Basic Digital Graphics Concepts
(restored and looked at in August of 2018, touched up 3/03/2006, partially revised 4/25/03 from the 1997 slide-show original)
Hi --I'm Craig and this "show" is a presentation of "the world of digital graphics" --from WAAAY back, and from my own peculiar and jaundiced points of view  :-)

Micrografx was once the Cadilac of PC graphics programs, so much of my work is still created with their 1992 "Picture Publisher" and 1994 "Designer" (vector graphics --integrated with the bit-map graphics of PPub-3.1). But: I also have and use recent Mgfx programs plus several of those more recognizable name plates like: Adobe and Paint Shop Pro plus ray tracing programs.

There are many and confusing concepts involved in this realm of endeavor, plus the field of topics and format types is being added to by frequent software developments.  The history on these programs and their outputs  has been pretty much defined by what's technically possible and practical.

Whatever you decide to use, it's important to run at least one newer program which keeps abreast of new formats and the translation between them.

Let's start with image file types (or: "formats"). In a way, all file types are image files, since the consequence is some kind of a visual presentation, but we're concerned with those that contain specific shape information --and often other attributes like color, brightness, and included masking information.

Fonts are a special class in that they're resident as a set of vector graphics, either in the program, the printer, or both.  They're often an essential part of technical work and can play havoc when moved to another program that doesn't have them in its repertoire  Odd fonts must be separately supplied when a work is given over to another company for printing or other work.  The most common and trouble-free fonts are "True Type", which scale smoothly and give a good monitor presentation that relates closely to what will be printed.  I consider TT fonts to be "vector graphics" (see below).

Line Art is the earliest and most memory efficient graphic type.  Most graphic formats (such as "TIFF", "PCX", "GIF") support line art in "native" and compressed file sizes.  This is called "1 bit" graphics because a pixel element is either on (black, say) or off (white).  Any 2 "colors" can represent a line art graphic.  This is one of a family of "bit mapped" graphics, so called because its elements are plotted within a rectangular field of pixels (picture elements).  The number of pixels in an image and often the physical size of the field are specified within the file.

Grayscale art is a type of "8 bit graphics" since it contains 8 binary bits of information per pixel.  For gray scale images (black through white), this has been considered sufficient for professional grade reproduction, although more bits may be used to provide more reach or "color depth", perhaps to store shadow details that would otherwise be lost into "black".  (Scanners run 10 bits internally.)

This standard of 8 bits counts in binary fashion to 256 levels/shades of brightness (or gray) and came about because display devices (monitors) and printers can only reproduce densities within a range of about 100 to 1.  When that range is divided into 256 steps, the transition between steps appears to be a smooth continuum.  (A convention of "Post Script" printing is implicit in this.  See below.)

Confusion arises because although a file might contain an 8 bit smooth grayscale graphic, the presentation (an ancient VGA monitor display, say), printer output, or the parent document the graphic's placed in might only support 1 bit or 4 bit ("16 color") graphics. (Hardly ever seen now in the 21st Century.)

"Bit mapped" graphics contain only just so many pixels and resolution.  When a bit map is enlarged such that the resolution of the display or printer is significantly greater than that of the graphic itself, the image starts looking rough and jagged.

Another phenomenon common to all bit map graphics is that when they're manipulated for contrast and such, levels of gray are lost.  You can't stretch the tones too much before you start seeing the gray "steps" and the image goes grainy/motley.

Color bit mapped images are simply the combination of 3 "grayscale" images using --instead of white through black-- the primary colors of red, green, and blue.  They're consequently called "RGBs".  The professional level for color has been set at 3x "8 bit" for many years, though "deeper" formats are sometimes heard of and are used within graphics devices.  Obviously, the permutation of 256 shades (or steps) of R, G, and B yields 256 x 256 x 256 = 16.7 million different color possibilities.  This image "depth" is also called "24 bit color".  While "30 bit color" might refer to a scanner using 3 internal 10 bit color channels, "32 bit color" usually refers to the "subtractive" primaries cyan, magenta, and yellow plus the printer's black ("K") channel: "CMYK".  (Unless one is responsible for actual pre-press final image setting or "color separation" work, the independent graphics technician has no business using, specifiying, or adjusting the K channel.)

To create image files that are more manageable in size, the color content might be reduced to 64,000 ("64K") colors, which easily passes for 24 bit or "true color".  By carefully selecting and adjusting the "color palette", full color images are often reduced as low as 256 total colors with fair results (GIFFs).  Each image contains only those colors actually needed to reasonably reproduce the image.

Obviously, an old 16 color "plain VGA" monitor display will garbage up any "continuous tone" 8, 16 (64K), or 24 bit image.

(Unfortunately, recent graphics programs/manuals have taken to referencing 24 bit RGB color as "32 bit" --perhaps to one-up others using standard terminology.)

(Notice how this color image falls apart when enlarged too much.)

Vector Graphics constitute a large family that includes the many "3 dimensional" forms now being used to produce digital animation and rotateable technical illustration.  The basic idea is that all lines proceed from defined points, along defined angles, and for defined distances.  These lines can have 24 bits worth of color definition and many other attributes.  The most exotic form of vector graphics are within "ray tracing" programs which actually define a 3 dimensional environment, its light sources, atmosphere, and assign qualities of reflection, transparency, and refraction to surfaces defined by vector graphic lines and areas.

Simple vector graphics started out as "CAD" (computer assisted design) programs and solved the problem of creating large high resolution works with limited memory.  Their greatest advantage is that such images can be re-sized with no loss of resolution.  Their weakness is that they're devilishly complex in construction and you can't just take a graphic "eraser" to a section you want to trim off.  You CAN, however, distort such images as easily as bit map types.

While vector graphics take up little room on a hard drive or diskette and transfer quickly, they're math heavy and would often "draw" slowly (with old computers) when called up for display and manipulation.  Sophisticated "virtual reality" ray trace creations use to take hours of run time for a single display output, since an entire little universe has to be replicated and interpreted for a particular point of view from "within" it.

A multitude of creative graphics programs have evolved: from an assortment of limited purpose applications (that might not work well from one to another) to feature laden program "suites" that combine functions or at least smoothly switch between among a family of applications.

Traditionally, you'd have a graphics program (Mgfx Picture Publisher, Adobe Photoshop, say) that would work bit maps and offer everything from a poor scanner tie-in to many modes of image manipulation and a finished graphics file output.  This bit map final would be in only one of several formats (TIFF, TARGA, PCX, GIF, BMP and such), but most programs now give you a choice of nearly all formats, both for import and export plus compression options.

That graphic could then be taken into many other applications:  graphics capable word processing, a "designer" type program (Adobe Illustrator, Corel Draw) that can handle bit maps as well as the more usual vector graphics, or a publishing program (like Ventura Publisher or Adobe PageMaker).

Graphics services normally expect to see your image work "encapsulated" into a "page" from one of these programs (say: an "EPS" Encapsulated PostScript or an "AI" Adobe Illustrator file).  Recently, it's become easier (as well as more sensible) to just hand over the raw RGB bit map image file.  (A vector graphic work will be in one of the designer type formats, of course.) (Adobe "PDF" format has popularly replaced other formats for small scale commercial printing, but is seldom used for authoring personal desktop documents and graphics. A word processing format or a "Power Point" type show format is more common.)

Let's review graphic formats and practice one more time:

* Those of us who use a simple (comes-with browser, say) viewer often get ONLY a screen rez display.  Although you might conceive of your image as (say) 2.5" wide for each frame AT a resolution of 150 DPI, a simple viewer or web page display will present a pair of frames that are each 5" wide --and the person looking at it will have to scroll around.  (Even the rotten comes-with-Windows "Paint" program lets one change display scale, however.)

With a halfway decent graphics program, an image at 150 DPI often looks a lot sharper in its hard-edge details than it would (size of size) at screen (72 or 75 DPI) rez.

* When you create art work in "Paint" or a good Designer program, large areas are often made up of exactly the same color value.  "One bit" or "line art" is, of course, made up of either black or white pixels.  When such images are made as or converted to either the "GIFF" or "PCX" formats, they automatically compress --and compress VERY well --often better than if "JPEG'd".  When made as or converted to a "TIFF" or "TARGA" graphics document, they can still be saved in the compressed mode --which applies approximately the same kind of compression as is inherent in PCX and GIF files.  (This is called "LZW" compression.)

* Photographic type, or "Continuous Tone" ("CT") images often have next to no areas that are exactly the same color value.  Trying to "LZW" compress such images might reduce them a tad, and add a lot of delay time for reassessing them.

Instead, we usually JPEG such images when they need to be compact.  JPEG is a wonderful type of compression that, if you have a good graphics program, allows you a very wide range of choices for the degree and kinds of compression.  Often these choices are mostly made for you and you're presented with "Less Compression", "Normal Compression", or "More Compression".  I suspect that most of us will prefer "Less" compression, because "JPEG" is a "lossy" (not lousy) way to go.  The compression program tries to pick out stuff you might not notice --and throw it away to save space.

        > Once JPEG'd, the discarded information is lost, so save a TIFF version if the image is important to you.

        > If you bring up a JPEG'd image, then re-save it as a JPEG, the image takes 2 hits --and it always shows the damage.  Although "Less" compression might look identical to a TIFF, 2 successive "Less" JPEG compressions will make the image look motley --especially where hard edge details lie against smooth areas.  Always go to the first compression as your source --or save the opened image as a "TIFF".

The TIFF, PCX, and JPEG, formats are not always the same, compressed or not.  There've been a series of developments for most graphics formats and sometimes a TIFF saved by one brand/version of graphics program won't open right in someone else's.  That's where those converter/translator programs come in handy.  (I know you've got a good one.  I use "ViewPrint" here.)

Most graphics formats support both "True color" --meaning 24 bit color: 8 bits each for Red, Green, and Blue; AND "8 bit color" --which is also called "indexed color" and "256 color pallet".  GIFFs, however, are always 256 bit color, and JPEGs are either 256 grayscale or 24 bit color.  (Sometimes a graphics program can't access the grayscale version of JPEGs, and there's really little reason to use it.)

Obviously, if an image is only gray scale, it looks perfect as 8 bit "color" --because there's only one color: black-to-white = gray.



In the interests of attaching photos to e-mail, posting photos to web pages, sending photos to printers (perhaps as an "Encapsulated Post Script"/EPS or Acrobat PDF format file to a commercial printer), and having image utilities well within the contrast ball park. it might be necessary to trouble with your graphics card driver (assuming your computer has a discrete/separate graphics card, but it still has a driver).

I've paid very close attention to gray scale/tracking, contrast gain, and colorimetry equivalents, but strangely, and despite 13 years of computer graphics, I've seldom even thought about "system" gamma --or, rather: the gamma throughput to the screen of one's monitor. My graphics manuals counseled me to adjust my graphics program gamma (dynamic contrast) at 50%, then leave it alone --and that simply worked.

Lately, however, I've noticed significant disparities between the default gamma of my system (which is used for utility displays like e-mail image viewers and special image editing windows) --and the main editing window in my graphics programs --both old and new. Without much of an effort to research the why of it, the problem seems to have been where my monitor driver's gamma --or base contrast-- was set.

As to the regular monitor contrast control, one starts with that at 100% --and you leave it there. The regular brightness control is turned up until blacks begin to budge, then back a hair.

Next you set your office/studio lighting to where you'd normally have it.

At this point you can use the monitor gamma utility supplied with Photo Shop (see PS Helps). That sends a gamma order to your monitor driver (if it's up to date).

Alternatively, find your monitor driver's settings via Control Panel then Display and get on the gamma adjustment. Go to:

--and pick a gamma value to test with. The author argues persuasively for the Mac standard of 1.8, which value cured my graphics problems. He provides a link to a Java applet web page for to double check/confirm your average gamma setting (light to dark). Don't be surprised if that number disagrees with the gamma number you read out of your monitor driver settings, or the number indicated by graphics program utilities (including Photo Shop's).


Click *HERE* to continue