The Image-Editing Survival Manual
wherein Photoshop means "any pixel-based layers-capable image editing program," except when it doesn't--
Okay, there seems to be enough interest to warrant doing this so here's the first installment. Hopefully it will be helpful, and not the reverse! (And sartorias, nobody is too stupid to do graphics, particularly not anyone who can tack up a horse and not get the snaffles mixed up with the woffles - computers and especially graphics programs like Photoshop et al are very complex machines that do a lot of different things, many of them at the same time, and which come with very poor documentation. "Here's your new airplane, here's your sloppily-written and untested manual, have fun!" -- it's really like that. Not that I'm necessarily going to be any better at explaining it, but I've been through it all, including the sneering "help" and clickety-clickety-now-you-know-[L]user, type of demos.)
Parameters:
I'm working on a Dell Dimension B110, Celeron 2.53 GHz processor with 256 MB of RAM and a 60 GB hard drive, integrated video card with no extra memory and the default Dell 17" conventional monitor that came with it, running Windows XP Home Edition. Not ideal, but it's what I've got. Everything in these tutorials (except the 3D stuff) could also be done on my old Pentium II machine running Windows 98SE with 96 MB of memory and an 8 GB hard drive, only a lot slower.
I'm using Adobe Photoshop 4, partly because that's what I've got (bought openly & legally, from remaindered stock, when I was a lot less poorly off), and partly because I still like it better in almost all respects than later versions, for several different reasons. Photoshop 3 was massively unstable and had serious screen redraw slowness, but 4 improved on it in every respect and would even run on my ancient 486, with care. And 4 loads quickly, superquick on this beast, no going for cups of coffee, and while it doesn't have all the bells and whistles of later versions, it does everything I need it to do, quickly and with very few crashes (usually due to having too many files in too many programs all open at once, which happens with every version of every program.) The two features I miss from later versions are 1) multiple undos (save often!) and 2) the ability to lock layers; but I'm used to being careful, after using this for almost ten years now.
A lot of the features of Photoshop are going to be similar or identical to other image editing programs, and I'm going to try to deal with the more basic, but obscure, features that are common to all but the most basic ("get rid of red-eye & crop your own pics") image-editing programs, with quick'n'dirty techniques for when you don't have a lot of time or money or either - and with the philosophies, or headspaces, which underly all pixel-based graphics work. If you already know image-editing inside and out, if you already have a thousand-and-one filters to do any effect you can think of, or if you want to know secrets of the most advanced features of the latest versions, this probably won't offer much to you - but maybe you'll find the one obscure Mystery revealed that will solve a problem you've been fighting with for years. Happens to me nearly every day. Some of the commands may be a little different on Macs, but everything should pretty much work the same way, especially if you're using Actual Photoshop, because Adobe has made a consistent effort for about 10 years now to keep both platforms parallel. If you're using Linux, sorry - afraid I never have and can't offer any Linux-specific advice. But I would assume that if you can get Photoshop running on Red Hat etc, that it would work the same inside...
There is only one piece of extra equipment which I consider nearly essential, besides a scanner - I use an old Canon USB scanner that was $96 new quite a few years ago, and which is better than some more expensive models I've used - and that is a pen-mouse. I would say as well, that it's worth spending the extra money for a Wacom - I had an Aiptek which cost about half as much and used it for years, but it was flimsy and the last two years it was held together with tape, it had serious problems with speed and responsiveness, which meant that sometimes you'd get no lines at all - and other times it would start scribbling across the page like an Etch-A-Sketch gone mad. Very bad. I also don't see the point in getting a giant tablet even if I had the money for it; I use the smallest Intuos and find it plenty of drawing room, I've used a large one and found it very tedious and clunky - but then, I learnt to draw freehand on small pocket sketchbooks and the margins of my notebooks, so it may be all what you're used to. The other great thing about Wacoms is that they don't use external power supplies - the stylus doesn't need any batteries and the tablet gets its juice through the USB cord. This is a bigger boon than you might think, until your stylus stops working and you have to go and buy another AAA in order to finish your project. But a pen-mouse of any sort not only is more comfortable (and avoiding repetetive motion injury is important) but it allows for insanely better accuracy when drawing or erasing. You can speed up the time needed and reduce the frustration involved in silo'ing a photo exponentially, with one. It's kind of like your trusty machete, in the image-editing jungle, or your tricorder, on an Away Team mission...
First things: Check your Photoshop [substitute your own image-editing program] preferences.
Under General make sure that "Interpolation" is set to "Bicubic (Better)", that "Anti-Alias Postscript" and "Save Palette Locations" are turned on, and the rest as you please, but they're not really important one way or the other. What "Interpolation - Bicubic" does, rather than the "Nearest Neighbor/Bilinear" options, is when you scale an image or part of an image up or down, instead of doing a faster, but less accurate, stretching on the picture, instead it does a calculation on each pixel and kind of feathers it out based on a formula. It's just one more way of keeping your pictures crisper and more accurate, and you don't, as far as I can tell, take much of a performance hit. "Anti-Alias Postscript" means that when you import a Postscript graphic - EPS or .ps file - it will be slightly feathered along the edge, instead of jaggy. You probably won't use this much, but it is important when you do it, and it saves the time of going "Oh, !@#%!" and having to re-import after changing the settings. "Save Palette Locations" is just what it says - after you arrange your toolbars on your workspace the way you like them, you don't have to set them all up again unless you have a really bad crash which eats your preferences. (It happens.) Again, just another little thing to make your life more efficient & less frustrating.
Under Saving Files I just go with "Save Previews" as the default - I only know one situation where it apparently caused an issue in printing, and that was years ago with a now-discontinued line of large-format inkjet drivers. I can't imagine it being a problem 99% of the time, and thumbnails are handy to have when trying to make sure you're opening the right file, for both you and the printer who may be using your file. How far backwards compatible do you need to go? Not sure - is anyone out there using anything older than Photoshop 4? Basically, if you are the only one going to need to change your file, then it's not going to be an issue one way or the other; if you're going to send it to someone else to make changes to the raw file and its layers - this is different from printing/outputting - then you should probably ask what version they're using anyway. It is not something that should come up much, and certainly not in prepress - you shouldn't be sending printers your layered Photoshop files, but rather TIFFs, unless they tell you explicitly otherwise. (The exception to this, for 3D rendering folks, is when you are offering layered files for the express purpose of people playing around with them, and in that case you want them back to at least version 4, so that everyone will be able to use them.) I don't know what "Save Metric Color Tags" even means, so I just leave it unchecked, on the basis that TMI can be a crash-causing thing, with files. (I told you this was a "rough guide"...)
Under Display and Cursors make sure that "CMYK composites" is set to "smoother" and turn off all the other check boxes. Set the "Painting Cursors" to "brush size" and the "Other Cursors" to "precise". Otherwise you get e.g. a big cartoony paintbrush or magnifying glass icon, and you can't be sure what will happen when you click down, where the stroke or selection will start.
Transparency & Gamut - just leave this alone, it's not worth fussing with imo.
What "Transparency" is about, is what color the background of your workspace is, so that you can tell what parts of a picture have been cut away and are now clear through. I used to go and reset the squares for the background display to a pale sherbet orange, because the gray was hard for me to see sometimes, esp when working on BW photos - but any color ends up being hard to see in some situation, and it's a pain to go and change it every time.
"Gamut settings" is picking a color that will be the warning color for if you have picked colors that won't print, because there are different color ranges that can be displayed 1) on monitors 2) printed on paper, and there are areas where they don't overlap (frex, you can't get a monitor to display true cyan (printers' blue for full-color process" printing) and you can't get that bright lime green or electric royal blue color on paper unless you buy neon inks and have it printed special.) I used to worry about this, but it never really helped me avoid creating something that wouldn't print right, it just made scary blotches on the screen - basically, remember that aqua blues look duller on screen and that neon colors can't be made by mixing red/blue/yellow/black, and you're fine. And if you're only doing work to be posted online, no worries! None of that is relevant anyway for you. But if you like tinkering, go right ahead and play with these settings, it's a free-will universe & you can't really hurt any thing by changing them.
Units & Rulers - set them for "inches" or "centimeters" which ever you are most comfortable with - not pixels, it will drive you insane trying to figure out where you are on a page measured in the 1000s, and picas are completely irrelevant in the modern digital prepress world. Seriously, every printshop I have dealt with or worked at, whenever a document came in set to "picas" first thing we did was change it to inches. It's an archaic unit of measure that is really counterproductive when you're trying to print things on paper measured in inches or centimeters, and it doesn't make you look more knowing or sophisticated (as one designer told me when I tried to tell her not to do it, because changing the preferences on 30 different documents in order to bind them in one booklet was a timewasting PITA.)
The "Column Size" things don't matter, because you are not going to be setting lots of copy in Photoshop. No, seriously, you're not. I will come hunt you down with a wet snappy towel, if you do, in the name of all prepress people everywhere who have had to try to cope with the nightmare that results when people insist on using inappropriate programs to build text-heavy pages.
Set "Point/Pica Size" to "Postscript," because for the little bits of lettering that you will be doing in Photoshop, you are better off going with a round number than a fraction of an archaic measure that we're stuck with. Not that a fraction of a seventy-second of an inch is a very big deal, but it's one less bit of prepress tweakiness, and as a matter of principle it's better to avoid tweakiness. (What is "tweakiness"? You'll know it when you spot it, but anything that just doesn't really make sense and leaves you going "why on EARTH did they set it up that way?" Applies to both OS/Software and Other People's Files.)
Guides & Grid - like the Transparency/Gamut settings, I don't fuss with this any more, but if you want to make the Guides your favorite color, or use the Grid option, go right ahead.
Plug Ins & Scratch Disk - The plug-ins folder is where the program by default goes and looks for any plug-ins you have installed. Some come with Photoshop, others you have to buy, but there are also a lot of free ones out there to be downloaded, developed by other artists. Most of them are imo pretty useless, or redundant (including some of the pay ones!) but some of them are handy, even if it's just for some specialized thing. Plug-ins, aka "filters" in Photoshop, add sfx like blurring or rippling or texturizing, applying a sort of "handpainted" or drawn effect, make a "glow" and the results are sometimes surreal, sometimes "fine-art" and you can play around with them in various combinations to, frex, rescue a lower-quality photo sometimes. But they have to be copied into the right folder to be able to access them - only here is where you decide what the "right" folder is, you might want to have your filters all in some common directory, or just leave them in the default folder. It doesn't matter, so long as you put them all in the same place and tell Photoshop here where to look for them. But if you can't remember where your filters go to install a new one, you can always go to this screen and check...
"Scratch Disk" - this is a biggie, for avoiding crashes and maximizing your efficiency, if you can afford it. What the scratch disk is, is the place where Photoshop stores its temp files while you are working. The more space for the scratch file, the bigger an image you can work on, but if you run out of space, you can have a nasty crash and lose it all. By default the temp files go into your main hard drive with the program, but if you have another drive installed, or another partition, you can specify to work in that one, and get more space, and if your temp files are not on the same drive as your programs are working from, you're less likely to get crashes and corruption too. So here is where you tell Photoshop where are the workrooms to spread out in, so to speak.
Memory & Image Cache this is another biggie, especially the "Physical Memory Usage" - what that does is tell Photoshop how much of your RAM to set aside just for Photoshop to use, the more you can give it the better it will run, but then everything else, including your OS, your internet browser, and any audio software you're listening to, will be fighting over the scraps. I have it given 50% of the approx 230 available MB, and I can usually run one, or maybe 2 other programs, like Acrobat and DAZ Studio, or Wordpad and Quark, at the same time, without it crashing. But not always - that also depends on how big the files are that you have open in each program, and what processes you're running, how system-resource-intensive they are (frex, "searching" is a very demanding process, whether you're searching your hard drive or inside a file, and that can really bog everything down or even cause lockups if you're, say, simultaneously trying to apply a filter to a big picture.)
----
Now that we've got our machine fine-tuned and ready to roll, what does it actually DO???
One reason that I think people have so much trouble with graphics programs is that the interfaces and manuals were created by non-linear-thinking aliens that nobody does a very good job of explaining how they actually work - what the virtual reality they are VR'ing is supposed to be, of defining the headspace of them, so to speak. The eplanations don't seem to explain - at least, that was my experience, poring through the Photoshop and Corel Paint manuals and computer magazines(and yes, that is the right way to spell that kind of "poring"--!!!) reading/rereading about "additive and subtractive color" or "layer masks" and trying to relate all this stuff to ANYTHING of my prior dozen-odd years of formal and informal visual art training and pragma... I coined the word "Gibberese" to internally describe the technical language of these things, and the amount of frustration I underwent in trying to learn how to, e.g., make a simple drop shadow back in circa 1996, under deadline, cannot be described.
Right now we are working with what is known as bitmap graphics - and right there we have a problem. I prefer the term "pixel-based" as in opposition to "vector based," because just as some hawks are falcons and some falcons hawks depending on what way you're using each term, "bitmap" can refer to either a class of files, or a particular type of file within that class of files. The word "raster" is also used to refer to this kind of art, but since that's a neologism it doesn't really help until you've worked with this stuff long enough to know that "raster graphics" means things made up of pixels - but then, we have another problem, because on one level, everything on the computer is made up of pixels, that's what you're looking at all over your monitor, and if it isn't pixels then you can't SEE it, right?
So here is my attempt to thrash out the dialectical kudzu of digital art: Think of a "pixel" as being like a mosaic tile, which is a mathematical plane made of protean Ur-Matter - it has no depth, no fixed size, no fixed color, the only attribute it has is being square* and being pure potentiality, capable of taking any color and shade of an array of "millions of colors," and being made to be any size in relation to any object, given the right command set. This "pixel," this virtual unit of imagery, is like a musical note - what is a "note" -- ? Nothing, until you set down a specific pitch and duration, with a specific array of prior commands that tell the musician reading it which pitch and how long in relation to still other notes - and even then, in what way does it exist, until someone plays it? And the duration is not fixed, it has no absolute value, because the conductor or performer may decide to speed up or slow down the tempo, try it allegro or andante, and the pitch is not even fixed and absolute, because you might play the same musical notation on a different instrument, or one that is tuned high or low--!
In such a way we have "screen pixels" and "image pixels" and yet none of these things do we actually ever see - because a pixel is not identical with either a phosphor dot on a monitor or a scrap of ink on a printed page, those are simply the means by which pixels are represented - and so we use the word interchangeably to mean all of them. And thus a "pixel" in your jpg can be one size or another, or rather represented by screen dots at one size or another, depending on whether you have used a "zoom" command, or have set your monitor resolution differently - and that is why your 100 x 100 pixel icon can look one size in your browser and another when you are working on it in Photoshop, or one size on one computer and another on a different one.
It is all conventional, in the strict sense of that word - and yet it is still a reality, it is just a very fluid and Platonic one, and that makes it hard to think about abstractly - but we must, in order to master it, music or digital art.
It is also helpful to undrestand what is going on with "pixel-based/bitmap/raster-graphics" by compare/contrast with the alternative, "vector-based graphics", which just to make you more insane, are also represented by pixels of the monitor sort and "rasterized" when printed out on paper. But a vector graphic works by a different kind of internal Platonic architecture than does a JPG, TIFF, or bitmap - a vector graphic works by plotting out a series of points on a virtual plane, and flinging lines or loops between the co-ordinates depending on what command has been used to define the point, as a straight angle joint or a [Bezier-] curve joint. This means it can be scaled up, in theory, infinitely large, with no degradation of quality (in practice, no, because you can only make it as big as you can find something to print it with) whereas a pixel-based graphic, no matter that you can tell it to make the pixels any size, still has limited dimensions because it must be made up of a certain fixed number of tiles - and the "bigger" you make the tiles in relation to the picture formed by them, the more rough and jagged the image comes out, and the "smaller" they are, the more of them it takes to make the same picture, even though just as with real mosaics you get a much smoother and more natural image, less crude and blocky - and the more "expensive" it becomes in terms of memory needed by the computer and printer to display it.
Moreover, pixel-based images, like mosaics, can be "made" of more expensive materials, so to speak - you can buy plain stone tiles, or semi-precious stone tiles, or stained glass tiles, or even ones with gold leaf bonded into them, which are very pricy indeed. Pixels can come in plain binary, black or white, and those are the "cheapest" when it comes to file size and memory needed to render them onscreen or printed. This is what is meant in Photoshop when you have the option to convert a b/w image to a "bitmap" under the "Image/Mode" tab, btw. This does give jagged edges, as well as being only solid line art, with no shading, but since the pixels are very "light" file-size-wise, you can make a picture that is 600, even 1200 ppi (more on that later, too) and not have it be enormous. This can be useful for certain graphic purposes, depending on what you are trying to print and how much RAM/hard drive space you have...
Then you can have the 256-shade pixels, which may be either grayscale only, or "indexed color," as in GIF files. Grayscale images only have 256 shades, because although there are an infinite number of possible shades between black and white, 256 is all you really need to print/display a convincingly-smooth 3D monochrome photograph or drawing. So that is all that computers use, in order to reduce system strain, and all you ever see when you see a printed b/w picture on a page (as opposed to an actual photo printed from a negative in a darkroom. Some fine-art printers get around this logistical limitation by printing "2-color black-and-white," where a gray ink is used to make a duotone picture, and yes, you can see a difference, if you have examples of both side-by-side for comparison, but this is expensive to do and esoteric stuff, because of the expense and also because printing duotones and controlling the outcome is a bit finicky, same with "six-color printing" for full-color, partly because of the extra expense involved which means people aren't used to doing it...)
Now, while there are likewise an infinite number of hues which can be seen by the human eye, and pretty much so with what can be rendered in paints, pigments and dyes in combination, the computer graphic is limited to what monitors and video cards have been devised to do - I remember when we went from "binary" to 256-color 8-bit VGA and wow, was it ever cool when you could first play games in all of those crude bright tones instead of just pure amber-on-black (or maybe light green-on-black, depending) or vice-versa, and then we got SVGA and full 24-bit color, for a while a range of several tens of thousands of colors was available, and now we are up to 32-bit color & what is called "millions of colors," the exact number of which I don't recall, but which is nevertheless a multiple of eight, you will notice that it all works on powers of 2, 2 cubed being 8 etc - to get better quality both in video and in sound, you start out with binary choices and you compound them, making "richer" pixels etc. The 256-indexed color palette of GIF-format images is a compromise to get an illusion of full color while not taking up too much file size, and ergo bandwidth particularly - it works by dropping out or combining, using any of several possible algorithms, the colors that are most similar to each other, and that is why they are often very grainy-looking and/or posterized, with one But! and that is, the closer your original picture is to monochrome, the more likely it will have the least amount of discernable quality reduction. If I have written the preceding paragraph correctly, you will understand why.
Because GIFs are what is called "indexed color," which is something I kind of understand abstractly, but can't explain very well, but which means that you only get the colors to paint with that you see on the screen already, you don't want to try to edit a GIF file as-is. If you want to manipulate a GIF, first convert it to RGB mode, which will give you the "millions of colors" option, so that whatever shade you make with your color-picker, you have it to paint with. Otherwise you will be stuck only working with what is there, and that is why if there is no, say, red already in the .gif, and you want to slap (eg) a red X over something in it, it just won't work, you can't get a red because red is not one of your 256 choices. You must convert to RGB, and paint in red, and then convert it back to a gif, in which case some other color shades will get thrown out, or else save it as a JPG, which doesn't have the same limited color palette.
Why would anyone use gifs, you might ask, with all their limits? Because a) they're often much smaller than JPGs, in terms of file size - the exception being, if you have a b/w photo or a grayscale scan of a drawing! - and b) you can make one color of the 256 be "transparent" and thus have a picture which is silhouetted (hereafter "silo'd," the technical prepress abbreviation) on whatever background color or texture you want on your webpage. The original Myst game also used indexed color to have very rich-looking scene backgrounds that would still work on even a fairly basic, non-graphics-designed computer. But ordinarily, and in all print applications (!!!!) you want to use RGB files, whether JPGs or TIFFS or something else.
(If you are going "but what about CMYK?!!?" then why are you reading this? -- just hold on, eventually I'll get around to explaining how CMYK works and also how, in digital graphics situations, it's just more virtuality. Monitors are not CMYK-enabled structures, so CMYK-mode image files are "really" RGB until they actually are burned to film or plate...yes, it's all very crazy-making, sorry. It's like transposed music, kind of.)
JPEGS, which stands for "Joint Photo Experts Guild", btw, are made of full-range-of-color "millions of colors" RGB pixels - which are your mosaic tiles made of gold-leaf or lapis-lazuli, as it were - that save on file size by a different method, they use a different type of "compression" which is an unhelpful term but means that you're doing something to reduce the amount of digits it takes to write the picture to a disk, and which can be either "lossy" (ie the picture is lower-quality afterwards in one way or other) or "lossless" which means that redundent information has been taken out, if there was any to take out. Frex, large areas of flat color - instead of writing over and over again something like "Row 1, Pixel 1: yellow; Row 1, Pixel 2: yellow; Row 1, Pixel 3, yellow" all the way down to row 35 on your mosaic grid, you have the file write "Row 1, Pixel 1 to Row 35, pixel 70: yellow" and that takes up a LOT less characters than typing it out hundreds of times.**
It's like the difference between an abbreviation which you understand, and one you don't fully get - e.g., "e.g." vs BRDHD on a license plate - does the latter mean "broadhead" or "beard head" or "board head" or what? If it's on a surfer-themed jeep, you might guess at it being "BOARD HEAD" but you're still missing key information and having to guess at what goes in between, and maybe it's pretty obvious but maybe it isn't. And that's what happens with "lossy" compression, and how sometimes-critical details can be lost out of JPGS, if they're crunched down too low. OTOH, if you save a JPG with no loss, then you may just have a huge file and no space/bandwidth saved at all. It depends on the individual image. But a good rule of thumb is, if there is a lot of color change, and a lot of variegation, then you won't get much savings with lossless compression, and how far down you feel you can go before it starts to look too bad, is something you will need to experiment. (This is why you should always make a backup copy of a picture, before you start playing around with it - it's too easy to forget, save something lossy and not be able to go back to the higher-quality again...)
TIFFs (from Tagged Image File Format) are (mostly) lossless and/or uncompressed at all, they can come in a couple different flavors but if you only work on the internet, you won't prolly deal with them ever. They are primarily (historically and ontologically) used for professional printing, and until recently you couldn't even view them in a browser - Quicktime has included a TIFF viewing plugin of late, but you don't want to use them in web pages, because they aren't made to work with browsers and HTML and vice versa. However, they are superior, as a rule, for prepress, and conversely, until relatively recently, you couldn't use anything but TIFFs when it came to pixel-based graphics in professional desktop publishing, there were plugins to allow Quark and Pagemaker to read JPGs etc but they were glitchy, and the results when you printed out one of these web-intended images on a printing press were...unreliable, shall we say? Sometimes you might get a negative image; sometimes you might just get big black blotches. A lot of times your system just crashed. So, now Quark and Indesign, Pagemaker's successor, can read other formats, but you're still better off in terms of quality and safety using TIFFs, if you're going to print - and not compressed tiffs, either, because the lossless compression mode, LZW, can't be read by a lot of printer engines. Either it crashes or you get a blank space where your picture should be, and that can be very expensive and frustrating.
And now, here's where it gets really insane - you can also save a millions-of-colors tiff as a "bitmap," because the word also means a type of file language that can be read by a lot of different programs (tho' not web browsers - unless saved as an .ico file! there's always an "unless") because it's one of the oldest formats around on personal computers (there's also a whole range of high-end professional graphics stuff that I'm not even getting into, Scitex/RAW etc, altho' I'll try to explain it if anyone needs to know...) and can be opened by, e.g. MS Paint. And it has its own compression possibilities, too! But it doesn't change (afaik) the quality to save something as a bitmap - the only thing is, bitmaps of this sort are REALLY bad news to try to print with, because you can't set the page-printing dimensions internally very well. They're kind of the "pidgin" of pixel-based file format languages.
And this is how the theory and the pragma collide, because we are incarnates working with physical stuff, but you have to be able to conceptualize it to know how to work with it, but "if you can dream it, you can do it!" is not always true - not even with an unlimited budget, tho' that makes it easier. Sometimes people just haven't figured out how to build a machine that can do something "simple", or trying to make one do one thing has unintended consequences elsewhere. Which is not just true of digital art.
So, a really quick little explanation of what is going on and why there are so many different types of graphics formats, and what they're for, and when to use/not use them, which hopefully didn't confuse everyone too much and left everyone with useful information. For comparison, here's the relative file sizes of a blank 100 x 100 pixel image - your standard icon/avatar, in the different modes:
Bitmap (pure b/w): 2k
Grayscale (256 possible shades): 10k
Index (256 possible shades): 10k
RGB (millions of possible shades): 30k
These seem very small, but they will add up quickly, if you multiply them on each side!
-Okay? Everyone got their water, anti-malarial-tablets, pocket-knife-with-the-implement-for-getting-stones-out-of-horses-hooves, spare clothes, and towel? Now we can get to the good stuff - next lesson will be about Alpha Channels and Layers, although we may make side excursions into whatever strikes my fancy. We'll learn how to go from this to this, which will neatly cover both layers and channels, at least some aspects of both.
*Mostly, but we're not going to talk about the rectangular ones now because they are an abomination unto the Production Department, and I don't know why you get jpgs with them sometimes, only that they are a) distorted and b) Evil.
**This is not how colors are actually written, this is just a "natural language" approximation of it - if you want to see what it actually looks like open up a JPG in a text-editing program like Notepad...but save first, because it can sometimes crash your system. There are a LOT of little letters and numbers in even a small picture!
Special Lesson Bonus Link
Some cool free tiling filters (Win only, sorry).
A free "Gothic" effect plugin for MAC/Win here
Okay, there seems to be enough interest to warrant doing this so here's the first installment. Hopefully it will be helpful, and not the reverse! (And sartorias, nobody is too stupid to do graphics, particularly not anyone who can tack up a horse and not get the snaffles mixed up with the woffles - computers and especially graphics programs like Photoshop et al are very complex machines that do a lot of different things, many of them at the same time, and which come with very poor documentation. "Here's your new airplane, here's your sloppily-written and untested manual, have fun!" -- it's really like that. Not that I'm necessarily going to be any better at explaining it, but I've been through it all, including the sneering "help" and clickety-clickety-now-you-know-[L]user, type of demos.)
Parameters:
I'm working on a Dell Dimension B110, Celeron 2.53 GHz processor with 256 MB of RAM and a 60 GB hard drive, integrated video card with no extra memory and the default Dell 17" conventional monitor that came with it, running Windows XP Home Edition. Not ideal, but it's what I've got. Everything in these tutorials (except the 3D stuff) could also be done on my old Pentium II machine running Windows 98SE with 96 MB of memory and an 8 GB hard drive, only a lot slower.
I'm using Adobe Photoshop 4, partly because that's what I've got (bought openly & legally, from remaindered stock, when I was a lot less poorly off), and partly because I still like it better in almost all respects than later versions, for several different reasons. Photoshop 3 was massively unstable and had serious screen redraw slowness, but 4 improved on it in every respect and would even run on my ancient 486, with care. And 4 loads quickly, superquick on this beast, no going for cups of coffee, and while it doesn't have all the bells and whistles of later versions, it does everything I need it to do, quickly and with very few crashes (usually due to having too many files in too many programs all open at once, which happens with every version of every program.) The two features I miss from later versions are 1) multiple undos (save often!) and 2) the ability to lock layers; but I'm used to being careful, after using this for almost ten years now.
A lot of the features of Photoshop are going to be similar or identical to other image editing programs, and I'm going to try to deal with the more basic, but obscure, features that are common to all but the most basic ("get rid of red-eye & crop your own pics") image-editing programs, with quick'n'dirty techniques for when you don't have a lot of time or money or either - and with the philosophies, or headspaces, which underly all pixel-based graphics work. If you already know image-editing inside and out, if you already have a thousand-and-one filters to do any effect you can think of, or if you want to know secrets of the most advanced features of the latest versions, this probably won't offer much to you - but maybe you'll find the one obscure Mystery revealed that will solve a problem you've been fighting with for years. Happens to me nearly every day. Some of the commands may be a little different on Macs, but everything should pretty much work the same way, especially if you're using Actual Photoshop, because Adobe has made a consistent effort for about 10 years now to keep both platforms parallel. If you're using Linux, sorry - afraid I never have and can't offer any Linux-specific advice. But I would assume that if you can get Photoshop running on Red Hat etc, that it would work the same inside...
There is only one piece of extra equipment which I consider nearly essential, besides a scanner - I use an old Canon USB scanner that was $96 new quite a few years ago, and which is better than some more expensive models I've used - and that is a pen-mouse. I would say as well, that it's worth spending the extra money for a Wacom - I had an Aiptek which cost about half as much and used it for years, but it was flimsy and the last two years it was held together with tape, it had serious problems with speed and responsiveness, which meant that sometimes you'd get no lines at all - and other times it would start scribbling across the page like an Etch-A-Sketch gone mad. Very bad. I also don't see the point in getting a giant tablet even if I had the money for it; I use the smallest Intuos and find it plenty of drawing room, I've used a large one and found it very tedious and clunky - but then, I learnt to draw freehand on small pocket sketchbooks and the margins of my notebooks, so it may be all what you're used to. The other great thing about Wacoms is that they don't use external power supplies - the stylus doesn't need any batteries and the tablet gets its juice through the USB cord. This is a bigger boon than you might think, until your stylus stops working and you have to go and buy another AAA in order to finish your project. But a pen-mouse of any sort not only is more comfortable (and avoiding repetetive motion injury is important) but it allows for insanely better accuracy when drawing or erasing. You can speed up the time needed and reduce the frustration involved in silo'ing a photo exponentially, with one. It's kind of like your trusty machete, in the image-editing jungle, or your tricorder, on an Away Team mission...
First things: Check your Photoshop [substitute your own image-editing program] preferences.
Under General make sure that "Interpolation" is set to "Bicubic (Better)", that "Anti-Alias Postscript" and "Save Palette Locations" are turned on, and the rest as you please, but they're not really important one way or the other. What "Interpolation - Bicubic" does, rather than the "Nearest Neighbor/Bilinear" options, is when you scale an image or part of an image up or down, instead of doing a faster, but less accurate, stretching on the picture, instead it does a calculation on each pixel and kind of feathers it out based on a formula. It's just one more way of keeping your pictures crisper and more accurate, and you don't, as far as I can tell, take much of a performance hit. "Anti-Alias Postscript" means that when you import a Postscript graphic - EPS or .ps file - it will be slightly feathered along the edge, instead of jaggy. You probably won't use this much, but it is important when you do it, and it saves the time of going "Oh, !@#%!" and having to re-import after changing the settings. "Save Palette Locations" is just what it says - after you arrange your toolbars on your workspace the way you like them, you don't have to set them all up again unless you have a really bad crash which eats your preferences. (It happens.) Again, just another little thing to make your life more efficient & less frustrating.
Under Saving Files I just go with "Save Previews" as the default - I only know one situation where it apparently caused an issue in printing, and that was years ago with a now-discontinued line of large-format inkjet drivers. I can't imagine it being a problem 99% of the time, and thumbnails are handy to have when trying to make sure you're opening the right file, for both you and the printer who may be using your file. How far backwards compatible do you need to go? Not sure - is anyone out there using anything older than Photoshop 4? Basically, if you are the only one going to need to change your file, then it's not going to be an issue one way or the other; if you're going to send it to someone else to make changes to the raw file and its layers - this is different from printing/outputting - then you should probably ask what version they're using anyway. It is not something that should come up much, and certainly not in prepress - you shouldn't be sending printers your layered Photoshop files, but rather TIFFs, unless they tell you explicitly otherwise. (The exception to this, for 3D rendering folks, is when you are offering layered files for the express purpose of people playing around with them, and in that case you want them back to at least version 4, so that everyone will be able to use them.) I don't know what "Save Metric Color Tags" even means, so I just leave it unchecked, on the basis that TMI can be a crash-causing thing, with files. (I told you this was a "rough guide"...)
Under Display and Cursors make sure that "CMYK composites" is set to "smoother" and turn off all the other check boxes. Set the "Painting Cursors" to "brush size" and the "Other Cursors" to "precise". Otherwise you get e.g. a big cartoony paintbrush or magnifying glass icon, and you can't be sure what will happen when you click down, where the stroke or selection will start.
Transparency & Gamut - just leave this alone, it's not worth fussing with imo.
What "Transparency" is about, is what color the background of your workspace is, so that you can tell what parts of a picture have been cut away and are now clear through. I used to go and reset the squares for the background display to a pale sherbet orange, because the gray was hard for me to see sometimes, esp when working on BW photos - but any color ends up being hard to see in some situation, and it's a pain to go and change it every time.
"Gamut settings" is picking a color that will be the warning color for if you have picked colors that won't print, because there are different color ranges that can be displayed 1) on monitors 2) printed on paper, and there are areas where they don't overlap (frex, you can't get a monitor to display true cyan (printers' blue for full-color process" printing) and you can't get that bright lime green or electric royal blue color on paper unless you buy neon inks and have it printed special.) I used to worry about this, but it never really helped me avoid creating something that wouldn't print right, it just made scary blotches on the screen - basically, remember that aqua blues look duller on screen and that neon colors can't be made by mixing red/blue/yellow/black, and you're fine. And if you're only doing work to be posted online, no worries! None of that is relevant anyway for you. But if you like tinkering, go right ahead and play with these settings, it's a free-will universe & you can't really hurt any thing by changing them.
Units & Rulers - set them for "inches" or "centimeters" which ever you are most comfortable with - not pixels, it will drive you insane trying to figure out where you are on a page measured in the 1000s, and picas are completely irrelevant in the modern digital prepress world. Seriously, every printshop I have dealt with or worked at, whenever a document came in set to "picas" first thing we did was change it to inches. It's an archaic unit of measure that is really counterproductive when you're trying to print things on paper measured in inches or centimeters, and it doesn't make you look more knowing or sophisticated (as one designer told me when I tried to tell her not to do it, because changing the preferences on 30 different documents in order to bind them in one booklet was a timewasting PITA.)
The "Column Size" things don't matter, because you are not going to be setting lots of copy in Photoshop. No, seriously, you're not. I will come hunt you down with a wet snappy towel, if you do, in the name of all prepress people everywhere who have had to try to cope with the nightmare that results when people insist on using inappropriate programs to build text-heavy pages.
Set "Point/Pica Size" to "Postscript," because for the little bits of lettering that you will be doing in Photoshop, you are better off going with a round number than a fraction of an archaic measure that we're stuck with. Not that a fraction of a seventy-second of an inch is a very big deal, but it's one less bit of prepress tweakiness, and as a matter of principle it's better to avoid tweakiness. (What is "tweakiness"? You'll know it when you spot it, but anything that just doesn't really make sense and leaves you going "why on EARTH did they set it up that way?" Applies to both OS/Software and Other People's Files.)
Guides & Grid - like the Transparency/Gamut settings, I don't fuss with this any more, but if you want to make the Guides your favorite color, or use the Grid option, go right ahead.
Plug Ins & Scratch Disk - The plug-ins folder is where the program by default goes and looks for any plug-ins you have installed. Some come with Photoshop, others you have to buy, but there are also a lot of free ones out there to be downloaded, developed by other artists. Most of them are imo pretty useless, or redundant (including some of the pay ones!) but some of them are handy, even if it's just for some specialized thing. Plug-ins, aka "filters" in Photoshop, add sfx like blurring or rippling or texturizing, applying a sort of "handpainted" or drawn effect, make a "glow" and the results are sometimes surreal, sometimes "fine-art" and you can play around with them in various combinations to, frex, rescue a lower-quality photo sometimes. But they have to be copied into the right folder to be able to access them - only here is where you decide what the "right" folder is, you might want to have your filters all in some common directory, or just leave them in the default folder. It doesn't matter, so long as you put them all in the same place and tell Photoshop here where to look for them. But if you can't remember where your filters go to install a new one, you can always go to this screen and check...
"Scratch Disk" - this is a biggie, for avoiding crashes and maximizing your efficiency, if you can afford it. What the scratch disk is, is the place where Photoshop stores its temp files while you are working. The more space for the scratch file, the bigger an image you can work on, but if you run out of space, you can have a nasty crash and lose it all. By default the temp files go into your main hard drive with the program, but if you have another drive installed, or another partition, you can specify to work in that one, and get more space, and if your temp files are not on the same drive as your programs are working from, you're less likely to get crashes and corruption too. So here is where you tell Photoshop where are the workrooms to spread out in, so to speak.
Memory & Image Cache this is another biggie, especially the "Physical Memory Usage" - what that does is tell Photoshop how much of your RAM to set aside just for Photoshop to use, the more you can give it the better it will run, but then everything else, including your OS, your internet browser, and any audio software you're listening to, will be fighting over the scraps. I have it given 50% of the approx 230 available MB, and I can usually run one, or maybe 2 other programs, like Acrobat and DAZ Studio, or Wordpad and Quark, at the same time, without it crashing. But not always - that also depends on how big the files are that you have open in each program, and what processes you're running, how system-resource-intensive they are (frex, "searching" is a very demanding process, whether you're searching your hard drive or inside a file, and that can really bog everything down or even cause lockups if you're, say, simultaneously trying to apply a filter to a big picture.)
----
Now that we've got our machine fine-tuned and ready to roll, what does it actually DO???
One reason that I think people have so much trouble with graphics programs is that the interfaces and manuals were created by non-linear-thinking aliens that nobody does a very good job of explaining how they actually work - what the virtual reality they are VR'ing is supposed to be, of defining the headspace of them, so to speak. The eplanations don't seem to explain - at least, that was my experience, poring through the Photoshop and Corel Paint manuals and computer magazines(and yes, that is the right way to spell that kind of "poring"--!!!) reading/rereading about "additive and subtractive color" or "layer masks" and trying to relate all this stuff to ANYTHING of my prior dozen-odd years of formal and informal visual art training and pragma... I coined the word "Gibberese" to internally describe the technical language of these things, and the amount of frustration I underwent in trying to learn how to, e.g., make a simple drop shadow back in circa 1996, under deadline, cannot be described.
Right now we are working with what is known as bitmap graphics - and right there we have a problem. I prefer the term "pixel-based" as in opposition to "vector based," because just as some hawks are falcons and some falcons hawks depending on what way you're using each term, "bitmap" can refer to either a class of files, or a particular type of file within that class of files. The word "raster" is also used to refer to this kind of art, but since that's a neologism it doesn't really help until you've worked with this stuff long enough to know that "raster graphics" means things made up of pixels - but then, we have another problem, because on one level, everything on the computer is made up of pixels, that's what you're looking at all over your monitor, and if it isn't pixels then you can't SEE it, right?
So here is my attempt to thrash out the dialectical kudzu of digital art: Think of a "pixel" as being like a mosaic tile, which is a mathematical plane made of protean Ur-Matter - it has no depth, no fixed size, no fixed color, the only attribute it has is being square* and being pure potentiality, capable of taking any color and shade of an array of "millions of colors," and being made to be any size in relation to any object, given the right command set. This "pixel," this virtual unit of imagery, is like a musical note - what is a "note" -- ? Nothing, until you set down a specific pitch and duration, with a specific array of prior commands that tell the musician reading it which pitch and how long in relation to still other notes - and even then, in what way does it exist, until someone plays it? And the duration is not fixed, it has no absolute value, because the conductor or performer may decide to speed up or slow down the tempo, try it allegro or andante, and the pitch is not even fixed and absolute, because you might play the same musical notation on a different instrument, or one that is tuned high or low--!
In such a way we have "screen pixels" and "image pixels" and yet none of these things do we actually ever see - because a pixel is not identical with either a phosphor dot on a monitor or a scrap of ink on a printed page, those are simply the means by which pixels are represented - and so we use the word interchangeably to mean all of them. And thus a "pixel" in your jpg can be one size or another, or rather represented by screen dots at one size or another, depending on whether you have used a "zoom" command, or have set your monitor resolution differently - and that is why your 100 x 100 pixel icon can look one size in your browser and another when you are working on it in Photoshop, or one size on one computer and another on a different one.
It is all conventional, in the strict sense of that word - and yet it is still a reality, it is just a very fluid and Platonic one, and that makes it hard to think about abstractly - but we must, in order to master it, music or digital art.
It is also helpful to undrestand what is going on with "pixel-based/bitmap/raster-graphics" by compare/contrast with the alternative, "vector-based graphics", which just to make you more insane, are also represented by pixels of the monitor sort and "rasterized" when printed out on paper. But a vector graphic works by a different kind of internal Platonic architecture than does a JPG, TIFF, or bitmap - a vector graphic works by plotting out a series of points on a virtual plane, and flinging lines or loops between the co-ordinates depending on what command has been used to define the point, as a straight angle joint or a [Bezier-] curve joint. This means it can be scaled up, in theory, infinitely large, with no degradation of quality (in practice, no, because you can only make it as big as you can find something to print it with) whereas a pixel-based graphic, no matter that you can tell it to make the pixels any size, still has limited dimensions because it must be made up of a certain fixed number of tiles - and the "bigger" you make the tiles in relation to the picture formed by them, the more rough and jagged the image comes out, and the "smaller" they are, the more of them it takes to make the same picture, even though just as with real mosaics you get a much smoother and more natural image, less crude and blocky - and the more "expensive" it becomes in terms of memory needed by the computer and printer to display it.
Moreover, pixel-based images, like mosaics, can be "made" of more expensive materials, so to speak - you can buy plain stone tiles, or semi-precious stone tiles, or stained glass tiles, or even ones with gold leaf bonded into them, which are very pricy indeed. Pixels can come in plain binary, black or white, and those are the "cheapest" when it comes to file size and memory needed to render them onscreen or printed. This is what is meant in Photoshop when you have the option to convert a b/w image to a "bitmap" under the "Image/Mode" tab, btw. This does give jagged edges, as well as being only solid line art, with no shading, but since the pixels are very "light" file-size-wise, you can make a picture that is 600, even 1200 ppi (more on that later, too) and not have it be enormous. This can be useful for certain graphic purposes, depending on what you are trying to print and how much RAM/hard drive space you have...
Then you can have the 256-shade pixels, which may be either grayscale only, or "indexed color," as in GIF files. Grayscale images only have 256 shades, because although there are an infinite number of possible shades between black and white, 256 is all you really need to print/display a convincingly-smooth 3D monochrome photograph or drawing. So that is all that computers use, in order to reduce system strain, and all you ever see when you see a printed b/w picture on a page (as opposed to an actual photo printed from a negative in a darkroom. Some fine-art printers get around this logistical limitation by printing "2-color black-and-white," where a gray ink is used to make a duotone picture, and yes, you can see a difference, if you have examples of both side-by-side for comparison, but this is expensive to do and esoteric stuff, because of the expense and also because printing duotones and controlling the outcome is a bit finicky, same with "six-color printing" for full-color, partly because of the extra expense involved which means people aren't used to doing it...)
Now, while there are likewise an infinite number of hues which can be seen by the human eye, and pretty much so with what can be rendered in paints, pigments and dyes in combination, the computer graphic is limited to what monitors and video cards have been devised to do - I remember when we went from "binary" to 256-color 8-bit VGA and wow, was it ever cool when you could first play games in all of those crude bright tones instead of just pure amber-on-black (or maybe light green-on-black, depending) or vice-versa, and then we got SVGA and full 24-bit color, for a while a range of several tens of thousands of colors was available, and now we are up to 32-bit color & what is called "millions of colors," the exact number of which I don't recall, but which is nevertheless a multiple of eight, you will notice that it all works on powers of 2, 2 cubed being 8 etc - to get better quality both in video and in sound, you start out with binary choices and you compound them, making "richer" pixels etc. The 256-indexed color palette of GIF-format images is a compromise to get an illusion of full color while not taking up too much file size, and ergo bandwidth particularly - it works by dropping out or combining, using any of several possible algorithms, the colors that are most similar to each other, and that is why they are often very grainy-looking and/or posterized, with one But! and that is, the closer your original picture is to monochrome, the more likely it will have the least amount of discernable quality reduction. If I have written the preceding paragraph correctly, you will understand why.
Because GIFs are what is called "indexed color," which is something I kind of understand abstractly, but can't explain very well, but which means that you only get the colors to paint with that you see on the screen already, you don't want to try to edit a GIF file as-is. If you want to manipulate a GIF, first convert it to RGB mode, which will give you the "millions of colors" option, so that whatever shade you make with your color-picker, you have it to paint with. Otherwise you will be stuck only working with what is there, and that is why if there is no, say, red already in the .gif, and you want to slap (eg) a red X over something in it, it just won't work, you can't get a red because red is not one of your 256 choices. You must convert to RGB, and paint in red, and then convert it back to a gif, in which case some other color shades will get thrown out, or else save it as a JPG, which doesn't have the same limited color palette.
Why would anyone use gifs, you might ask, with all their limits? Because a) they're often much smaller than JPGs, in terms of file size - the exception being, if you have a b/w photo or a grayscale scan of a drawing! - and b) you can make one color of the 256 be "transparent" and thus have a picture which is silhouetted (hereafter "silo'd," the technical prepress abbreviation) on whatever background color or texture you want on your webpage. The original Myst game also used indexed color to have very rich-looking scene backgrounds that would still work on even a fairly basic, non-graphics-designed computer. But ordinarily, and in all print applications (!!!!) you want to use RGB files, whether JPGs or TIFFS or something else.
(If you are going "but what about CMYK?!!?" then why are you reading this? -- just hold on, eventually I'll get around to explaining how CMYK works and also how, in digital graphics situations, it's just more virtuality. Monitors are not CMYK-enabled structures, so CMYK-mode image files are "really" RGB until they actually are burned to film or plate...yes, it's all very crazy-making, sorry. It's like transposed music, kind of.)
JPEGS, which stands for "Joint Photo Experts Guild", btw, are made of full-range-of-color "millions of colors" RGB pixels - which are your mosaic tiles made of gold-leaf or lapis-lazuli, as it were - that save on file size by a different method, they use a different type of "compression" which is an unhelpful term but means that you're doing something to reduce the amount of digits it takes to write the picture to a disk, and which can be either "lossy" (ie the picture is lower-quality afterwards in one way or other) or "lossless" which means that redundent information has been taken out, if there was any to take out. Frex, large areas of flat color - instead of writing over and over again something like "Row 1, Pixel 1: yellow; Row 1, Pixel 2: yellow; Row 1, Pixel 3, yellow" all the way down to row 35 on your mosaic grid, you have the file write "Row 1, Pixel 1 to Row 35, pixel 70: yellow" and that takes up a LOT less characters than typing it out hundreds of times.**
It's like the difference between an abbreviation which you understand, and one you don't fully get - e.g., "e.g." vs BRDHD on a license plate - does the latter mean "broadhead" or "beard head" or "board head" or what? If it's on a surfer-themed jeep, you might guess at it being "BOARD HEAD" but you're still missing key information and having to guess at what goes in between, and maybe it's pretty obvious but maybe it isn't. And that's what happens with "lossy" compression, and how sometimes-critical details can be lost out of JPGS, if they're crunched down too low. OTOH, if you save a JPG with no loss, then you may just have a huge file and no space/bandwidth saved at all. It depends on the individual image. But a good rule of thumb is, if there is a lot of color change, and a lot of variegation, then you won't get much savings with lossless compression, and how far down you feel you can go before it starts to look too bad, is something you will need to experiment. (This is why you should always make a backup copy of a picture, before you start playing around with it - it's too easy to forget, save something lossy and not be able to go back to the higher-quality again...)
TIFFs (from Tagged Image File Format) are (mostly) lossless and/or uncompressed at all, they can come in a couple different flavors but if you only work on the internet, you won't prolly deal with them ever. They are primarily (historically and ontologically) used for professional printing, and until recently you couldn't even view them in a browser - Quicktime has included a TIFF viewing plugin of late, but you don't want to use them in web pages, because they aren't made to work with browsers and HTML and vice versa. However, they are superior, as a rule, for prepress, and conversely, until relatively recently, you couldn't use anything but TIFFs when it came to pixel-based graphics in professional desktop publishing, there were plugins to allow Quark and Pagemaker to read JPGs etc but they were glitchy, and the results when you printed out one of these web-intended images on a printing press were...unreliable, shall we say? Sometimes you might get a negative image; sometimes you might just get big black blotches. A lot of times your system just crashed. So, now Quark and Indesign, Pagemaker's successor, can read other formats, but you're still better off in terms of quality and safety using TIFFs, if you're going to print - and not compressed tiffs, either, because the lossless compression mode, LZW, can't be read by a lot of printer engines. Either it crashes or you get a blank space where your picture should be, and that can be very expensive and frustrating.
And now, here's where it gets really insane - you can also save a millions-of-colors tiff as a "bitmap," because the word also means a type of file language that can be read by a lot of different programs (tho' not web browsers - unless saved as an .ico file! there's always an "unless") because it's one of the oldest formats around on personal computers (there's also a whole range of high-end professional graphics stuff that I'm not even getting into, Scitex/RAW etc, altho' I'll try to explain it if anyone needs to know...) and can be opened by, e.g. MS Paint. And it has its own compression possibilities, too! But it doesn't change (afaik) the quality to save something as a bitmap - the only thing is, bitmaps of this sort are REALLY bad news to try to print with, because you can't set the page-printing dimensions internally very well. They're kind of the "pidgin" of pixel-based file format languages.
And this is how the theory and the pragma collide, because we are incarnates working with physical stuff, but you have to be able to conceptualize it to know how to work with it, but "if you can dream it, you can do it!" is not always true - not even with an unlimited budget, tho' that makes it easier. Sometimes people just haven't figured out how to build a machine that can do something "simple", or trying to make one do one thing has unintended consequences elsewhere. Which is not just true of digital art.
So, a really quick little explanation of what is going on and why there are so many different types of graphics formats, and what they're for, and when to use/not use them, which hopefully didn't confuse everyone too much and left everyone with useful information. For comparison, here's the relative file sizes of a blank 100 x 100 pixel image - your standard icon/avatar, in the different modes:
Bitmap (pure b/w): 2k
Grayscale (256 possible shades): 10k
Index (256 possible shades): 10k
RGB (millions of possible shades): 30k
These seem very small, but they will add up quickly, if you multiply them on each side!
-Okay? Everyone got their water, anti-malarial-tablets, pocket-knife-with-the-implement-for-getting-stones-out-of-horses-hooves, spare clothes, and towel? Now we can get to the good stuff - next lesson will be about Alpha Channels and Layers, although we may make side excursions into whatever strikes my fancy. We'll learn how to go from this to this, which will neatly cover both layers and channels, at least some aspects of both.
*Mostly, but we're not going to talk about the rectangular ones now because they are an abomination unto the Production Department, and I don't know why you get jpgs with them sometimes, only that they are a) distorted and b) Evil.
**This is not how colors are actually written, this is just a "natural language" approximation of it - if you want to see what it actually looks like open up a JPG in a text-editing program like Notepad...but save first, because it can sometimes crash your system. There are a LOT of little letters and numbers in even a small picture!
Special Lesson Bonus Link
Some cool free tiling filters (Win only, sorry).
A free "Gothic" effect plugin for MAC/Win here