoom, for example, and other super-wide lenses on the the NEX-7. These casts can also arise as unwanted effects in tilt/shift work with technical cameras or with specialised tilt/shift lenses. Without the plug-in from Adobe Labs, such colour casts are extremely difficult to fix. There's also a separate program called "Cornerfix" that works like the "DNG Flat Field" plug in. It too, however, requires the image to be in DNG format.

DNGs work for me, but I'm not blind to the difficulties. I just haven't mentioned them here. For that you'll have to wait for part 2.