AB: All-Round Nice Guy

Here’s a quick Photoshop tutorial based on a recent project of mine - typesetting a 900 page book for a ReBusiness client, complete with a new cover design and matching web site…

The problem was that the book had to be publicised well before the printers had completed their print run - so the book had to be quickly mocked up (faked) in Photoshop for the promotional material - a fun half-day job. Here’s how I did it…

Step One - Get something close to play with.

We’re not going to spend time modelling a book in a 3D application, and play with the ambient light sources for a half a day - the best way to get going quickly is to find something similar, and use it for reference. In this case, it was easy - a previous year’s book was on hand, and my trusty Fuji S2 (a 4 year old digital SLR camera) is always with me.

We wanted the final image to be fairly dynamic - we were going to promote the fact the this year’s book would be 20% bigger than the previous year - so we got in tight to try to get the front edge appearing a little larger than the rest of the book. In the end, a better balance was to show a little more of the front cover - to let the promotional material begin to promote the new design.

The shot below is done entirely on site at the clients’ premises, under flouro lighting, with only a slight care for white balance - we’re after the reference, not looking for a Pulitzer Prize

Step Two - Draw a path around our book.

Anyone reaching for the Magic Wand tool here, or the Magnetic Lasso tool - will be summarily shot, drawn and quartered. The *only* tool of choice for ‘drawing around’ something is the Pen tool. Why? Many reasons, not the least of which are accuracy and repeatability. Once you save a path, you can reuse it time and time again. The uptake of the Pen tool is harder than the Magic Wand, but will leave everything else in the dust once conquered.

Zoom into 100% or 200%, and get a real feel for what you’re about to trace around. Some areas are indeed perfectly straight line segments, so a simple click, scroll, click will give you a perfectly straight path between those points. But take some time around any features that are, well, features. Granted, there weren’t many here - the two notable ones are the little flick of the top right corner (a creative license - it wasn’t there half as much), and also the little indent of the fold line (the score) in the top left corner - both add that tiny bit more than if you’d simply made the book from 6 quick clicks.

Once you complete the path, double-click the path in the paths palette to save it - call it whatever you like. But then select that path by clicking on the little marquee circle at the bottom of the paths palette. You’ll now have the book ’selected’.

There’s a few ways to make a mask at this point - a layer mask is more usually the common way. But in this case, I didn’t want to work with one mask on one layer - I wanted to be able to move all my layers and masks around freely - all was well - since the path had been saved, we could reselect it and reapply it again and again. So, with the selection active, invert the selection (command-shift-I) to select the ‘outside’ of the book, and remove all the old refernec ebackground by pressing ‘delete’ (backspace for PC). Our reference book was now precisely the same as our path. Make a new layer, fill it with white, and move the layers around so that the old book now sits on the white bg layer, something like this:

Step Three - Get your new artwork ready - all about file sizes.

Using InDesign for the new book cover, there is an option to export directly to a jpeg, but in this case, it was only at screen resolution - as if we’d simply taken a screen shot ourselves. Going directly to the ultimate method, I just make the cover into a print-resolution (print-quality) PDF, then opened that PDF from Photoshop and rasterised it to a decent size. I say ‘decent’, because the final image only needed to go on an A4 as a square finish (ie - 20cm x 20cm). ‘Print quality’ for an A4 page is 300dpi (more accurately, it doesn’t need to be any more than this value), although the same image size at 240dpi would have done just as well, and would have been near half the file size and CPU load.

The final rasterised dimensions of the cover had to be a little bigger than planned, simply because the PDF of the cover included the back cover as well. If we had made the new artwork 21cm wide at 300dpi, then the front cover would have been a little underdone. I made the rasterised file 40cm wide at 300dpi - plenty of size to spare, but no so large that it takes the whole day to rasterise.

Here’s the cover PDF, rasterised:

This is probably a good spot to ensure that the new artwork (the digital photo of the refernece book) is/was the right size for the job - you may think this is checking the sizings a little too late, but the camera was a 12 mega pixel camera running in 6 mega pixel mode, and the maths is very easy to figure out: a 6mp camera results in an 18mb RGB image size, or a 24mb CMYK image size. And since we only wanted our image to be about half of an A4 - and since *every* designer/photographer knows that the only file size you’ll ever need for *any* job in the whole world is 33mb, we were in the clear.

What that? You didn’t know that the only file size you’ll ever need in the entire world is 33mb? hmmm… Let me explain.

Consider a magazine cover, or a page from a fine art book. Let’s assume the page is A4: a 21cm x 30cm final physical size. (We’ll ignore bleed for now.)

Now, to the other end of the equation: what our eyes can perceive as good quality, and what we perceive as being sub-standard quality.

The human eye can resolve dots on a printed page, down to about 1000dpi. That’s pretty darn good. So to fool the eye, we have to make sure that the apparent dpi is equal to, or more than 1000dpi. Any less, and we’ll be able to see where the designer is trying to make an curved line out of a grid of horizontal and vertical dots. That is, we’ll be able to percieve the line *as* a grid of dots, and not think that the line is a solid, curved line.

For bitmap images (purely black and white), we definately need to go over this 1000dpi ‘vertical limit’, otherwise, the purely black and white dots will become noticable. A convenient value is 1600dpi, and, lo and behold, a 21cm x 30cm bitmap image in Photoshop at 1600dpi *is* 33mb.
For practical purposes, bitmaps at 1200dpi do just the same job, and are 2/3rds the file size and CPU load. Let’s assume that 1200dpi is plenty enough resolution from this point, and take the savings in file size and CPU load as a bonus.

With greyscale (or monotone) images, the defining quality is that any image has a range of possible colours, going from one colour (generally white) through to another colour (generally black!). This allows for a curved line to be drawn, and all the ’steps’ in the grid of dots are ‘filled in’ with dots of varying strength. They essentially smooth out a curve so our eye doesn’t twig to the fact that it still is a square grid of dots.
The great news is that to fool the eye with a greyscale or monotone image, you only need about 600dpi. Any more is a waste of hard drive space and CPU cycles. And, guess what? A 21cm x 30cm greyscale in Photoshop at 600dpi *is* 33mb.

Finally, onto full colour images - and wonderful news - with the dazzling array of colours at our disposal, it’s very easy to literally blindside the human eye into thinking that there’s more pixels than there really are. 300dpi seems to be just about enough to fool the human eye - the eye is simply so overwhelmed by the colour that it doesn’t really care how good the resolution is - it just looks *damn fine*. And, you guessed it, a 21cm x 30cm Photoshop image (in CMYK) at 300dpi *is* 33mb.

Well… almost.

In reality, full colour printing is usually done with via a CMYK printing process: that is, Cyan (blue), Yellow, Magenta (the red/purpley colour) and K (black). These four coloured inks (or toners) can make up a very wide range of the colours that the human eye can see - and typically all the colours you’ve ever seen in anything printed, well, ever.
Now for the maths behind it, though: if 300dpi apparently fools our eye for colour, *and* there’s four colours making up each dot, then there’s actually 1200dpi there, so it’s no wonder that it can fool the eye.

In the end, if *anything* is 33mb, that is, if any A4 image contains some 9 million dots (do the maths yourself) then our eye is completely fooled.

Q: But if you need anything bigger, isn’t that going to be even bigger?
A: No. An poster migiht be 2 metres tall, but it certianly isn’t 300dpi. An ad for the side of a bud is huuuge, but, up close, it’s about 2dpi. Open a new document in Photoshop, and play with the image size and the dpi values, and watch the file size change. The only real issue is working with A3s, or double page spreads - the file needs to be 66mb if you want the end used to be able to stick their nse right up to the final printed page.

Quickly wrappping this little tangent up - working *backwards* from the magical 33mb, that means that all you need is a digital camera that can do 8 megapixels, since that will give you a 24mb RGB, or a 32mb CMYK. Of course, you’ll want more to be able to have room to recrop your images, and you’ll want to have bleed and slug on your final printed page, but that’s the general theory. Hence my dig camera that can do 12 megapixel when required, but does 6mp a whole lot sharper.

SO… getting back to our artwork. Finally!

Ah - I’m lost, too - oh yes! Make your image the *right size* for the job: crop the book as tight as you’ll need it, then resize it to 20cm wide at 300dpi. sue 240dpi if you would have otherwise needed to resample the image up - it’ll pass the test just as well

Step Four - Placing the front cover.

Select only the front cover of the new artwork, and paste it into the new book file. It’ll become a new layer, and you can go ahead and make its opacity to 50% in the layer’s palette - that’ll help us line things up in a minute.
Go Command-T (or control-T for PCs) or go Edit/Free Transform. Do a control-click (or right-click) on one of the corners of the Free Transform box, and choose ‘distort’. Now, get to work, and line up the four corners of the trnasform box with the four corners of the reference book. Allow for a little more width than for a little less - it’ll be easier to crop the book to our saved path than to have to manufacture extra book width. Press enter to commit to the tranform when you’re happy.
Head to the path palette, select our outline path, make it a selection, then remove the outsides, just like before. Restore the opacity to 100%, if you haven’t already done so.

Step 5 - Place the spine.

Same as before: select the spine only in the new (flat) artwork, paste it into new artwork. Free transform it, (command-T or control-T) and control- or right-click the handles at the corner to be able to choose ‘distort’. With only the four corners, match it up - take your time - the opacity thing would have helped, if you didn’t do that already. Make the path into a selection and clear away any ‘overage’.
At the place where the spine meets the front cover, figure out which of the two edges is the best to go in front, and swap the layers if needed - have a care lining up any features of the new book design that need to match up, like the black horizontal lines in this case. I think I may have matched up the bottom half first, then roughly selected the top half of the spine, and nudged it a bit with the arrow keys, to match up the top black line as well.

Step 6 - Add some shadows.

Nothing in Photoshop can ever be done well with one large, obvious effect. Almost without fail, it’s better to build up a series of subtle effects than to try one effect at 100% strength - this is especially so with drop shadows.

Now, we all know that drop shadows are simply constructs in our mind - we *know* that a Photoshop drop shadow isn’t a real shadow, but if everyone plays along with the construct, then no-one gets hurt. The only aim is to make sure your drop shadow looks more on the ‘nah-it-doesn’t-bother-me’ side than on the ‘that’s-not-a-real-drop-shadow’ side. That is, fake it, but make it at least half possible.

In this case, I used cues from the original digital photograph, plus the above methodology and made one tight, darkish shadow:

…followed by one wide and shallow shadow at a very light opacity - both on separate layers, so I could tweak them all within the layers palette.

They were both made by using our saved path on a new payer, and making a solid shape from the selection. Since the book is also ‘cut’ from this same selection, there’s no chance of any of this new shape being visible - only the resulting shadow. I then simply used the layer effect, create drop shadow tool. Long gone are the days back in Photoshop 2.5.1 where only a very small group of people had knowledge of how to make drop shadows. Those were the days…

Step 7 - Add some grunge.

In effect, we’re done, if you wanted the book to be pristine. It’s pretty darn close to reality, but with a tiny bit more work, we can add some shading some other elements to make it a little more real.

Taking a look at the channels of our original book (choose the original photo layer then go to the channels pallette) we can see that the Red plate has some lovely shadows there on the spine:

Let’s grab that grunge and apply it (subtlely) to our new book. In the Red channel, my best practise is to simply go Select All (Command-A or Control-A), Copy (-C, anyone?), New (-N),’Enter’ to accept the new image size, and Paste (-V). I’d usually also do a Flatten (-E) by default, since my fingers seem to do that Copy/New/Yep/Paste/Flatten thing: day in, day out. The benefit here is that we can play with this image in a new file as if it were a sandbox - we can’t do any damage whatsoever to the real file, and we can then copy and paste it back into the new file later on - both files are exactly the same dimensions, so there’ won’t be any realignment issues at all.

With this new ‘grunge’ file, we want to isolate that darker portion only - and since it’s blacks we’re after, it’s really easierto work in greyscale mode. So go Image/Mode/Greyscale. That gets rid of anything do with colour. Go Levels (Control or Command-L), and really crank the thing up: bring the top left-hand triangle (the black point) down to the left hand edge of the first (leftmost) mountain. Then bring the top right-hand triangle (the white point) WAAAY up to the left, right up to somewhere nearing that same mountain, above. You’ll see the book mostly dissapear, apart from the grunge that we want to keep, plus some other artefacts of dark grey hanging around. When you’re happy with the grungy area, you’re good to go. Commit the Levels, and with a large white brush, paint away any unwanted dark grey artefacts, until you get something like this:

From here, reverse the process to get this image back into the main file. Select All (-A), Copy (-C), Close (-W, don’t save if you’re happy with it), then Paste (-V). BUT - we may have left the old file in the Red Channel. Make sure you’re back in ‘all channels’ before you paste (the top, ‘RGB’ Channel in the channels palette).

Set this new pasted layer to ‘Multiply’ in the layer palette’s Modes pop-up - this usually says ‘Normal’, if you can’t find it. Change the opacity of the payer so that the new grunge is visible, but just out of obivious sight. Squint a little to see if that changes things, and try the image at different zoom levels to make really sure.

For this task, I also duplicated the grunge layer (Command or Control-J), blurred it a fair bit, and varied the opacity on the duplicated layer to get it just right.

I also ran a black line down along the score mark (next to where the front cover meets the spine) - to reinvent that feauture of the reference book, hoping desperately that there would be a score mark in this years’ book, too. The line was simply blurred and had it’s opacity changed so that it looked about right.

Step 8 - Add some highlights, and we’re done!

Grunge is good, but shine is fine. Lastly, we’ll add some highlights to the right hand top and bottom corners of the cover. The reference book had some reflections of lights, but in reality, it’s just another construct, like our friend the drop shadow - to enhance the overall image. How many times would you have ever seen a ‘four window panes’ highlight on an image of a balloon, and never taken a second glance?

This time, we’ll get out a fresh layer, and very slowly build up the highlights using a white brush, set to very low opacity - say 2% or 3%. the older Photoshop ‘fade in X steps’ brush option has gone (I can’t seem to find it!) - so the next best option is taking time to paint over and over again with a 2% brush. This gives you total control to paint the highlights - start from the brightest part of the hioghlight, then work outwards, going a different direction each time. Use the number keys, or the number pad to change the opacity of the brush, and the square brackets to change the size of the brush on the fly. Make the layer set to ’screen’ mode (where the ‘multiply’ mode was on the layers palette), and vary the overall opacity to get the right balance. If you make a mistake, use the eraser set to the same size and opacity, and slowly remove excess highlights. Doing it all this way lets you average out your brush strokes, and slowly craft your artwork.

Finally (finally!) I added a few very subtle horizontal lines to the ‘pages’ of the book, still visible from our old reference photo. It’s very subtle, but makes that half-of-half-of-1% difference.

Here’s one of the final promotional pieces that it was used in:

So - that’s it. That’s enough, surely! You’ve learnt a lot more than ‘how to fake one book in Photoshop’, and I’ve typed for a lot longer than it actually *took* me to do the artwork in the first place.

If you have any comments, queries, or alternative methodologies, feel free to share them here.

AB out