I80

,0091 Hi

THE 100% RULE

Finally, before we get started on some practical examples of problem images and solutions, let's agree on a simple rule you need to follow from now on: the 100% rule. Always check your images at 100% magnification on screen in your chosen image editor. That way, you are looking at the actual pixels and you will see defects much more clearly than at a lower magnification.

NOISE

Time now to turn to some specifics of problems with images that lead to rejections. We will start with public enemy number one—noise. To be a big noise in microstock, avoid noise!

Digital noise is the number one reason why images are rejected by our photo screeners.

Jon Oringer, CEO, Shutterstock

Noise takes different forms and is usually most clearly visible in areas of even color, such as a clear blue sky. Luminance noise looks like film "grain" (Figure 5.5). Variations in the sensitivity of individual photosites (the bits of the chip that record light) in digital cameras can cause the appearance of grain in an image. When this is even and very fine, it may not detract from an image at all; in fact, there are a few circumstances when adding a little grain can be beneficial (as we will see later). Generally, however, this grain-like noise is bad, and too much will lead to your images being rejected. Compact digital cameras use smaller sensors with poorer light-gathering properties, and this can lead to more noise (Figure 5.6).

Another tip is to expose to the right. This as an advanced digital camera shooting technique that you can try if your camera has a histogram display that you can use to check the exposure of your photographs (most do). Before explaining how to do this, it is worth discussing the theory.

There is a noise floor in all imaging devices. Think of this as being like background noise affecting your stereo system. The more you turn up the music volume on your stereo, the more difficult it will be to hear any background noise from the electronics. Similarly, with a digital camera, the more light you let in, the more any noise will be suppressed. But the risk is that if you take this too far, you will overexpose highlight parts of your images, leading to those horrible white blank areas. So, it is a balancing act to "max out" the exposure without burning out image highlights.

Figure 5.7 illustrates histograms of digitally captured images. The histogram on the left has a gap to the right, which means the

FIGURE 5.5 "Spring Field Background." Pictures like this, with large areas of a single color, are more likely to show noise. © Douglas Freer/ Shutterstock

FIGURE 5.5 "Spring Field Background." Pictures like this, with large areas of a single color, are more likely to show noise. © Douglas Freer/ Shutterstock

image has been underexposed. While this can be corrected later, in the process of making adjustments to brighten the image using curves or levels in Photoshop, you will increase visible noise in the image, particularly in the darker shadow areas of the image, especially if you are working in 8 bits. The histogram on the right, Figure 5.7B, is much better, with lots of data from the shadows to the highlights, but the highlights are not "burned out." The only difference between the two images used for these examples was the exposure setting.

Of course, you might find that your shot is overexposed, in which case you'll need to reduce the exposure, not increase it. Try if you can to expose so that there are data all the way to the right but not beyond it. Not all cameras allow you manually to set the exposure, and generally it is not necessary to do so. However, it is normally possible to

FIGURE 5.6A and B Two sky sections: A showing noise (quite fine, not too bad), and B with the noise removed using noise-reduction software. B is more likely to be accepted.

FIGURE 5.6A and B Two sky sections: A showing noise (quite fine, not too bad), and B with the noise removed using noise-reduction software. B is more likely to be accepted.

FIGURE 5.7A and B A illustrates a "bad" histogram from an underexposed shot. B is a lot better, with no large gaps at either end. Generally aim for a histogram where there are data right up to but not touching the right side.

use exposure compensation to increase exposure, so a good workflow might be as follows:

• Select a low ISO setting, where possible. Make sure, though, that you have a high enough shutter speed to avoid camera shake. (You could try a wider aperture to achieve this.)

• Compose and take your shot; check the histogram.

• If the histogram shows problems, adjust the exposure to get the best histogram and then reshoot.

• If you are in any doubt, bracket your exposures; that is, take several versions of the same image at different exposures and choose the best later.

COLOR BLOTCHES

Another sort of noise is color noise that can look like uneven blotches, as seen in the next image where an attempt has been made to extract shadow detail from a church interior shot (Figure 5.8). Provided noise in luminance or color forms is not too extreme, you can zap it in a number of ways.

FIGURE 5.8A and B Before (A) and after (B) filtration of an image with blotchy color noise.

NOISE-REDUCTION SOFTWARE

The simplest solution for noise is to use a good specialized noise-reduction software package like Neat Image (www.neatimage.com) or Noise Ninja (www.picturecode.com). These products work by measuring image noise through samples taken from within the image and

FIGURE 5.8A and B Before (A) and after (B) filtration of an image with blotchy color noise.

then applying complex algorithms to remove the noise without destroying too much image detail (Figures 5.9 and 5.10).

Photoshop CS3 has its own built-in noise reduction program, accessible from the Filter menu (Filter > Noise > Reduce Noise). It is not as complete or, in my view, as effective as specialized noise-reduction programs, but it is better than nothing. Try experimenting with the settings.

One of the problems with all noise-reduction software is that even when you exercise great care, it can still remove some useful image detail when it removes noise, leaving the image looking too smooth and "plastic." To avoid this, experiment by using different amounts of noise reduction. Use the "fade" command in Photoshop to reduce the amount of filtration after the event if necessary. Use only the minimum amount of noise reduction necessary to remove the worst of the noise.

FIGURE 5.9 Neat Image working on a high-resolution JPEG. The green box is the sample area used to build the initial profile. This can then be fine-tuned.

FIGURE 5.9 Neat Image working on a high-resolution JPEG. The green box is the sample area used to build the initial profile. This can then be fine-tuned.

FIGURE 5.10 The Noise Ninja Photoshop plug-in profile window.

Noise Ninja Unlicensed

Noise Nn)a Profiles View Help

FIGURE 5.10 The Noise Ninja Photoshop plug-in profile window.

Noise Ninja Unlicensed

Noise Nn)a Profiles View Help

0 Lumrtanc* Q Coton

Aulo pro Pled

Prod* Flter Noise Bruîh Actions Annotions Image info Noise indtïi s< >» chrom»)

Prod* Flter Noise Bruîh Actions Annotions Image info Noise indtïi s< >» chrom»)

Mill lHl< ill I I lllh Jllh llln lllli All lifo Ith I III n II Ilk I ^^ I

Hi I (l I, l! 111 III 11 Uli I |]I|, 11 II. Il II. Ai Ikjft r.lllta I Profile char I I

Iti. It. llliilllnllliillliillliilllJtllli! I i.lllii -

ii . lu. iln.ilfeithiliiliii

0 Lumrtanc* Q Coton

Aulo pro Pled

Another tricky issue can be that noise is often only visible in part of the image, say, in the shadows or the sky. In Photoshop, there are a number of ways to make sure that noise reduction is applied only to those parts of the image that really need it

• Quickmask. In Photoshop, use the quick mask tool and paint over areas you don't want noise reduction to impact before running the noise-reduction software. This is fast and effective, but you have to judge in advance which areas to exclude from noise reduction.

• Selection. Use the lasso tool to do the same job as Quick mask; then feather the selection before running noise reduction.

• Historybrush. After you have run your noise-reduction software, select a history state from the time immediately before you ran the software, and use the history brush on any areas where too much detail has been lost. You can build up the effect by selecting less than 100% opacity for the history brush and then painting over the area where you want to reduce the effect of the noise reduction software. I suggest a low value, say 20%, for the history brush.

• Layers. My personal favorite. Duplicate the layer and then run the noise-reduction software on the bottom layer. Use the eraser tool (select an opacity of about 20%) with a soft edge to gently rub out those parts of the top layer where noise is intrusive; or you can even vary the opacity of the entire layer until you are satisfied that you have the balance between noise and detail correct.

Sometimes getting the balance right between noise removal and detail retention is difficult. If you overfilter an image, the microstocks will reject it. If you have too much noise in your image, the microstocks will reject it. Talk about being caught between a rock and a hard place! Using the above tips will help you minimize the risk of rejection; but, if all else fails, you have to get devious and add noise to the filtered image.

"Whoa!" I hear you murmur. "Add noise? But you just mentioned that noise is bad!" Strangely, adding some noise can add realism back to the filtered image, if done with care. It is not cheating. All that matters is that the image looks good to reviewer and buyer alike.

My suggestion is to use a free Photoshop plug-in from Richard Rosenman called the Grain Generator, available from www.richardro-senman.com. Choose either minimum or standard, setting between 2% and 7%, and see how the finished article looks at 100% on the screen. It only works on 8-bit images, but you'll be saving your files as 8-bit JPEGs before submitting them, so that's no problem. I think you'll find that adding a little noise can sometimes have real benefits. I have used this trick quite a few times to good effect, and the images look better for it, even when printed! Of course, it is better to prevent noise in the first place, but shooting conditions may not always allow for that.

An alternative to the Rosenman filter is to use the noise filter in Adobe Photoshop (Filter > Noise > Add Noise). Select Uniform and around 5%. I find this is easier to use than the grain filter in Photoshop (Filter > Artistic > Film Grain).

ARTIFACTS

I hate artifacts. They are more trouble than the sort of noise mentioned in the preceding section. Most often they are the result of compression applied when saving images in JPEG format.

Figure 5.11 is an example of part of an image with heavy JPEG artifacts visible. The most common cause of visible JPEG artifacts is saving a JPEG file at a quality setting that is too low. The JPEG file format uses lossy compression. File size is reduced by throwing away some image data. At the highest-quality settings, the data lost are often

FIGURE 5.11A and B Two small sections from a larger image. A clearly shows JPEG artifacts on sharp edges. Avoid this at all costs!

FIGURE 5.11A and B Two small sections from a larger image. A clearly shows JPEG artifacts on sharp edges. Avoid this at all costs!

almost invisible; but as the quality settings are reduced, more data are discarded in the quest for a smaller file size.

Unfortunately, the effects quickly become apparent to image inspectors and, worse still, the compression artifacts increase each time a JPEG file is saved. This is one critical reason why it is worth converting a JPEG format file from your digital camera to a TIFF format, preferably a 16-bit TIFF format, while you are editing the image data. Only convert back to JPEG when you have finished editing your file and are ready to submit it to your microstock library.

Some noise-reduction programs can be set to reduce JPEG artifacts, including the built-in noise-reduction filter in Photoshop (mentioned earlier). However, I have found they rarely work well. Avoidance is better than any cure.

In summary, my advice to avoid or minimize artifacts is the following:

• Use raw format whenever available.

• Use the highest in-camera JPEG settings.

• Convert JPEGs to TIFFs (16 bit preferred) and only resave as JPEGs once you are done editing your image and ready to send to the microstock library.

• Don't repeat save JPEGs as this causes cumulative image degradation.

CHROMATIC ABERRATIONS

Called CA for short, chromatic aberrations are color fringing on edges in images caused by lenses not focusing different wavelengths accurately on the sensor. The microstocks are sharp on too much CA, so check your pictures, particularly the edges of the frames, for this problem.

Take a look at Figure 5.12A and the cropped section indicated in red. Figure 5.12B illustrates a cropped section showing fairly bad edge-of-frame CA; notice the color fringing at the areas of high contrast. Figure 5.12C shows the corrected version.

If you shoot raw, then your raw file converter may include the facility to remove or reduce CA before decoding the file. The built-in raw converter with Photoshop CS3 has just such a facility. If not, or if you shot the photo in JPEG, then you can still remove CA by judicious application of Photoshop's Lens Correction filter (Filter > Distort > Lens Correction). I used the built-in CA-removal tool with my standard raw converter, Silkypix, on Figure 5.12B.

MOIRÉ AND COLOR NOISE

Moiré is similar to the color aliasing normally visible on fine detail as the limits of the sensor's resolution are reached. It manifests itself as

FIGURE 5.12A-C A illustrates buildings and cutouts. B and C show before and after shots of chromatic aberration.

FIGURE 5.12A-C A illustrates buildings and cutouts. B and C show before and after shots of chromatic aberration.

a rainbow pattern effect and can be quite difficult to deal with. But deal with it you must before submission.

Most digital cameras are fitted with an antialiasing (AA) filter to reduce moiré effects. This works by slightly softening the image. The downside is a slight loss of resolution. A few cameras take a chance on moiré and do not use an AA filter. This results in a sharper image, but it leaves you spending more time in postprocessing removing not just moiré but other forms of color noise.

Figure 5.13 illustrates a small part of an image opened up in Photoshop with pretty bad moiré. Figure 5.14 shows the image being worked on to remove the moiré, using a simple solution that is usually good for the job. Here is what you can do:

• In Photoshop, select the Paint Brush and then from the Mode menu, choose Color.

• Choose the eyedropper tool, and select a part of the image that is closest to the average color. In this case, it's a bronzy pink shade.

FIGURE 5.13 Moiré sample.

FIGURE 5.13 Moiré sample.

• Then, using a soft brush (see the white circle in Figure 5.14) of reasonable size, gently paint over the affected area. If the color looks wrong, go back and fine-tune it until it looks right. With a bit of patience and practice, you will get the hang of it.

What you are doing is coloring the image with the average local color using the paint brush. In the process, you are removing (coloring) the moiré color pattern. Unfortunately, the actual pattern itself remains, but it is usually not noticeable if the color is removed. The effect is nondestructive for other parts of the image. If you work on a layer, you can always erase through any areas where you have overdone it. This method can also be used to deal with a variety of other color anomalies.

Of course, there are software tools that can help. Adobe Camera Raw includes a color noise slider that reduces color artifacts, including moiré, but it affects the entire image. Other raw converters have similar tools, and these are fine for minor cases, but when the chips are down, I prefer my method.

FIGURE 5.14 The Photoshop Paint Brush in color mode removing moiré.

FIGURE 5.14 The Photoshop Paint Brush in color mode removing moiré.

DUST BLOBS

You have spent a lot of money on your digital camera, so it can come as a nasty surprise to see how easily dust blobs make an unwanted appearance in your images. Being electrically charged devices, camera sensors tend to attract small particles of dust that stick to the surface of the glass filter in front of the sensor. The problem is much more acute with interchangeable lens digital single-lens reflex (SLR) cameras because dust enters the camera body each time you change lenses. But any digital camera can potentially be affected, as can scans of film.

As with other image defects, dust blobs, hairs, or other unwanted horrors need to be removed; otherwise, your images will be rejected. There are two great Photoshop tools that can help you achieve image perfection: the Spot Healing Brush tool and the clone stamp.

SPOT HEALING BRUSH

The Spot Healing Brush works by sampling the area surrounding the brush and then replacing the area selected with the sampled data. It is very effective for blobs and marks in areas of even color, such as a blue sky, but less effective in areas where there are more complex patterns. Figure 5.15 shows what dust blobs might look like against a blue sky.

Choose the Spot Healing Brush from the Tool menu, and using a brush size just a little larger than the defect you want to remove, brush over the dirt. For large areas, try using the Patch tool from the same menu.

CLONE STAMP

The Spot Healing Brush is a relatively new addition to the Photoshop armory. The clone stamp has been around longer, but is almost as useful. The difference is that you select the area from which the sample is taken, then paint over—or clone in—the selected data to remove the visible dirt blob. A disadvantage of the clone stamp is that it does not try to match the color of the area being replaced as the Spot Healing Brush does, so with sky, for example, you need to make sure you select an area very close to the dirt mark being removed or you will see a difference in the color. An advantage is that it copes better, with practice, in areas of more complex detail, where you have more control over what data you use as your clone sample.

There are some cameras now being sold that have built-in dust-reduction systems. They use high-frequency vibrations to shake dust off the camera sensor, and they work quite well. If you do have dirt

FIGURE 5.15 Dust blobs against a blue sky.

FIGURE 5.15 Dust blobs against a blue sky.

or blobs on your sensor, you can return the camera for service and cleaning or try one of the do-it-yourself solutions now available on the market. Some products use chemical solutions to "wash" the sensor clean; others used static charged brushes to gently lift off dust particles. Whatever solution you use (I use a static charged brush), take great care to follow the manufacturer's instructions as it is easy to permanently damage your camera or its sensor.

REMOVING COPYRIGHT SYMBOLS AND LOGOS

Copyright symbols and brand names that you might hardly notice walking down the street suddenly seem to be everywhere when you are out shooting for stock, and the last place you need them is in your image. The clone stamp tool can be used to remove unwanted parts of your image, such as roof-mounted aerials and copyright logos or names that, if left in place, could lead to your image being rejected by the microstocks. With some libraries, even street signs can get you in trouble. Remember this formula: logo = microstock rejection.

You need to check your photos for copyright information, brands, names, and the like. Look carefully. They lurk in many a dark corner! If you find any, ruthlessly expunge them using whatever tool you are most comfortable with; my preference is the clone stamp tool. While you are at it, zap any visible faces for which you don't have a model release.

You will take care not to include copyright material or faces in your microstock shots, but don't be too concerned as it is really quite a simple process to remove the offending material in postprocessing.

Check out Figure 5.16. The main part of this composite example shot is bounded in red and is the final image after the road sign was cloned out using Photoshop's clone stamp tool. The area on the right within

FIGURE 5.16 A road sign in this image the blue line shows how the road sign was originally visible, and the was cloned out using Photoshop's clone stamp tool. © Douglas Freer

area bounded in a green line is an enlargement of the offending area.

SHARPENING YOUR IMAGES

Your microstock images need be sharp. Are yours? If not, they are sure to be rejected.

Sharpening your images is not a get-out-of-jail for poor focus. There is little point in trying to salvage a poorly focused image by using a lot of sharpening. It won't work. You must start with an image that is properly focused, at the point where it should be focused. With a portrait shot, you will normally focus on the eyes (not the end of the nose, as can happen!)—so with portraits, start by checking the eyes are sharp by examining them at 100% magnification on your computer monitor. Also, check for any camera shake, which can be caused by using a shutter speed that is too slow for the available light and the focal length of the lens.

But even properly focused shake-free images may not look quite sharp enough to microstock image inspectors.

A number of factors affect image sharpness such as the following:

• Lens quality. Not all camera lenses are equally sharp. Lenses that cover a wide range of focal lengths, such as some consumer-grade superzooms sold in kits with digital SLRs and some powerful compact camera lenses are "softer" than their less ambitious (and usually more expensive) counterparts—particularly when used near their maximum aperture.

• Default in-camera sharpening. If you shoot in JPEG format, the camera wil apply some level of sharpening. This can normally be adjusted by the user.

• Dirty or greasy marks on the lens. Get cleaning!

• Poor focus or camera shake. No real solution, so please avoid at all costs.

Assuming you are working with an image without a terminal problem such as camera shake (unless used intentionally as a special effect) or poor focus, then you should apply some sharpening to the JPEG file as a last step in your post-production process and before submission of the image to the microstocks.

My workflow involves turning sharpening down low in-camera and (when shooting RAW file format) in my raw decoding software. Sharpening too early in the process can emphasize noise or artifacts that I generally want to deal with first before I apply any significant amount of sharpening.

Knowing how much sharpening to apply is tricky. If your file looks sharp on-screen at 100% magnification, then you will only

FIGURE 5.17 Dialogue box from Photoshop CS3 Unsharp Mask. The original image, from the cover of this book, is already pretty sharp, so not much extra sharpening is required. © Douglas Freer

need a little sharpening tweak. In Photoshop, use Unsharp Mask (Filter > Sharpen > Unsharp Mask) at a setting around radius 0.6 to 1.2, amount 50-175%, and a threshold of 1. Figure 5.17 shows the Unsharp Mask filter in action on the cover photo. Play around with the settings to achieve the best result.

If you have to apply more sharpening than the maximum suggested, then the image might simply be too unsharp to salvage.

If you only want to sharpen a selected part of your image, not the whole of it (to avoid emphasizing noise in shadow areas, for example) then you can use the same techniques I suggested above when using noise-reduction software so that you only apply sharpening where it is needed. Do be careful not to overdo sharpening. Too much can be as bad as none at all.

There are a number of specialized sharpening programs available to purchase, some of which operate as plug-ins or actions within Photoshop or other image editors. Links to examples are included in the appendices.

Note that some traditional non-microstock libraries, like Alamy, warn against applying any sharpening at all. Each library has its own quirks. My experience with the microstocks has shown that careful sharpening, but not too much, is desirable.

With your well-composed images now free of noise, dust blobs, and copyright logos and properly sharpened, you are ready to start down the path toward microstock nirvana. But hold on a moment, that path is taking you past a camera store, and wouldn't a nice new piece of kit help you take better photographs? Read on; Chapter 6 has it covered.

This page intentionally left blank

Photoshop Secrets

Photoshop Secrets

Are You Frustrated Because Your Graphics Are Not Looking Professional? Have You Been Slaving Over Your Projects, But Find Yourself Not Getting What You Want From Your Generic Graphic Software? Well, youre about to learn some of the secrets and tips to enhance your images, photos and other projects that you are trying to create and make look professional.

Get My Free Ebook


Post a comment