All posts by Jeremy Schultz

Removing Backgrounds, Part 1: The Background Eraser Tool

This first of a four-part series covers how to remove backgrounds using a seldom-used variant of the Eraser tool

fig1Figure 1: The Background Eraser icon.

It’s buried in the toolbar and I don’t see many users working with it, but the Background Eraser tool is an interesting little piece of Photoshop engineering that can help you out when you’re caught with certain kinds of backgrounds that other tools can’t touch (or can’t touch without mangling the foreground object you’re trying to preserve!). This article is the first in a four-part series that will cover four tools that you may not be aware of and will help you solve some tough situations.

When to use the Background Eraser

The Background Eraser tool (see Figure 1 for its icon) works by Option/Alt-clicking to sample the color that is at its center (or, hotspot) and then deletes all pixels of that color or close to that color that lie within the eraser’s boundaries. This makes it a great tool to use for erasing backgrounds that are a different color than the foreground object, even when the background may have dark and light elements like clouds or other things that the Magic Wand tool or other value-sensitive tools would leave out. This is a great thing to remember when a background has some stray pixels that you can’t grab in one pass with the Magic Wand tool: use the Background Eraser and remove them all at once!

Use the Background Eraser to remove backgrounds that are a different color than your foreground object!

The downside to using color as your basis for pixel removal, as all you colorists already know, is that in the real world color is getting thrown all over the place. Shadows and highlights on an object often carry some of the color from the light or the reflected light coming off a colored object, and Photoshop is not smart enough to know what color to keep and what to delete. That is the Background Eraser’s downfall, and on some images you may find using this tool requires some precision in order to preserve some highlights or cast colors that need to remain on the foreground object.

fig2Figure 2: Subject and background, separated by color.

Figure 2 is a good candidate for the Background Eraser tool. The blue sky and red lighthouse are very different in hue so it should work well. Problem spots will be the door, glass and metalwork at the top of the lighthouse as well as the white window frames on the sides, which have a color cast created by the blue of the sky. As we erase close to the horizon we’ll also encounter the ocean in the distance and the tree, fence and grass to the left, all of which has blue elements and will make the job tougher.

The options

fig3Figure 3: The Background Eraser tool’s Options settings.

Figure 3 shows the tool settings in the Options bar (I assembled them vertically for this visual aid):

  • The Brush pop-up allows you to change size, hardness and other features of the eraser’s shape
  • The Sampling buttons are new to Photoshop CS2, which in CS and earlier versions were kept in a simple drop-down menu. These buttons control how the Background Eraser tool samples the color to be erased:
  • Continuous (the icon with a moving eyedropper) will sample a new color every time the cursor is moved
  • Once (the icon with a target and eyedropper) will sample color with the first Option/Alt-click and will continue to delete pixels of that color until a new color is sampled
  • Background Swatch (the icon with the foreground/background swatches) will enable the Background Eraser tool to delete pixels containing the same color as the background swatch
  • The Limits drop-down menu controls what pixels the Background Eraser tool can edit at any given moment (see Figure 4):
  • Discontiguous edits any and every pixel of that color that fall within the eraser’s boundaries
  • Contiguous edits only pixels of that color that are in the region sampled by the hotspot
  • Find Edges edits exactly like Contiguous except it also preserves the sharpness of shape edges and can create some strange shapes if it comes across hard-edged elements or pixels that are a bit darker or lighter than
  • Tolerance, the tool’s most important setting, controls how sensitive the eraser will be when deleting pixels
  • The Protect Foreground Color checkbox allows you to preserve pixels that are the same color as the foreground color swatch

fig4Figure 4: The effects of the various Limits settings.

The two options you will use most often will probably be the Tolerance and Limits options, as they control how sensitive the eraser is and which pixels you want to edit within the image. I usually keep my Sampling set for Continuous, my Limits set for Discontiguous and my Tolerance set anywhere from 25% to 40% or 50%, depending on the image at hand.

To the lighthouse

My task with the lighthouse image in Figure 2 is to delete the sky background and insert another one later. This image is flattened (meaning there’s only one layer, the Background) but the Background Eraser tool erases to transparency (there’s one other tool that does this, can you name it?) so once I start working with the tool my Background will convert to a regular transparent layer. I’m going to save Photoshop the work and do it myself, and set up a little visual aid to help me work:

Step 1: Double-click the Background in the Layers palette and then click OK in the New Layer dialog box in order to convert the Background into a transparent layer named “Layer 0″.

Step 2: Hold the Command key (PC users use Ctrl) and click the New Layer button on the bottom of the Layers palette to create a new layer below the image layer.

Step 3: Fill the new layer with a vibrant or bright color so when you erase the image you will have visual feedback of what you are deleting. I am using a bright yellow in this example.

These first three steps are optional, but I like to use them because the usual checkerboard pattern to show transparency sometimes doesn’t show up well and you can’t tell exactly what pixels are being deleted. Now for the actual erasing:

Step 4: Set the Background Eraser tool with a 800-pixel hard-edged brush, Discontiguous limits and Tolerance of 35%.

Step 5: Click on the blue sky and drag as the Background Eraser tool removes the blue sky and preserves the red lighthouse.

Some tips:

  • Areas with color variations, like the clouds, will be tougher than the flat sky. Click on these stubborn areas to erase them and others like them in the same brush area.
  • Lower your brush size to attack small or hard-to-reach areas. To erase the area by the fence and tree at lower right, I had to shrink my brush down to 200 pixels or I would have erased the blue in the window frame.
  • Leftover pixels will show up as faint discolorations or “dust” on your background color. In my case, I could see faint blue and green discoloration on my yellow color. A few more passes with the Background Eraser tool will clean it up. A higher Tolerance at the outset will help too.
  • The ocean at the lower right ended up being too similar to the sky to make the job easy, so I used the Marquee tool to select the ocean, then selected Select -> Inverse to select everything except the ocean, then erased as usual.
  • You’ll notice some of the yellow is showing up on the glass panes beside the lighthouse lantern. This is the reflected blue light coming off the glass; my eraser brush was large enough to catch it as well as the blue sky between the metalwork.
  • After I have erased all the pixels that I can see, I like to zoom in, set my Background Eraser tool at a high Tolerance such as 80%, then shake the mouse around any areas that look like they may have stray pixels. Fast, jerky movements on the problem areas ensure that I catch as many pixels as I can while my Sampling is set for Continuous.

A new day is dawning

fig5Figure 5: The Background Eraser tool does its job!

Figure 5 is the result, a red lighthouse with a bright yellow background! And of course from here what you do with your image is up to you. In my case I have a nice photograph of clouds over the sea (Figure 6), and in Figure 7 I made it my new background by simply dragging the image onto my edited lighthouse image, then moving the clouds layer behind the lighthouse layer. And because the Background Eraser tool knocked out some of the reflected light on the glass, the color of the new sky shows up in its place and automatically creates a convincing reflected color that matches the sky. Easy!

fig6Figure 6: My new background.

fig7Figure 7: Incorporating this new background is a breeze!

About next time

On Sunday, November 13 I will publish Part 2 of my series on removing backgrounds. We will feature another tool, I won’t tell you what it is here but it is the only other tool in Photoshop that by default erases an image to transparency rather than white (I mentioned this earlier in this article). It’s another tool that is relatively obscure but once you start using you’ll wonder how you got along without it!

SERIES INDEX

Part 1: The Background Eraser Tool
The Background Eraser tool removes backgrounds according to color, which makes it useful for images where the foreground object and background object have different color hues.

Part 2: The Magic Eraser Tool
The Magic Eraser tool is a combination of the Eraser and Magic Wand tools, which makes it valuable in situations where backgrounds have a uniform color and value.

Part 3: The Magnetic Lasso Tool
The Magnetic Lasso tool seeks edges with high contrast, making it a tool that can snap to a foreground object’s edges quickly in order to remove the background.

Part 4: The Extract Filter
The Extract filter is a complex tool for eliminating the toughest backgrounds, including those dealing with hair, fur, leaves and blurred edges.

Photoshop TV Reviewed: Good But Not Great

The mix of two parts goofy humor to one part Photoshop bothered me a bit

I just got through watching the first two episodes of NAPP’s Photoshop TV (click here for the main Web site or click here to subscribe via iTunes). The comments on the Photoshop TV Web site are all exceedingly positive, but I have to say that I was underwhelmed. Each episode is 30 minutes long and in that time you’d think they would have at least five or six good tutorials and a chunk of news, but there’s usually only three tips and some news (not necessarily in-depth) and the rest of the time is spent joshing around, telling goofy jokes, forgetting lines or spying on the slightly annoyed NAPP staff with Dave Cross’ Dave Cam. Now I know that goofy humor is part of what makes Scott Kelby, his books and NAPP popular, but there can be such a thing as too much.

I have to give credit where it’s due though: Photoshop TV is big now and it will only get bigger, as iPods continue to saturate the market and users use them more and more for recorded shows like these as well as than music. The show is good now, so any Photoshop user who’s even remotely interested should subscribe and check it out. And Scott And Friends made some improvements in the second episode, getting rid of their original decor (the first episode looked like it was shot in Scott’s living room; the second, in NAPP’s studios) and tweaking some other things for the better.

But I think a lot more could be done:

Use two hosts (or even one) instead of three. Too often Matt and Dave are just standing around while Scott is talking, and don’t have much to do except smile and laugh at the jokes. It can be painful to watch. Or if cutting someone out isn’t an option, at least don’t keep everyone in frame all the time: bring Matt or Dave in to share their tips, but leave Scott as the grand impresario!

Remember the info. There were a few times in each episode when Scott, Dave or Matt simply forgot an important date or tidbit of information, such as the date for Bert Monroy and Ben Willmore’s next events. Use some cue cards or something to get that information down pat.

More tips! Thirty minutes is a lot of time to kill for a podcast, and for me two, three or even four tips is not enough. Let’s see some more! I am a fan of Scott’s Down & Dirty Tricks books mostly because every page has several great tips and tutorials. Compared to those, Photoshop TV has a lot of fluff. I think Photoshop TV can be humorous and fun and still have six or seven excellent tips!

I’m going to be holding onto my subscription to Photoshop TV for a long time. I feel it’s good but not great at this point, but NAPP is working hard on it so I’m excited to see how future episode evolve!

The Nikon D200 Is Coming

The camera made waves by accident, now will it make waves for real?

I just heard this on the October 31 episode of Photoshop TV: Nikon had apparently posted information on its brand-new D200 camera by accident, quickly pulled the page but not before the information was able to spread around the world. Doesn’t it seem the Internet always turns little mistakes like these into global rumor grist?

Well, I just visited the Nikon USA Web site and it looks like the D200 is almost out of the gate: click here for the official page.

The D200 has 10.2 megapixels and the body is $1,700, so look for it to be a popular camera for professionals and high-end prosumers. Look for a release date sometime next month.

Photoshop Radio Becomes Photoshop TV

Just in time for the video iPod, NAPP takes its show to the (little) big screen

You may no’t find it by actually visiting the National Association of Photoshop Professionals’ Web site (the “Photoshop TV” icon goes to the wrong domain!), but if you visit photoshopguys.com you’ll find Photoshop TV, which replaces Photoshop Radio as NAPP’s vehicle for bringing Photoshop tips and info to iPod users.

It looks like a pretty cool weekly videocast (it comes out on Mondays), and the hosts are none other than Photoshop veterans Scott Kelby, Dave Cross and Matt Kloskowski so you know the info is top-notch. The first episode featured tips on sharpening and special effects, Apple Aperture and some Photoshop trivia to top it off.

It’s not without its growing pains, however: the files are huge! I opted to download the MP4 files straight to my desktop and the files are 106MB and 140MB respectively. That’s a lot of waitin’. I’m still waitin’, so I haven’t been able to actually watch an episode straight through. I’m going to post again once I have, but until then I wanted to get the word out.

Response from users has been pretty positive so I think Photoshop TV is here to stay!

ZOMBIES! or, How To Add Some Color To The Living Dead

Thanks to amateur digital photographers, we see dead people.

Digital photography has changed the world of Photoshop and photography in amazing ways, with designers, sales reps and even clients now able to grab their digital camera and take a product shot or portrait shot with the resolution to print in a fine magazine. Who needs professional photographers, now that a $500 camera can capture enough pixels for just about any computer application, which itself is only a few hundred dollars off the shelf, and that digital image can be used practically anywhere? Who needs professional photographers? Some might argue that we ALL are professional photographers.

The difference, though, is that professional photographers have years of training and experience. Now that photography has become accessible to all, us print professionals are seeing a glut of bad photographs coming across our desks and into our inboxes. Low resolution, bad lighting, cell phone photos—we’ve seen it all. But for this Halloween I will focus on one bad photograph in particular, seen in Figure 1.

fig1Figure 1: What color is his skin?…neutral gray?!

The scary part? The fleshtone reads as 207R/207G/207B—neutral gray. How can this be? Nothing, not even a color cast? This poor man, in the prime of life, forced to walk the world (and be printed in a society magazine) looking like the undead? Not if I could help it, and I was determined to bring this man back to life, Frankenstein-style.

This is an exercise in color correction, and it’s a simple three-part procedure of setting the shadow, highlight and fleshtone colors. If you want more information on these and other color correction techniques, pick up Dan Margulis’ Professional Photoshop.

Step 1: Use the Eyedropper tool to pinpoint the shadow, highlight and average fleshtone in the image. See Figure 2 to see where I set these points.

fig2Figure 2: My eyedropper points.

Step 2: Make sure the Info palette is visible, and you can see the color information for all three points. The Info palette will display each point’s color information at the bottom of the palette; I’ve included Figure 3 below to illustrate.

fig3Figure 3: The Info palette.

Step 3: Open the Curves dialog box (Cmd-M/Ctrl-M) and press Cmd-1/Ctrl-1 or select Red in the Channel drop-down menu. Then drag the shadow point on the curve either left or down until the red color number for the shadow point reads 15.

Step 4: Select Green (Cmd-2/Ctrl-2) and do the same with that shadow point, then select Blue (Cmd-3/Ctrl-3) and do the same. When you are done the color numbers for the shadow point should read 15R/15G/15B.

Step 5: Repeat steps 2 and 3, but this time drag the highlight point either up or right, and the goal is to attain color numbers of 245R/245G/245B for the highlight point. Click OK to exit the Curves dialog box. See Figure 4 for the image with updated shadow and highlight points.

fig4Figure 4: Just a couple changes, but the man is beginning to revive….

The reason for steps 2 through 4 is to remove any color cast. If the image’s shadow and highlight are neutral (and in RGB, neutral colors have equal color numbers) then a color cast cannot exist.

Our test subject is starting to show signs of life due to the radical change in highlight point. The man’s white undershirt was baby blue when it should have been white, and fixing that also zapped some of the blues in the face, leaving behind reds and yellows. Now we’ll focus on the fleshtone point by converting to CMYK and making its color numbers a formula fleshtone color.

Step 6: Convert the image to CMYK color (Image -> Mode -> CMYK Color).

After the CMYK conversion, the fleshtone point reads as 0C/14M/20Y/0K. The black value is okay, but there should be a few points of cyan and maybe 30-40 points of yellow and magenta, with more yellow than magenta.

fig5Figure 5: My curves for Step 7.

Step 7: Repeat step 3, but now you will work with the cyan, magenta and yellow channels separately. Raise the curve where each color number in the fleshtone lies. Look at Figure 5, which is my curves for cyan, magenta and yellow. You’ll see how I took each color number for the fleshtone and raised it to satisfy my fleshtone formula. You will be able to see your changes to the image if you have the Preview box checked in the Curves dialog box, and if the changes don’t look quite right you can tweak them as necessary.

fig6Figure 6: It’s alive! It’s alive!

Figure 6 is the final image. This image still can benefit from some further color correction (working in LAB mode would be very helpful here) and sharpening, but our goal was to bring this image back from the dead and with some curves corrections we did just that. Curves is a great way to correct a range of tones and introduce nice color where absolutely none exists.

SPOOKY! Building a Halloween Typeface in Photoshop

Here’s a fun Halloween typeface you can build right in Photoshop in five easy steps!

Step 1: Type something up in your favorite spooky font and kern it as needed. I used Manson in Figure 1.

fig1Figure 1: The original type layer.

Step 2: Duplicate the type layer by dragging it to the New Layer button at the bottom of the Layers palette. Make the original type layer invisible by clicking on the eye icon next to the layer.

Step 3: Select your duplicate layer and select Filter -> Distort -> ZigZag. You’ll be asked if rasterizing the type is okay, which it is, and then you’ll see the ZigZag dialog box. There are many ways you can distort your type with this filter, and I suggest you play around with the settings as there’s some really cool tricks you can do here. In this case I used the settings of Figure 2.

fig2Figure 2: Settings for the first of two ZigZag filter effects.

Step 4: Repeat the filter if you want more modulation. I made a second pass with the settings in Figure 3.

fig3Figure 3: The second ZigZag filter effect.

Step 5: Turn on the visibility of your original type layer to achieve Figure 4, a very spooky typeface for Halloween! Depending on the typeface and the filter effect, you may want to kern your original type layer again to match the filtered text (see Figure 5).

fig4Figure 4: The two layers combined! It’s alive!

fig5Figure 5: The final type, after some kerning.

Dan Margulis’ Photoshop LAB Color

Dan Margulis’ new book focuses on Photoshop’s nuclear color space. But who is it written for—the novice or the expert?

lab color book
A few weeks ago I wrote a very favorable review for Ben Willmore’e Up To Speed: Photoshop CS2, and one thing I really liked about the book was its specialized focus and in-depth coverage. Too many Photoshop books cover every teensy-weensy feature of the application and each with only a passing comment or paragraph. Now Dan Margulis, whose “Professional Photoshop” book has been the bible of color-correction for many years, has put out “Photoshop LAB Color: The Canyon Conundrum and Other Adventures in the Most Powerful Colorspace”. It sounds like a doctoral dissertation, but it’s actually a substantial book devoted to a single obscure menu item that almost no one understands or even pronounces correctly. However, as Dan spends much of the book trying to prove, LAB is a valuable “nuclear option” for the arsenal of any Photoshop professional.

LAB: A Primer

You’ll find “Lab Color” in the Image->Mode menu. Don’t let Photoshop’s mixed-case “Lab” fool you: according to Dan’s book, it should be in all caps and pronounced as letters, like CMYK or RGB. LAB is a color space that is unique in Photoshop in that the L channel carries all contrast and value information, while the A and B channels compose the color (A controls green and magenta, B controls blue and yellow). The LAB gamut is much, much larger than CMYK or RGB and can create colors not possible in our universe. But, as Dan’s book teaches us again and again, in retouching it’s sometimes these unreal colors that end up achieving the best final results (and, as “Professional Photoshop” readers will attest, Dan Margulis is nothing if not insistent on creating the very best images). LAB is the best color space for retouching (Chapter 11 of the LAB book is simply titled “The Best Retouching Space”) because of its ability to separate detail from color and make weak color into vibrant color.

A book written for both the novice and expert

Dan made it clear to members of the Applied_Color_Theory user group that a major challenge with “Photoshop LAB Color” was in writing a book so novice users could grasp the basics of LAB, while expert users would out of the same book gain bleeding-edge knowledge and techniques that make normal users feel like they’re in fourth-year astrophysics. LAB is such a foreign and advanced topic that it presents its own conundrum: cover just the basics, and you bore the advanced users; cover every detail and point of theory, and you confuse the novice users. Dan’s solution was to split each chapter into two sections, each one with its own typeface and approach. The first half of every chapter is a gentle walkthrough of LAB techniques and some explanation of how LAB can be used in real-world situations, and the half ends with a short review and exercises. Then the second half begins, and “Professional Photoshop” veterans will probably say this is where “Photoshop LAB Color” starts to feel like a Dan Margulis book, with exposition on color theory and color theory history, lots of color numbers and image-by-image comparisons that seek to create a LAB expert in each of us.

Dan has been writing long enough to make both sections work well, though inevitably I think everyone who reads this book will gravitate to the sections they understand and perhaps see the other sections as something of a waste. I enjoyed the advanced sections but only skimmed the novice sections. And there will be other readers who are intrigued by the novice sections but end up discouraged and confused by the expert sections. This book could almost have been split into two publications, one for beginners and one for experts.

The design: Jam-packed

This book is full of large images, large type with large x-heights and large columns. Dan even modified the typeface, as he does in some of his other publications, in order to increase the type size; read the last pages of his books to learn just how he modifies the typefaces he uses. But with all those elements of largesse, I find these and other Dan Margulis books to be tough to read. White space is nowhere. Big blocks of text and images are everywhere. I know Dan believes in large images and type, and I agree with him and especially with the large images part, because it’s so important to see what exactly is going on in those examples. But the book is so packed with images and text that it can be overwhelming and difficult to read, not because the material is tough to grasp but because it takes a concentrated effort just to slog through the design of it all!

Dan has publicly grappled with publishers before over the design-versus-size issue. Photoshop User magazine, the official publication of NAPP, runs his Makeready columns but, according to one of Dan’s posts on the Applied Color Theory group, “insists on running excessive white space and declines to accommodate the larger images I prefer to use.” For that reason Dan considers Electronic Publishing, which also runs that column, to be the official source of Makeready. Maybe Photoshop User is on to something in regards to design. From this designer’s standpoint, the layouts could use some fresh air.

Magic color

Despite questions about the layout and the organization of the book, I think “Photoshop LAB Color” is a great Photoshop book and essential for anyone who does color-correction. “Professional Photoshop” gives substantial space to LAB, and Dan has seminars at Photoshop World and others that focus solely on LAB, but this is the only book devoted to LAB that I know of and it’s quite exhaustive. Dan himself said that some of the techniques in these pages were developed and studied in the course of writing this book, which means they are new tricks that we are the first to employ. Some techniques that I got the most out of were:

  • Building masks with LAB
  • Eliminating moire
  • Changing or matching product colors
  • Blending the A or B channel into L to modify colors without selecting
  • Improving portraits and fleshtones, covered in detail in the last chapter

There’s plenty more, but these were the ones that I was drawn to the most. They are real-world techniques a Photoshop user will go to again and again. And despite the split in chapters between hand-holding for the novices and rocket science for the experts, there is more than enough good writing and examples to teach anyone how to harness the power of LAB. This is a book for experts and novices alike, and everyone will benefit.

Conclusion

There were doubts before the publication of “Photoshop LAB Color” that it would be successful: the thinking was that the subject matter was too narrow, too specialized, and/or too advanced to write a book that would sell well. Thanks to positive feedback from the Applied Color Theory group, Dan did indeed write the book and both he and his publisher were surprised when word came a few days after the unofficial release date that it had already taken the top spot on Amazon.com’s bestsellers list for computer and Internet books (the official publication date wasn’t until Photoshop World, a few days later). I’m not surprised by the book’s success: the Photoshop book market is starved for well-written books for the advanced user, focusing on specialized topics. For Photoshop users who have already learned much of what the application can do, finding a book like “Photoshop LAB Color” is like finding an oasis in the usual bookstore Sahara. The water may not be crystal-clear, but it is good to drink.

Space Monkey and Other Photoshop Easter Eggs!

Photoshop has a long history of Easter eggs and other oddities

Hold down the Command key (PC: Ctrl), select Photoshop->About Photoshop… and you’ll see this:

space monkey

It’s the latest Photoshop Easter egg, those goofy little jokes software developers work into their products just for fun! Photoshop and ImageReady has had a long history of Easter eggs, and you can learn how to uncover some more in the earlier versions by clicking here.

If you have other Easter eggs for Photoshop CS2 and other CS2 applications, drop us a line and we’ll post them here!

Tutorial: Manipulating Grayscale Images for Better Bitmaps

What do you do when you need a bitmap image but only have a grayscale image, possibly with midtones? Here are three techniques to get you good line art

My recent tutorial on eliminating jaggies from bitmap images was responded to by Mac Wizard,

Rather than trying to convert back to bitmap, you could add an adjustment layer of curves and pull the diagonal line vertical, black point to 125 and white point to 125. By adjusting the horozontal position of the vertical line (does that make sense?) you can adjust how much or little the white fills in. ie. Black and white point at 20, or 200.

It’s a great technique for making a grayscale image more like a bitmap, whether for a bitmap conversion later or just to clean up your grayscale image. I promised Mac Wizard I would present a tutorial on grayscale images, and here it is. I am going to cover three different techniques that anyone can use to drop out gray values and end up with black and white images that are very suitable for a bitmap conversion.

Curves

fig1Figure 1: The Calligaris logo, our specimen for this tutorial.

This technique was shared by Mac Wizard, and here we’ll use it on Figure 1, the Calligaris logo, a real-world example that I had to clean up after the client could not produce a specimen of better quality.

Step 1: Open the Curves dialog box. Keyboard command is Cmd-M (Mac) or Ctrl-M (PC).

Step 2: Click-drag the Eyedropper tool over a the dark and light areas and note where they lie upon the curve. As you click-drag you will see a circle along the curve line where those values lie (see Figure 2). This is an optional step; you can experiment with steps 3 and 4 if you like until you find a final image that looks best to you.

fig2Figure 2: The Curves dialog box, testing the logo with the Eyedropper tool.

Step 3: Drag the black point of the curve horizontally until it is beyond the range of dark values.

Step 4: Drag the white point of the curve horizontally until it is beyond the range of light values.

fig3Figure 3: The technique at work.

At this point you will have a steep straight line in the Curves dialog box, with previously dark values falling beyond the black point and light values falling now beyond the white point. See Figure 3 for an illustration of the technique at work on the Calligaris logo. If you see specks of white in the black areas (or black specks in the white areas), then a few pixels were not in the original range from Step 2 and can be cleaned up by dragging the black and white points farther in and steepening the line (see Figure 4). The line can go so steep as to be vertical, with the black and white point having the same value on the grid; this is what Mac Wizard mentioned in his post. Tip: You can still use the Eyedropper tool to sample the stray pixels and see where they lie on the curve line. Even though they appear black or white, that is the Curves preview at work and using the Eyedropper will reveal their original values.

fig4Figure 4: A second pass using Curves.

fig5Figure 5: The new Curves and logo.

At this point the grayscale image (Figure 5) is now clean and, after eliminating jaggies it will be ready to be converted to line art.

Levels

I prefer using Levels to accomplish the same thing Curves done above, because the Levels histogram makes for an easily-read visual interface that makes this job a cinch. If you are a visual learner or find it easier to work with things you can see and manipulate, then this technique may be better for you.

We’ll use the same Calligaris example, Figure 1 above.

Step 1: Open the Levels dialog box. Keyboard command is Cmd-L (Mac) or Ctrl-L (PC).

You’ll see two large peaks with a deep valley between them; see Figure 6. Images with a large dark region and white region will always produce a histogram like this.

fig6Figure 6: The Levels dialog box for the Calligaris logo.

Step 2: Drag the black input slider to the right until it is beyond the left peak.

Step 3: Drag the white input slider to the left until it is beyond the right peak.

fig7Figure 7: The new Levels.

fig8Figure 8: The new image.

Figure 7 is the altered Levels dialog box; Figure 8 is the resulting image. Levels changes the pixel values in exactly the same way as Curves did, but the quick visual feedback of the histogram makes it quick and easy. Levels (and Curves, for that matter) do not work as well with more difficult grayscale images such as Figure 9, which has some midtone grays as well as darks and lights and produces a Levels dialog box (Figure 10) with a smaller valley than the Calligaris logo. The technique remains the same, but it will be harder to find the right positions for the two sliders and you may need to fiddle with them as we did for the Curves technique by bringing them even closer together to eliminate any gradations between black and white.

fig9

Figure 9: A much tougher image to clean up.

fig10Figure 10: The Levels for Figure 9, which show the lack of a well-defined gap between dark and light values.

50% Threshold

The Curves and Levels techniques above are useful for preparing a grayscale image for a clean conversion to a line art bitmap using the 50% Threshold method. However, if the original image is fairly clean to begin with, with large and relatively uniform areas of dark and light values, it may not even be necessary to use Curves or Levels to bring the darks to black and lights to white. Just let the bitmap conversion do it for you!

In Figure 11 we are back to our original Calligaris logo.

fig11Figure 11: Back to our original logo in Figure 1.

Step 1: Select Image->Mode->Bitmap… and convert the grayscale image to a bitmap, making sure “50% Threshold” is selected under Method.

fig12Figure 12: The bitmap conversion of Figure 11.

Easy enough—dark values become black, light values become white (see Fig. 12). The resulting image has some stray white pixels and is not as clean than if you had used the Curves or Levels techniques to prepare it, but for some grayscale images, especially ones with uniform darks and lights, it may be easier to just execute the bitmap conversion. Note Figure 13 below, which is the Levels dialog box and histogram for Figure 12 after converting it back to a grayscale (Levels and other adjustments do not work on bitmaps). Since all values are either black or white, the histogram will show only one-pixel-wide peaks on each end, and trying to perform the Levels (or Curves) technique on this image will produce no results. Even if you drag a slider in Levels all the way to the other side of the histogram, the preview will show change but click OK and the image will show no change.

fig13Figure 13: The Levels dialog box for Figure 12, which are practically useless!

Conclusion

Mac Wizard’s original post has dug up a few useful techniques for manipulating grayscale images for a clean and easy conversion to line art. I don’t know how many times in a production environment that I’ve had to take a photocopy, fax or lousy business card logo and make into a clean piece of line art that is actually press-worthy. These are the techniques I rely on to spin straw into gold, and I’m sure there are other techniques that have been developed in the field so if you have your own formulae for bitmap conversion make a comment below and perhaps we’ll revisit this topic in a future article.