Photoblog 2.0

[Note: this is a reposting of a story originally published in August 2005.]

Fresco Flower, photo by Harold Davis.

I thought this photo looked like a painting when I took it, and I did my best to accentuate the effect in post-production processing.

First I applied Photoshop’s Fresco filter, then the Film Grain filter. After I cleaned the image up, I cloned some of the detail in the heart of the flower back in from the orignal image (proving, which I didn’t know before I tried, that you can use the Clone Tool between images, not just intra-image).

Check out some more of my recent flower pix (all post-processed in Photoshop, different manipulations and level of work involved):

When you use Photoshop, nothing is more important than the ability to make selections. As a simple example, when you want to make a composite by moving a person from one background onto another, you need a way to select the person—in other words, to tell Photoshop which pixels you want to move. Besides this kind of large-scale selection, selection is used on a much smaller scale in Photoshop to control the areas you want to change in an image, and for a variety of other reasons.

Since you need to be able to select in Photoshop, Photoshop provides a number of nifty selection tools. These include a Magic Wand Tool, and the ability to select on the basis of color range. But it is a sad fact of life for Photoshop addicts like me that we spend much of our life on time consuming and tedious hand selections. The task is made even more difficult by the issue of edge treatment. The edges of a selection should usually be blended using a gradient effect, or your selection will appear jagged and unnatural.

It’s not intuitively obvious, but making a selection in Photoshop is logically equivalent to creating a layer mask. Both selections and layer masks are stored as grayscale information. If you have a selection, you can convert it to a layer mask (I’ll explain how in the course of this review); going the other way, it’s easy to convert a layer mask to a selection (by applying the layer mask). For the real technophiles among my readers (I was almost going to say “alpha geeks” but restrained myself), a layer mask is the same thing as an alpha channel. Hmmm! Selection = Layer Mask = Alpha Channel.

My preference is to work with layers and layer masks because this is a more flexible option than working with a selection cut out on a single layer. But to get the layer mask I need, I often have to start with a complex and time consuming selection.

With this background in mind, I tried out Fluid Mask, a Photoshop plugin (it also runs in standalone mode) from Vertus that is supposed to ease the pain of selecting. Read on for an overview of how the software works, and for my evaluation.

View this image larger.

The flower you see above on black, a Gaillardia, was originally on a busy green background (because I photographed it in my cramped side garden). The original image following RAW processing is shown below with the background I wanted to replace.

As a first try with Fluid Mask, I decided to try selecting the flower to cut it out and put it on a black background. But before I show you how this worked, I need to make a disclosure. Disclosure: Vertus provided me with the Fluid Mask software for free, as well as some complimentary individual training (but see the note below about the availablity of free online training for anyone).

Once Fluid Mask has been installed, you’ll find it as a menu item on the Photoshop Filters menu (at the bottom of the menu items, along with any other third-party plugins you may have installed). With an image open in Photoshop, when you fire up Fluid Mask, Photoshop temporarily closes (”to save resources”) and the Fluid Mask window shown below opens. Note: As a matter of general Photoshop practice I’d advise creating a duplicate layer before invoking Fluid Mask.

Fluid Mask drew blue lines on this image, indicating the edges that its automated analysis found. (You may need to look closely at the screen capture to see these lines.) The basic Fluid Mask idea is to draw on your image using three different colored brushes. Green means that you are keeping the area, red means that you are throwing the area away, and blue is used for an edge that needs to be treated specially. Here’s how this might look on a corner of the Gaillardia image:

So, this was my first solo flight with Fluid Mask, and I didn’t get everything quite right. While the Gaillardia is really a pretty easy selection problem, to do a good job using the normal Photoshop tools would have taken me an hour or two. In Fluid Mask, it was about five minutes. But I did cut too much out of the flower interior by mistake, as you can see by comparing this cut out with the original or final versions:

When I applied and saved my work, Photoshop opened back up, and I was able to quickly fix my botch. I converted my selection to an alpha channel (layer mask) by clicking in the thumbnail in the Layers palette with the Apple key held down (Option key in Windows), and then clicking the little alpha channel icon on the bottom of the pallette (third from left, you can see this in the capture below). Then I added a black layer between my original flower and its masked duplicate. A reveal-all layer mask on the black layer allowed me to “paint-in” the few small areas from my original that I had missed in Fluid Mask. All this took about two minutes.

Bottom Line
I recommend this product. I expect to use it every time I have a reasonably complex masking or selection challenge, and I’m sure that I’ll become a more skilled user over time. For me, owning this product is a no-brainer. (o:

Pros: Greatly speeds selecting and leads to better quality selection.

Cons: Complicated to learn, without an adequate manual. The good news: online interactive tutorials are available free five days a week.

Product details: Downloads from Vertus cost $239 (in the US) for either Windows or the Mac.

Like World without End, my image of endless doors, I shot the base photo for this image at Fort Point in San Francisco. In the shadow of the Golden Gate Bridge, Fort Point is one of the oldest examples of military architecture in California. The brick vaults shown here were likely built when California was part of Mexico, and were contested during the war of independence that made California a republic.

View this image larger.

The Photoshop technique I used here is to create larger and small versions of the original, and layer them together to make the arches appear to go on forever (you probably need to view the image larger to see this). To enhance the illusion I created, I selectively sharpened the arches in the distance more than the closer arches. Ideally, this image should be viewed as a very large print to get the full effect.

Is there really a vanishing point?

[Photoshop composite at four magnifications. Original image created using five exposures at durations from 1/2 a second to 15 seconds, each capture using a Nikon D300 with a 12-24mm Zoom lens at 14mm (21mm in 35mm terms), at f/22 and ISO 100, tripod mounted.]

Nicky on Black, photo by Harold Davis.

Nicky and I were fooling around taking his picture. I gave him the flower-on-black treatment: I covered a high-backed chair with black cloth, and put him on the chair. Then I radically underexposed to enhance the depth of black in the background. I was careful to focus precisely on Nicky’s right eye.

In post-processing, I pulled Nicky out of the darkness layer by layer. Next, I selectively softened the image using the Smoothness setting on the Filter palette of the Noise Ninja Photoshop plug-in: both to compensate for the noise in this high-ISO image and to create a pleasing softness in the face and hair.

Finally, I shifted the image to LAB color mode, and carefully used the luminance channel and layer masking to modestly sharpen Nicky’s eyes, and only Nicky’s eyes.

[Nikon D300, 18-200mm VR zoom lens at 95mm (142.5mm in 35mm terms), 1/250 of a second at f/5.6 and ISO 2,500, hand held with vibration reduction turned on.]

In November of 2007, while I was clambering around Zion Canyon at night, exploring the Wave, and getting lost in the desert, Phyllis fielded a business call from Ringing Cedars Press. Ringing Cedars is the English language publisher of a series of books by Vladimir Megré.

The Ringing Cedars series conveys the wisdom, strength, and experience of Anastasia, a woman found naked upon the Siberian taiga. Anastasia provides insights on a wide range of topics ranging from health and utopian lifestyles to the measures needed to save our earth. Apparently, there’s a mysterious energy encoded in Anastasia’s words, and the more you read the books “the better you’ll feel.”

In any case, the Ringing Cedar series is a massive bestseller in Russian. While an English edition was in print at the time the Ringing Cedars publisher contacted us, the publisher was interested in creating a completely new and elegant design for the United States market. To this end, the publisher had hired noted book designer (and artist) Bill Greaves and conducted a massive artist search.

The goal was to find an artist with a body of work that could used as cover art. The cover art had to convey inspiration, and that it was both natural and imbued with a strong, creative life force. In addition, the art needed to be unique, distinctive, instantly recognizable, and cohesive. With these requirements in mind, the Ringing Cedars publisher was interested in my Digital Photogram series, which they had found on the Web. You can read more about some of the techniques I used to create this style of image in Xrays, Photograms, and Cross Processing, Oh My!

Here’s a product shot of the nine Ringing Cedars covers in a group.

The deal that I eventually negotiated with Ringing Cedars for the cover art was interesting because it was one part licensing, and one part assignment. Six of the images that wound up being used on the series cover were licensed, with minor modifications in some cases. On the other hand, I created three new cover images to fit the specific needs of the series titles. I always enjoy this kind of creative image creation, which usually leads me into some neat places in the process of fulfilling the needs of my client.

In the same way that the business arrangements were both fish and fowl (licensing and assignment), in a very real sense all the Ringing Cedars cover images involve both photography and digital painting. Each cover image is different in terms of where it falls on this spectrum. For example, the sunflower used for the cover of the first volume (”Anastasia”) is pretty much a digital photo, whereas the butterfly used on the cover of the fourth volume (”Co-creation”) is mostly digital painting from an original photo. That said, I think the team consisting of the publisher, the designer Bill Greaves, and Phyllis and myself, did a wonderful job of coming up with a cohesive look across a wide range of subjects. Generally, I’m appreciative of how well this team worked together. It’s rare in my professional experience to have a group of creative people working together with so much good will and positive energy.

Without further ado, here are the nine Ringing Cedars covers (along with some links to stories about how the images were created).

I blogged the image used on this cover here. We ultimately cut the flower off its stem to make it “float” on the black background. At the request of the client, I also worked in Photoshop to enhance the red glow in the center of the flower.

I blogged the image used on this cover here. More dragonfly images in this series.

I blogged the image used on this cover here.

I blogged the image used on this cover here.

I blogged the image used on this cover here.

I blogged the image used on this cover here.

I blogged the image used on this cover here.

I blogged the image used on this cover here.

The Lithographers Club of Chicago is a printing industry trade group. They selected an image of mine to use for the cover of their May/June 2008 magazine issue (the cover is shown below). Printing was contributed by Komori (they are a large manufacturer of printing presses) and demonstrates the Hexachrome process-color system. I’ll explain Hexachrome color later in this story.

View this cover larger.

Here’s my photo Wind, taken as a full horizontal, but cropped for this usage to fit the vertical cover format. You can read the story of how I came to make this image originally.

View this photograph larger.

Hexachrome color is a six-color process devised by Pantone in the 1990s. As you likely know, most process color printing is done with four process-color inks. The four process colors are Cyan, Yellow, Magenta, and Black (or CMYK for short).

The two colors that Hexachrome adds to CMYK are Orange and Green, so the six colors in Hexachrome process color are CMYKOG.

What’s the point of this? From the viewpoint of the digital photographer, the range of colors, called the gamut that CMYK can reproduce is limited. If you’ve ever had your photos reproduced via offset, you’ll know that some of them come out OK, but others lose a substantial part of their color range. CMYKOG (Hexachrome) has a much greater gamut than CMYK (see diagram below). In fact, the Hexachrome gamut falls somewhere between the sRGB and Adobe RBG 1998 gamuts, making photos reproduced via Hexachrome look very comparable to what you see on your RGB monitor (and your camera LCD). If you’ve ever flipped the color profile of an image in Photoshop from RGB to CMYK, and watched a huge color shift occur, you’ll know that this is potentially a big plus for digital photographers.

Commercial printers may not care so much about fidelity to an on-monitor RGB version of a photo. But they do have concerns about efficiency. For these printers, an important benefit of the Hexachrome process color system is that most of the solid colors included in the Pantone Matching System can be simulated using Hexachrome color. Simulating PMS colors means you don’t have to stop printing presses to prepare special plates for spot colors that have been specified using the PMS system (a common occurence in applications such as packaging). This translates to a potential boost in efficiency and cost savings.

So how does my photo do in CMYKOG? This is an image that has been reproduced a number of times in books. Each time, I’ve had to work on the CMYK files. But this time the reproduction shows the full range of colors in my photo and looks pretty much just like it does on my calibrated monitor—straight from the RGB file, with no special prep work on my part. In other words, I like it.

If you’re like me when it comes to wild and wacky Photoshop creations that start as digital photos, sometimes you don’t remember how you got to your end point. A related issue: sometimes in the process of experimenting one of the intermediate “experiments” seems better than the final version, so was it saved? How do you get back to it? Some of the good photocompositions get away, and this is all a matter of that apparently dull topic: workflow.

I’m not addressing the issue of Photoshop workflow in this story in detail, although I do have it pretty well down by now. Put simply: I archive all interesting variants of everything. This is a large topic, and means tracking a great many files and variants. I do plan to write more about this subject, which might generally be called post-processing, perhaps in that Photoshop book that I plan to get to some day.

A related issue is structuring and organizing storage for one’s digital images, and I do have a detailed blog story planned on the topic, so stay tuned.

The aesthetics, purpose, and (if you will) composition of photocomposition comprise another subject that I’m passing on for now. Sometimes it is best to approach imagery in a simple frame of mind: Does it do something for you, or not? This is related to the Supreme Court’s famous test for obscenity: Justice Stewart couldn’t define obscenity, but he knew it when he saw it.

No, this story merely shows the photographic pieces that I used to assemble the photocomposite I’ve called Event Horizon (immediately below).

View this image larger.

To head into a black hole of a tangent, Event Horizon is the title of rather silly 1997 science fiction film. According to the Wikipedia, in general relativity an event horizon is a boundary in spacetime, an area surrounding a black hole, beyond which events cannot affect an outside observer. Light emitted from inside the horizon can never reach the observer, and anything that passes through the horizon from the observer’s side is never seen again.

Moving away from image titles, the genesis of Event Horizon is the relatively straight architectural photo of the Life Sciences Annex on the Universoty of California at Berkeley campus shown below.

View this image larger.

I photographed this building stopped down for full depth of field using a circular polarizer to bump the reflections at 22mm (33mm in 35mm terms), 1 second and f/25 at ISO 100.

After a series of straight exposures, I started fooling around zooming during the exposure (see In a What-If State of Mind for more on this technique).

Waterfall Method (below) is the result of zooming from 70mm to 24mm (105mm to 36mm in 35mm terms) with the camera still on the tripod while doing a 2 second exposure at f/25 and ISO 100.

View this image larger.

So here’s my recipe: Take the original architectural shot, add a zoom time lapse slightly overexposed photo, and sprinkle in a pinch of Photoshop layering and channel operations, and Voila!

Gaillardia x grandiflora, photo by Harold Davis. View this image larger.

This flower is a Gaillardia x grandiflora ‘Oranges and Lemons’. Gaillardias are native to North America, and are sometimes called Blanket Flowers because of their coloration.

I’m using an eight foot long raised bed in my protected side yard to grow flowers for photography, and this Gaillardia is the first subject. As models go, I think my flowers will prove to be very pretty and cooperative. Another benefit: they don’t seek modeling fees.

I photographed this flower on a black velvet background using diffuse natural sunlight. A previous experiment had convinced me that a single point of focus wouldn’t create an image that was sharp all over the flower. So I made twelve varying exposures at three focus points, and hand layered them together for an HDR and HFR image.

Some related stores: Falling in Love, Red Flowering Dogwood Blossom, Gaillardia, Digital Photograms.

[Nikon D300, 200mm f/4 macro lens (300mm in 35mm terms), 12 captures at shutter speeds from 1/2 of a second to 8 seconds, all at f/32 and ISO 100, tripod mounted.]

Looking Down, photo by Harold Davis. View this image larger.

I took the original version of this photo (you can see it below) from the top of a rotunda in San Francisco City Hall. You can see a couple of people reading notices by a courtroom door, and the faint shadow of people in motion walking the corridors, rendered wraith-like by the long three-second exposure (selected so I could get plenty of depth of field).

The circular opening in the photo was actually pretty narrow, and the railing was high. The problem for me was getting my tripod in position over this balustrade. Even so, some tripod shadow and a tripod leg ended up in the capture, and I had to Photoshop them out.

With a decent rotunda view in hand, I pasted in four successively smaller (each copy was 20% of the size of the previous version) copies of the orginal image, to create a composite with the illusion of endless depth. This is the same technique I used in Endless Stairs and World without End.

As Phyllis says, “Down, down, into the pits of Hell, each a circle of bureaucracy lower in the pit!”

View this image larger.

[Nikon D300, 10.5mm digital fisheye, 3 seconds at f/22 and ISO 100, tripod mounted.]

Door Knob Dome Scandal, photo by Harold Davis. View this image larger.

Everyone thought it a scandal when the door knob in my basement got together with the dome in San Francisco. But the dome and door knob were merely romantic, and invited a red rose, too.

Related image: Dream Stairs.

Nautilus in Black and White, photo by Harold Davis. View this image larger.

Finley Eversole contacted me for permission to use one of my chambered nautilus images in his book Art, Death and Transformation, to be published by Inner Traditions in 2009. Based on my conversation with Finley, he’s a man with personal experience of the abstract expressionists. I gather that his book is about the relationalship of the spiritual to art, particularly in the context of mid-twentieth century art. There’s an extensive discussion of Robert Smithson’s famous Spiral Jetty, constructed in 1970 and located in the Great Salt Lake, Utah.

In the context of Spiral Jetty, Finley wanted to illustrate spirals in nature, hence my chambered nautilus image. The catch was that he could only reproduce my image in black and white. Here’s the original image in color (you may notice that the black and white version is rotated).

I undertook the conversion to black and white myself because I was afraid that anything like a straight RGB to Grayscale switch would result in a murky image that lacked contrast. Converting RGB to Grayscale simply throws away the color information in the image, much as will happen when you discard the AB channels in LAB mode. Getting a little more sophisticated than this, I converted using a black and white adjustment layer. On the default settings, the results were predicably flat. Using the so-called Ansel Adams settings (explained here) and a channel mixer adjustment layer, I got slightly better results. By the way, my new book Light & Exposure for Digital Photographers from O’Reilly includes a case study that visually shows the impact of various black and white conversion techniques.

So I realized I needed to do a great deal of preparation, and essentially change the contrast structure of the image, before discarding the color information. I put the image through a variety of steps, but the most effective was the selective deployment in tritone mode (also explained in Light & Exposure and see Toned), using black, white, and Pantone metallic silver as the three colors.

By the way, I am planning to make prints of both the black & white and color versions of this image, so please let me know if you have an interest in owning a print.

I’ll leave the last words to Finley Eversole:

The spiral represents the unfolding of our hidden creative powers and symbolizes both self-realization and boundless expansion. Its secret is beautifully dramatized by the chambered nautilus whose in-dwelling life unfolds cyclically and periodically, mirroring in itself the expanding spirals and eternal patterns of the evolving cosmos.

Related stories: Nautilus 69, Spirals, Resistance to Spirals is Futile.

Vertigo, photo by Harold Davis. View this image larger.

I lay down at the bottom of the spiral staircase shown in After the Wedding. With my tripod fitting clumsily in a tight corner of the stair, I used my digital fisheye to take a vertiginous photo up the stair well. In Photoshop, I layered in an extension to the hallway and replaced the skylight at the top of the stairs. Finally, I duplicated the image and flipped it horizontally. I pasted the original and the flipped version together to create a symmetrical, but twisted, abstraction.

[Original photo: Nikon D300, 10.5mm digital fisheye, 5 seconds at f/22 and ISO 100, tripod mounted.]

Memory Palace, photo by Harold Davis. View this image larger.

This digital photo collage is an elaboration of Dream Palace, itself a manipulation of a photo of William Randolph Hearst’s over-the-top underground swimming pool.

A memory palace, also called a method of loci, was a mnemonic system used for complex memorization in the days before external memory devices were common.

If you couldn’t write something complicated down, how were you going to remember all the details? A memory palace mentally pairs rooms in the imaginary palace with segments of the material to remember. Often the best mental memory palaces are based on an actual physical place.

If you want to try to construct a memory palace to help remember something complicated, here are some instructions for going about it.

The image below is a digital collage, meaning a Photoshop composite created by blending layers. As these collages go, this is a fairly simple piece. All the elements in the collage come from a single dSLR capture using my Nikon D300 (see below). So I’m able in this story to show where each element in the final digital collage comes from, how I generated the elements from the one capture, and how I put them together.

View this image larger.

The original of all three elements of this collage is an ISO 6400 photo of the underground pool at Hearst Castle shown below (here’s more info about the photo):

The final image is a triptych of three versions of this original photo. The middle image is the original photo with a duplicate version flipped vertically and layered on top, like so:

The left image is an inversion of the LAB A channel of the original image, like this:

However, I didn’t like lights putting out an apparently blue light, so I painted the yellow lights from the original back over:

The right image is an inversion of the LAB B channel of the original image:

Once again, I painted over the lights (a black lamp isn’t any better than a blue lamp):

To combine the three “panels” of my triptych, I started a new image sized to be three times as wide as the original, and then I added each element as a separate layer.

With this kind of effort, one of the most difficult things is to know when to stop. I’ll be writing about knowing when to stop in another story.

Some related stories: When Is a Photo Not a Photo; Variations, Resistance to Spirals Is Futile; World without End.

I’ve been shooting in low light at ISO 6400, partly on the grounds that my Nikon D300 is remarkably low noise, and partly on the grounds that noise can be used as part of the aeshetic of an image. My photos of the jellyfish in the aquarium tanks at Monterey are examples of this kind of high ISO work. However, I do find I need to selectively post-process for noise with these images.

View this image larger.

I processed these jellyfish photos, all shot at extremely high ISOs, for noise using Noise Ninja as a plug-in to Photoshop CS3. (Noise Ninja can run as a standalone or inside the Photoshop environment.)

With one twist (I’ll get to my variation in a moment), I used Noise Ninja in its default mode. This means opening Noise Ninja, profiling the image by clicking a button, tweaking the filter settings for strength, and then applying the noise reduction.

My own deviation from the tried-and-true starts by working on a duplicate layer, rather than the original. This is a best practice for Photoshop in any case. Then I use a layer mask to hide the Noise Ninja noise-reduced layer, and selectively paint in portions of this layer. Typically, I’ll work with two noise-reduction layers at different strengths, because even a very noisy image isn’t necessarily noisy all over. I also want the freedom to apply Noise Ninja selectively, and at different strengths, to different parts of my photos. I’ll leave some areas untouched.

View this image larger.

It’s worth noting that I use a similar selectivity-via-layers-and-masking approach when it comes to sharpening. Furthermore, I only sharpen luminance (black and white) and not the chroma (color) channels of a photo. My main sharpening tool is the paradoxically named Unsharp Mask Photoshop filter. Leaving chroma channels unsharpened happens to have a beneficial effect on the aesthetics of noise, so this kind of selective sharpening is really a help when you start with a noisy image.

View this image larger.