Myerson Photo Blog

Stone wall texture

I’ve been trying to come up with a cheap and somewhat realistic stone wall background. You may be looking to do this for your own photography backgrounds, or for theater productions, for Halloween decorations, or even for a very different kind of wall surface for a kids room. For me, this was for a photography background. I don’t think this is it yet, but it’s getting close. Here’s how I got there.

First off, the materials and equipment. The surface is a 4×8′ piece of EPS foam. This is the standard home insulation foam available at all home improvement big box stores. I went with the inch and a half thickness, because I knew I wanted to be able to cut some deep grooves between the rocks. Depending on where you live, there will be different kinds of EPS foam available. Here in Arizona, it’s simply the sheets of white Styrofoam. Other parts of the US will have big pink or blue sheets. Each is slightly different, and I believe the pink and blue are easier to work than the white, but you play the hand you’re dealt.

Next – and most important – is the foam cutting system. I went with the Hot Wire Foam Factory pro system (this one, specifically). It’s not the cheapest way to go, but I’ve used it for a number of projects already, and I wouldn’t change a thing. Ever hear the expression “like a hot knife through butter”? That’s what you get with this. The foam carves away beautifully. It takes some practice to get the cuts you want, but once you get the hang of it, it’s a great system.

Finally, I picked up some paint – 2 cans of black spray paint, a gallon of a medium gray, and a quart of a light gray. Your home improvement center can color match paint for you from those little paint chip samplers. That is definitely the way to go. If you end up building a second wall section, you’re going to appreciate the ability to go back to The Depot to get another can of exactly the same color.

So lets get carving. One important note about this step is to make sure you have adequate ventilation. The melting foam stinks and breathing it in can’t be all that good for you. The first thing I did was to layout the lines between the stones. I tried to keep it pretty regular, with just a few irregularities. Depending on the look you’re after, you may want to go very rigid, or very loose.  Once the mortar lines were drawn in, I grabbed the Hot Wire cutter and carved them out. I used a combination of the straight knife and the freehand router to cut the deep grooves.

Next was the irregular stone texture. This is the part that I thought would have the biggest impact on the final look. It matters, of course, but not as much as the paint job does. So the bottom line is not to stress too much about this step. I just tried to make the surface irregular. I used a variety of techniques, including dragging the router, using the side edge of the knife tool, and gouging with the router. The good thing about rough stone is that it’s a very forgiving look. You can use a variety of techniques and still get the effect.

With the blocks defined, and the texture cut in, it was off to the paint room. First I covered the whole thing with a pass of the black spray paint. Most important is to cover the deepest grooves. With that done, I grabbed a paint roller, and covered the whole thing with my medium gray. The paint roller doesn’t go into the deepest grooves, of course, so they stay black. It also doesn’t get into all of the shallower texture “valleys”, so I did some stiff brush touch up afterwards. \

Once that was dry, I opened the light gray and dry brushed that on. This is the step that really brings the piece to life. Dry brushing is when you scrape most of the paint off the brush and lightly pass over the surface of the workpiece. This has the effect of depositing paint only on the highest points of the texture. Starting from the top of the stones and brushing downward, I was able to create a very realistic texture effect very quickly.

The last step was one more pass of the black spray paint, pointing up from the bottom of the piece. This has the opposite effect of the highlight dry brushing; it creates the shadows under the “mountains”

(I didn’t take enough in-progress images when I made this piece, so I’ll make another one soon and take pictures along the way to illustrate this article.)

All in all, I like the result, but there are a few things I’ll do differently next time. I’m not thrilled with the brick pattern I used – the irregularities look too planned for my taste – so I’ll try to make it look more realistic in that regard. The texturing is improved by the paint job, but the base texture work could still be refined more. I also want to do a little more finishing work to the piece: I’d like to spray it with a matte finish to protect it without giving it too much shine, I want to add in some small rocks and perhaps vegetation in some of the crevices, and I have plans to include some shackles and/or torches coming off the wall, so I’ll need to devise a mounting system for that.

One of the most common rejection reasons on iStockphoto is the “overfiltering” rejection, largely because it encompasses so many possible causes. In this new “Rejections” series, I’ll look at some of the more common causes of the various iStockphoto rejections, and what you can do to avoid the most likely pitfalls.

When it comes to “overfiltering”, one of the things I hate to see is poor editing or cloning. This often happens when the contributor is trying to remove some other rejectable element in an image – a visible person for whom there’s no model release, a logo or other copyright-protected element, or sensor spots, perhaps – but does it in a less than seamless way. Any editing of this sort has to appear natural and realistic, and not look like the file has been edited.

Let’s use this image of a tent in the backyard to illustrate the wrong ways to go about removing material, then we can look at a good solution. As you can see, the bottom corner of the tent door has a manufacturer’s logo and a hand-painted “W1″. The logo is a no-go for iStock, so it will have to be removed. The “w1″ is not really problematic in terms of iStockphoto’s standards, but it’s also not doing me any favors in the image. It’s a distraction, and this would be a more usable image without it, so out it goes.

When it comes to doing it wrong, the three main culprits are The Blur, The Splotch, and The Clone:

The Blur

This is the easiest and least effective way to edit out undesirable elements. The contributor often selects the element (using a mask or lasso selection) and just adds the Gaussian Blur filter. There’s almost no way to make this an effective edit; selectively blurred areas stick out like sore thumbs. Not only does it leave an ugly and distracting region in the image, it also looks completely unnatural. Even more delicate selections than I did here are usually unconvincing. They will, by definition, be more blurry than their surroundings, and more blurry than the lens would have made it. There are really very few applications of this that would look natural.

The Splotch

This is what I call those cases where the contributor opted to cover over the undesirable mark. In general, that’s a fine idea, but The Splotch is a bad way to do it. This is when it’s covered over with just a painted-on single color brush stroke. In the tent example, the areas surrounding the marks we want to remove are mostly monochromatic – the maroon surrounding the “W1″ and the off-white surrounding the manufacturer’s logo. So you might think that a single color stroke could work. But click on that thumbnail above to see how ineffective that is. The big problem with The Splotch is that it makes no effort to replicate the texture of the surface.

The Clone

The Splotch fails because it doesn’t take into account the texture of actual material, so why not use the actual material to cover up the marks? Great idea, and usually when I need to fix an area, that’s my first step as well. But it must be done right. As a click on the thumbnail to the left will attest, it’s very easy to do it wrong. This is more common than you might think. I probably end up rejecting up to five or so images a day for overfiltering of this nature. It comes from a lazy application of cloning techniques. The clone stamp source is set once, and the contributor just click-click-clicks until the offending elements are covered up.

So if that’s how to do it wrong, how do we go about getting it right? The short answer is that there’s no short answer. It really depends on the specifics of the image. For this tent image (the iStock version, by the way, lives here), I did a combination of cloning and healing brush.

Step 1 was some cloning onto a new layer (everything is done on new layers, of course).  The key here is to try to match the lights and darks of the fabric wrinkles. There is not going to be a single clone source that will do it all, so I make sure to continually set new clone sources, taking some highlights from one area, some shadows from another. Don’t be afraid, also, to cover over areas that are actually just fine in the original. Sometimes it takes doing that to make the edits appear to blend in seamlessly. Be on the lookout for obvious pattern repetitions. The crop below shows the results of the cloning step. It’s not bad, but I’ve highlighted areas that are too obvious.

So we move on to step 2. On another new layer, I start using the Healing Brush tool (J). The key to using this tool is to keep a small brush and use multiple small strokes. The tool is most effective when you’re using it away from any hard edges or abrupt color changes. With those guidelines in mind, I did some selective healing to the trouble spots above, and got this:

And there you have it. A much smoother and more seamless edit. The same techniques apply whether I’m removing logos, skin blemishes, wandering background strangers, or annoying sensor spots.

I touched on this briefly in a previous post, but after a recent conversation with a contributor, I decided to expand on this idea. When you get keyword rejections or keyword removals from your iStock uploads it’s usually because of one of three reasons:

1. The term was spam.
2. The term was a “multiple incompatible”.
3. The term was a stretch.

#3 is where I want to spend some time with you today. Stretch terms are those terms that are related to but not conveyed by the subject and context of the image. They usually take the form of concept terms. A real life example to get us all on the same page:

This is one stern math teacher. The kind one used to know in high school after being a little late with the homework.  Not that that ever happened to me, of course.

The terms currently applied to this image:

• Education
• School
• Classroom
• Only Mature Men
• 40s
• Teacher
• Serious
• Mathematics
• Mathematical Symbol
• Whiteboard
• One Man Only
• One Person
• Horizontal
• Photography
• Mature Adult
• Mature Men
• Necktie
• Looking At Camera
• Displeased
• Anger
• Indoors
• Head And Shoulders
• Color Image

What’s not on the list? “Homework”, for one. This image is not about homework, it doesn’t depict homework, and a buyer looking for “Homework” shots would pass over this image in a heartbeat. “Homework” would be a stretch term, and it would be removed by an inspector without hesitation. It’s the kind of stretch term that we see people try to use often.

Enter the concept of Opportunity Rejection. An Opportunity Rejection is a rejection that should spur you on to create and upload something new. By removing the term, the inspector has essentially said to the contributor that this image of a disappointed teacher doesn’t convey the notion of “Homework”. Contributors should take that as a challenge to then create an image of a stern teacher that does convey the notion of “Homework”. By putting the term “Homework” on the original image, the contributor made it clear that he felt there was a market for “Stern” AND “Teacher” AND “Homework”. If there’s a market for it, why on earth wouldn’t he create an image that is undeniably and unequivocally about those things?

The benefit of turning an opportunity rejection into a new separate upload is that you can create a file that is geared specifically to the terms you want to use. You can achieve that Laser Focus I mentioned last week. An image geared specifically to the concept will likely sell better for that collection of terms than will a generalized image. The image above is mostly a generalized image. The teacher is stern, but nothing in the image indicates why that is so, so it’s left open for a number of interpretations, making it a good general option. The flip side of that is that I can’t keyword for every possible interpretation – I can’t drill down to the specifics of “Homework” and “Excuses” and “Poor Performance” and “Class Clown” and every other concept that might disappoint a teacher (again, not that I’d have any first-hand experience with the things that annoy teachers).

I say “mostly generalized”, however, because I did make some specific choices as well. For one thing, he’s a white male adult. So those specifics preclude other specifics. But that kind of thing happens almost any time you use a model. The other specific choice was to make him a math teacher. The formulae on the board behind him put him pretty squarely into the math realm (maybe physics, but I chose math – again, it’s not both). If a buyer needed an image of a disappointed teacher of chemistry, this file would either not work for their purposes, or they’d generalize their search (”Teacher” instead of “Teacher” AND “Chemistry”) and edit the file for their own needs.

Sorry, this is not a post about achieving photographic focus. You and your camera will have to figure that one out without me. This is about achieving content focus in your iStock images.

Content focus is, in short, the subject of your image; the content or concept that the image is meant to convey. One of the most important factors in a stock image is how quickly, clearly and concisely the image conveys a concept. If the image is about a college graduate entering the workplace for the first time, for instance, you’re going to want the image to convey that concept quickly and unambiguously. Because if your image does not, someone else’s will and that’s where the customer will spend his credits.

Because of the way the latest iteration of iStock’s “Best Match” algorthm works (the so-called “BM 2.0″), images will move up the rankings the more clearly they are associated with a given term. Images that are muddled in their presentation or that do not clearly convey a concept will find themselves languishing near the bottom. If you want to appear near the front of a Best Match sort, be sure your file is unambiguous in its meaning.

Lets look at an example. This image is a pretty good match for “Paperwork”, showing up in about the 15th place in a Best Match sort for that keyword as of this writing. It’s also a fine match for “Drowning”, currently occupying the #1 slot there. It also appears in a search for “Emergence”, but while “Emergence” is an appropriate term, this image isn’t really the best image in our database to convey that concept. For exactly that reason, it doesn’t even show in the first 500 results for that term.

The million dollar question, of course, is how to get your own images to move up the rankings? There are a few techniques you can consider:

• Trim the fat: One of the best things you can do is remove keywords that are unlikely to be of any help at all. If I had “Life Vest” on that drowning image above, that would be a prime candidate for removal (and let’s be frank, a term like that should be removed before you even upload). There is no “Life Vest” in the image, and no one searching for “Life Vest” is likely to buy an image that doesn’t contain or even relate to the term.But what about “Photography”, “One Person”, “Vertical”, and those other meta descriptors? No one is likely to search for “Vertical” to buy this image, right, so should that be trimmed? Nope, those kinds of meta terms are fine to keep on because they are usually used in a boolean mode – “Drowning NOT Illustration” or “Paperwork AND Vertical”. Those terms are not subject and concept terms, so they don’t muddy the subject and concept focus of the image.
• Consider Underserved Niches: It’s a whole lot easier to move up the results rankings when there are only a hundred or so images between your image and the top. Lets say you’ve hired a model, and you have a prop closet and wardrobe full of options. You can dress your model in a suit and tie and do a “Businessman” series. Upload that as your concept and you’ll be up against some 77,000 files. Might take a while to climb that ladder. Or you can put him in jeans and a flannel shirt, plunk a guitar in his hand and sit him down by a campfire to make him a “Camp Counselor”. Go search on iStock and see how many files you’re up against in that case. I’ll wait here.Back so soon? Now, it’s clear that “Businessman” is a more common search term than “Camp Counselor”. I’ll grant you that. Is it 20,000 times more common? Let’s look at this from the view point of expected value. This is an Econ 101 concept that seeks to describe the expected potential for a given action. For easy math, let’s first say that the average royalty per sale is $1. Now let us imagine that the search term “Businessman” leads to 100,000 sales per day. That comes to $1.29 per “Businessman” image per day (assuming that sales, on average, will spread out to everyone in the search results. Which they will not, of course. If your image is on page 50+ for that search term, you’ll be very lucky to have “Businessman” lead to any sales at all).

  Now lets say that “Camp Counselor” gets 1% of the search traffic that “Businessman” gets. That would be 100 sales per day for that term*. If there were just 10 matching files, they’d each have an expected value of $10/day. Much sweeter return than the $1.29 “Businessman” average. Of course, this example uses some assumptions. You’d have to make your own assumptions about the numbers in order to determine your own expected value.

• Generic is A Focus: I was recently chatting with another inspector about the notion of a generic file and specific keywords. The file depicted a series of generic spice rack spice jars. The keywords included every kind of spice imaginable. Now, we had to remove all those spice keywords, because the image didn’t actually depict them. The focus of the image is “Spices” in the generic. There are plenty of times when a buyer needs an illustration depicting a parent term in the generic, and not the specific instantiations of that parent. A wise contributor, however, would see the opportunity in this. Illustrate the generic parent term, and keyword it just for the generic, then also create the specific instances of the child terms (the parsleys, sages, rosemaries, and thymeses). The contributor would have to be careful not to make the specific instance illustrations too similar to each other, to avoid the “Duplicate/Serial” rejection, but I think a talented illustrator should be able to do that without much problem.
• Keep The Content Clutter Free As Well: Clutter-free keywording is important, but let’s not forget the content of the image must also immediately convey the concept. Take care to include just those image elements that lead the viewer to that concept, and eliminate everything else that doesn’t support the concept. This isn’t to say that all images must be isolated on white. That would be horrible. Keep in elements that set the stage and make the photo convincing, but if that coffee mug on the table isn’t helping to set the stage or sell the concept, lose it.

If you find that you’d like to improve your Best Match standings, try implementing some of the above. Feel free to drop a line in the comments to let me know what you find.

——————-

* Cry Foul! How can I claim the term Camp Counselor gets 100 sales a day when the few files keyworded with “Camp Counselor” together have a total download count of less than 100? One of the assumptions in the calculations above is that the demand for “Camp Counselor” is 100/day. The existing files in that search may not be suitable for the customers looking for that term, so they may not be buying any Camp Counselor images at all.

For this first DIY lighting post, I decided to go uber-simple. We’ll be putting together a lighting box for shooting isolated products. It couldn’t be any easier.

Step 1: Get yourself a box.
Step 2: Don’t throw away the old vinyl liner from the shower curtain.
Step 3: Try not to cut your fingers with the utility knife.

That’s about it. I took a large box (24″ x 30″ x 24″), and cut away the top and a side. I also cut a window in the front to shoot through. Over the open top and down the open side, I draped an old vinyl shower curtain liner (My wife was kind enough to save for it me, rather than throw it away, when it became too disgusting to keep using). As you can see, I opted for the hand clamp method of connecting shower curtain to cardboard.

Inside the box is some posterboard to create my white sweep.

In the setup photo above, you can see the box setup works fine with just a single speedlight. The sample image below was shot with a single speedlight coming through the top of the box. I experimented with a two light setup – one above, and one from the side – but frankly one light does the trick here. If you can afford some noise reduction (that is, if you don’t plan on submitting to iStock, or if you plan to downsample before submitting to iStock), you can drop the speedlight power down to about 1/4 and bump up the ISO to make up for the exposure difference. That allows you to save on speedlight batteries, and have a nice fast recycle time.

The box I used allows for a set up large enough to shoot products anywhere from very small to the size of a small cat. My cat is largish and uncooperative, so you’ll see no shots of her here.  This would be a great set up for an eBay photo shoot, or even for a catalog of smallish objects. The big downside is that it’s a big box that doesn’t store well; there’s no way to collapse it down. The upside, of course, is that it was basically free because I used found objects. If I had to buy the box and the vinyl, I’d have been out maybe $15.

Back in part one of this article, we talked about some ways to help your images get found on iStockphoto. Those tips mostly centered around things you can do before you take your shot; how to plan your shoot and compose your shot to help separate from the pack. In this part, we’ll look at some techniques for after the shot is already online.

Once the image is approved and available on iStockphoto, you’ll want to maximize exposure. There are a few ways to go about this: keywords, lightboxes, and cross-promotions are the biggest ones currently available to you.

Keywords
We discussed keywords some in the first post, so I won’t spend too much time on it here. Instead, I’d like to propose a couple of great ways to help think of keywords for your images. The first is to try to find the image yourself. That is, do a search on iStock the way a customer might. Think of how a customer might start that search, what keywords might he use, which ones would he exclude with a Boolean search string. Imagine, for instance, you have an isolated shot of a toothbrush. Pretend now that you’re a customer looking for just that – an isolated toothbrush, nothing else. A search for “toothbrush” reveals 1500+ other toothbrush shots. As a customer, the first thing I might do is remove all of the shots that also include “Dentist” and “Brushing”. That brings the count down to a more reasonable 500 images, and should provide good reason to avoid using related keywords. If you had included “Dentist” and “Brushing” as keywords, your image would now be part of the 1000 images removed from my result set. Be the customer. It will help you think of words that should and should not be part of your set of terms.

The next handy tool for keywording is the Suggestinator at iStock-apps.com. It’s a simple tool that lets others suggest terms for your images. It’s easy to use, and generally reveals some pretty good options.

Finally, iStock has started to encourage keyword requests in the Critique Forum. Start a new thread there, with “:KEYWORD:” in the thread title, and get lots of great suggestions. I did it recently for this rocket ship illustration, and was rewarded with maybe a dozen new terms, all of them appropriate and accurate.

Lightboxes
Lightboxes are collections of images on iStock. There are two kinds: public and private. Public lightboxes are curated collections of files all around a specific topic, such as “Families”, “Coffee”, “Mountain Climbing”, well you name it really. Any topic that has images on iStock probably has a lightbox or two around it as well. The files in the lightboxes are the ones that the curator feels best suit the theme. They may be those images that have the most sales among the images that fit that topic, or those that have the fewest, or those that the curator likes best, or those that were contributed within a given time frame, etc. Curators are free to add whichever images they like to their lightboxes, and are really under no obligation to add any at all. If you find a public lightbox that you feel might be a good showcase for one or two of your images, send a polite sitemail to the lightbox manager and ask if they would add yours to it. Some will, some won’t. Showing up in the lightbox may help visibility, or it might have no effect at all. It’s hard to know, but early on, it’s fun to try to increase visibility, so why not go for it.

Cross-Promotion
This is a big one. Cross-promotion is simply providing customers with a link to see some more of your relevant images. In the description field of your isolated toothbrush image, you can give folks a link to see your other dental hygiene images. I wrote a pretty thorough tutorial on how to set this up for your files, which can be viewed here.

Any other great ideas for getting found on iStock? Leave me a note in the comments or send me an email, and I’d be happy to add it here!

I shot this image of a newlywed couple a few years ago. Generally, we were pretty pleased with the results, but there’s one thing about it that has bothered me from the very beginning. We shot this image with a single strobe off to camera left, reflecting some light back onto the models from the wall off to camera right. A very very yellow wall.

The yellow cast on the left side (camera right side) of the models’ faces has always kept this from being an image I can truly be proud of. It’s time to finally fix that. Using Hue/Saturation, this is one of the easiest fixes we can do.

I opened a new Hue/Saturation adjustment layer. I could have just performed the Hue/Saturation adjustment to the existing layer, but I am a firm believer in the “exit strategy” philosophy: make sure that any changes you make to the file are non-destructive, and can be undone. Or more importantly, make sure they can be partially undone, as we’ll see here in a moment.

The key to this fix is the drop down menu in Photoshop’s Hue/Saturation dialog box. Rather than adjusting the hue and saturation for the entire range of values in this image, we can use that drop down box to select just the ones we want to modify. In our case, that’s the yellow values. Select “Yellows” from the drop down, and you’ll see the color range selector at the bottom of the window will now indicate the smaller range of affected values. There are three sections to that range indicator: the left fringe, the middle bar, and the right fringe. Input values (those colors above the selection indicator) will be affected by the slider actions and become the output values indicated on the color bar below the selection indicator. Those input values above the left and right fringes will be modified to lesser degrees than those values above the middle bar – from 100% for those values immediately adjacent to the middle bar, down to 0% for those values at the end of the fringe (marked with a rhombus shaped handle).

If that last paragraph was a little kooky, try playing with the sliders yourself to see what I mean.

With the proper input value range selected, I pushed the saturation all the way up to +100. This gave me a good sense of what values would be affected. I had to tweak the fringe sliders somewhat to get the range just right. Once I did, I could start making the real edits. I dropped the saturation down, increased the lightness, and slid the hue over a few degrees to make the last traces of yellow turn more red. Here’s what that gave me:

Not bad at all, but as you can see, it took out too much of the bride’s skin tones, leaving her much too pale. This is where adjustment layers have direct adjustments beat, hands down. I just masked out the adjustment layer, to allow more of the bride’s original image coloration to pop through:

Voila! All traces of the offending yellow wall have been removed in just a few easy steps.

iStockphoto.com has become one of the largest Royalty-Free photo libraries out there. Contributors to iStock have a great opportunity to reach a global marketplace for their images. That opportunity is only great, of course, if the images can actually be found. As an image inspector for iStockphoto, I get to see thousands of images each month as they move through the inspection queue and into the collection. So how do you go about getting your image noticed in the vast sea of 2 million+ images (nearly 3 million as of this writing)?

There are no guarantees, of course, but here is the first part of a two-part series on things to consider for your images. It may not come as a great surprise to find that much of the most important work you can do to get your images found actually takes place before you click the shutter.

Don’t push into the crowded bus
You aren’t really jockeying for position against the other nearly 3 million images. You’re only up against the images that share primary keywords with your images. You may have just shot the world’s best image of an apple, perfectly isolated on an infinite white sweep. Congrats! It’s now one of over 5,000 images that comes up in a search for “Isolated Apple” on iStock. Designers will often search well beyond the first page of results to find the image they need, but even the hardiest of image researchers will hesitate to comb through over 5000 images.

Let’s say instead of “Apple” and “Isolated”, your image was “Apple” and “Fire”. Now it’s one of 19 images. That’s a much greater ratio. The downside? As of today’s writing, those 19 images have a combined download total of about 75. Not a whole lot of demand for Flaming Apples.

Your job, then, is to find that sweet spot. Look for the underserved image markets on iStock. Often all it takes is adding a key element to your image (”Fire”, in our example above). Rather than shooting an attractive model in a business suit isolated on white (and, yes, I know I’m guilty of that as well), shoot her doing something that isn’t overdone in the library. Not talking on the phone. That is overdone. Have her balancing the books. Or balancing on a tightrope. Or balancing another businessperson on her head. This, of course, has to be thought out in advance of actually creating the shot.

Keyword Thoroughly and Accurately

Each keyword you add to an image is another search under which your image will be found. It’s absolutely in your best interest to be thorough about your keywords. This image has the following terms
applied to it:

• Single Object
• Small
• White
• Toned Image
• Horizontal
• Scarcity
• Place Setting
• Plate
• Nobody
• Empty
• Dishware
• Sparse
• Green Pea
• Food
• Silverware
• Poverty
• Famine
• Hungry

You can basically break the list up into three categories: the literal, the conceptual, and the compositional.

Literal
The literal contents of the image (plate, silverware, green pea) are the easiest keywords to add. Just look in the picture and name the key elements that are in there. An important note, however, is to avoid “laundry lists”. As an inspector, I often have to reject images for keywords when they rattle off every minute detail on the image. An image of a person, for instance, probably does contain a Human Face, Human Lips, Nose, Human Eyes, etc., but if these aren’t what the image is about – if they’re not key elements of the image – don’t bother adding them.

Another point about literal keywords – consider using parent terms if they still make sense. For example, in the image above, I have “Green Pea”, and its ancestor term “Food”. What I didn’t include, however, were the parent terms between Food and Green Pea; The actual taxonomy is Food > Vegetable > Legume > Pea Family > Green Pea. “Food” makes sense for this image, as does the specific food “Green Pea”. Arguably, “Vegetable” makes sense, but I get the sense that someone searching for “Legume” or “Pea Family” is more interested in a shot more devoted to those keywords.

Conceptual
Conceptual keywords are harder to define, harder to use, and harder to inspect. This is a muddy and slippery region of the keywording landscape. I feel the image above portrays concepts such as “Hunger” and “Scarcity” pretty well. It doesn’t at all represent “Business” or “Friendship”. What about the gray areas? Agribusiness? Competition? One could probably make an argument for those words, but my feeling is that if one has to make an argument for them, they’re probably not the best candidates. Stick to the ones that you had in mind when you made the shot.

Compositional
These are the metadata of the metadata world. These terms represent the photographic details of the image more than the contents. Horizontal, vertical, diagonal, above, below, close up, wide angle, worms eye view, monochrome, cross-processed. Again, this should be a pretty easy set of terms to define and use.

Don’t Spam
Spamming is including terms that are outright incorrect, or are so stretched so thin they’ll snap. Spamming on iStock also includes incorrectly disambiguating your terms. Spammers will be drawn and quartered on sight, so don’t do it.

Be About Something
The way to achieve both of the points above is to have your image be about something. That is, before you take the shot, decide what the shot is about. Start with the concept. Be intentional about the photograph. Don’t just set out to take a picture of a tree. Decide what it is you want the picture of the tree to portray. In essence, this means preselecting a literal keyword and a conceptual keyword to describe what you’ll be shooting. Consider the following combinations of “Tree” and conceptual terms:

• “Tree” + “Childhood”
• “Tree” + “Despair”
• “Tree” + “History”
• “Tree” + “Satisfaction”
• “Tree” + “Frustration”

Could you envision an image that could be accurately keyworded with any of those combinations?
By being more intentional about your stock photography, you gain several advantages. The first is that you can actively choose to put your image in an underserved market niche. Let’s say your research shows that there are over 220,000 images of trees, but fewer than 100 also include a given conceptual keyword. If you feel as I do that that’s an available market niche, you could plan your tree shot to portray that concept. Now it’s not just a tree image, it’s a tree image about something, and what’s more, it’s one of only 100 images about that something.

Being about something also helps in the keywording phase. The conceptual keywords – that category that is often so difficult to do – is basically handled ahead of time. You have alreay decided that your tree image would portray a given concept. The keywording is guaranteed to be easy and accurate.

Finally, by deciding ahead of time what your image is about, you give yourself the ability to remove distractions and make a better image. Does that mailbox in the background really help your tree image portray “desolation”? Is that garbage at the base of the tree really the best way to show “Stately Elegance”? Stock photography is all about iconic images that read quickly. Your job (ok, one of your many jobs) is to eliminate those elements that prevent the image from being iconic.

Join us for part II of the series, which will explore those things you can (and those things you cannot) control within iStockphoto’s environment to help boost your image’s exposure.

A question that comes up periodically among new contributors to iStockphoto is “how do I attach multiple model releases to my upload?”

It’s a good question, as there is only one upload box in the interface for attaching a model release. If you upload one release using that box, then try a second, it’s only the second that gets attached.

The solution is pretty simple: Just combine your releases into a single JPG document, and upload that. Your model releases should already be JPGs; just open a new photoshop document large enough to accommodate your two (or three or four) releases, and paste them into it. I like to do mine side by side, I see lots of folks stack them vertically into a single long document. Once, when I had five or six releases all in one document,  I went one step further: I tiled them into a document 3 releases wide by two releases high, and included a reference key indicating which release went with which model. It’s not required, of course, but that little step certainly can make things nice for the inspectors.

Time for another photoshop graffiti tutorial. This time, lets look at how to turn yourself or a loved one into a 60’s socialist revolutionary icon, a la Che Guevara.

Lets dive right in. First off, we’ll need a few photos. I’ll be using this friendly face and wood fence background, both from my iStockphoto portfolio.

To begin, we’ll need to do some preparation work on the photo of the man. Just as with our previous graffiti tutorial, we’re going to want to have our graffiti look like it was stenciled on and spray painted. To do that, we’ll have to adjust the face down to one color. Photoshop provides an easy way to do this. It’s called Threshold.

I opened the image of the man and cropped it down to just the area around the face. Before I applied Threshold, I wanted to remove some of the extraneous background. The threshold function will eliminate a great many details, so it’s OK to do a rough job of isolating the subject. Besides, if we’re going for a stenciled look anyway, edge details are not too significant. Here’s what I had:

The magic “Background Eraser” tool took care of most of the isolation for me. It’s an underused tool, but it’s so useful. After isolating the subject from the background, I added a threshold adjustment layer.

The threshold adjustment turns the image into just black and white pixels. No shades of gray, no nuance, just black and white. The slider lets you determine the threshold (hence the name) at which tones are converted to black or white. Play with the slider to see what I mean.

After settling on a good level of black and white detail, I opened the Channels palette, and copied one of the channels into a new one (as the image is black and white, it doesn’t make any difference whether you choose the R, G, or B channel to duplicate).

I didn’t really like the hard-edged bottom, so I painted a few small edits in this new channel. I then inverted the colors, swapping black and white. The end result was this:

We’re almost done with the stencil preparation. I went back to the layers palette, and added a new layer filled with plain white. I went back to the channels palette and loaded the new channel as a selection. That made it so that I had a selection of the stencil, where all the pixels were white. I dragged that selection into my background (wood fence) image:

Well, as graffiti goes, that’s pretty horrible. Lets make it a little (or a lot) more believable. First off, I went into blending options and straight to the bottom of the dialog box, where it says “Blend If:”. I moved the white point slider for “This Layer” down a notch, from 255 to 254. Since all the pixels on this layer are 255/255/255 white, I only have to hide that one value to make the whole thing go blank.

Then I went into gradient overlay and added a gradient. I like to use gradients, rather than single color overlays. Nothing in real world imagery is truly one color, so nothing in my fakery should be either. The important step, then, is to set the blend mode of this gradient overlay to “Multiply”. As you can see here, that lets the background wood texture show through.

Much better.

At this point, we’ll want to make the paint look a little more stenciled. There are two things I like to do to help create that effect: paint drips and overspray.

Paint drips are just the excess bits of paint that dribble down off the bottom edges of the painted area. These can be painted on directly, or you can use some vector art or custom shapes to do the trick. I tend to paint it on, because it’s quick, and at low resolution it’s more about the impression than about the details. If this were a high res piece, intended for print, I’d probably fire off a few more credits at iStockphoto and get one of the high quality paint drip vectors there.

The other effect I like to add is overspray. This is the spray paint that hits the wall outside the far edge of the stencil. You can see the effect in a real world photo here:

And a close up of what I’m talking about:

Creating this effect in our Photoshop document is actually identical to how it’s created in the real world; we’ll make a stencil (mask) and “accidentally” paint outside it. First the stencil. I used the shape tool to make a rough rectangle around our face. I then rasterized the shape, and roughed it up some using the Liquify filter, the Blur tool, and the Smudge tool. The areas where the edge of the mask are blurry will result in blurry edges to our overspray, similar to the effect of a stencil not placed immediately up against the surface being painted on.

I made a selection of that mask, inverted it and selected a rough brush. Make sure you’re painting on your face layer, as that has the layer effects. When I paint overspray, I try to build up the corners, to help indicate where the edges of the stencil would have been. It’s easy to go overboard with this technique, so experiment to find what works.

Here are the results of the detail roughing. Yeah, I blurred part of the face as well. I think it makes it look a little more stenciled.

One final effect I like to add is a little dimensionality, by having the paint match the contours of the fence beneath it. To do this, I hid all of the layers except the fence, then went over to my channels palette. I picked the channel that had the greatest contrast between the wood surface and the black spaces between slats. For my image, that appeared to be the red channel, though that could be different for you if you used a different base image. I duplicated this channel, then gave that new channel a levels adjustment to leave me with just the black spaces between the slats (a little of the wood texture is fine as well). Invert the channel, and make a selection of it.

With the selection still active, add in a Solid Color adjustment layer, filled with white. The selection automatically becomes the layer mask. Unhide the fence and face layers so you can see what’s happening here. With our new white masked layer selected, use the arrow keys to move the white pixels over to the left. I only had to nudge it a few pixels; depending on the resolution of your image, your mileage may vary. Set this white layer’s blend mode to “Overlay”, and reduce the opacity of the layer to about 20 or 30%. You should see that the painted face now has some highlights where the wood slats round over. Now here’s the cool trick: in the layers palette, make a selection from the face layer’s pixels, invert the selection, and in the layer mask for the overlay layer, make that selection black. That will prevent the overlay from lightening the “unpainted” wood.

That’s it for this tutorial. Add a comment here and let us see how you used these techniques to turn an average Joe into a socialist icon. ¡Viva La Revolución!

Click the image for a larger view of the finished piece.