Scanned Images With MECCA 2000

Transcription

1 MECCA 2000 User Group Meeting Scanned images within Mecca 2000 Scanned Images With MECCA 2000 Photographic images may be manipulated to the proper image type within common desktop applications such as GIMP for FreeBSD or Adobe PhotoShop for the Mac and Windows platforms. Illustrated are several versions of image types created from a single source CYMK color photographic image. 1 Artwork was first converted from CYMK to grayscale, then secondary conversion to B/W image with a 50% threshold for bi-level appearance. 2 Artwork was first converted from CYMK to grayscale, then secondary conversion to B/W image with a diffusion pattern dither to maintain rough photographic appearance. 3 Artwork was converted from CYMK to grayscale allowing highlights, mid-tones and shadows to fully remain. 4 Original artwork file, saved as Desktop Color Separations (DCS). Individual Cyan, Magenta, Yellow and Black channels converted into Mecca 2000 as single composite image allowing all previous color correction to be maintained. Ideal file type for CYMK Process color printing. 5 Artwork was converted from CYMK to RGB color mode (16 million colors +). Ideal file for display on screen or web-browsers. Not recommended for use with CYMK printing. Doing so will dramatically overcompensate for missing black values causing image to print too dark on press. 6 Artwork was converted from CYMK to 256 Indexed colors. 16 million colors compressed down to 256 colors maximum within embedded color palette. Ideal file for display on screen or web-browsers. Not recommended for use with CYMK printing. Doing so will dramatically overcompensate for missing black values causing image to print too dark on press. 7 Artwork was converted from CYMK to 128 Indexed colors. 16 million colors compressed down to 128 colors maximum within embedded color palette. Fewer colors available for display cause posterization effect within the image. Ideal file for display on screen or web-browsers. Not recommended for use with CYMK printing. Doing so will dramatically overcompensate for missing black values causing image to print too dark on press. 8 Artwork was converted from CYMK to 32 Indexed colors. 16 million colors compressed down to 32 colors maximum within embedded color palette. Fewer colors available for display cause posterization effect within the image. Ideal file for display on screen or web-browsers. Not recommended for use with CYMK printing. Doing so will dramatically overcompensate for missing black values causing image to print too dark on press. 9 Artwork was converted from CYMK to RGB and saved as JPEG image format with Maximum Compression. JPEG uses a form of compression referred to as "lossy". This type of compression discards similar color range pixel information each time opened and saved. Repeated opening and saving of the image file degrades quality of the image each time. Maximum compression stores the largest amount of color information, thus causing the least amount of image deterioration. 10 Artwork was converted from CYMK to RGB and saved as JPEG image format with Medium Compression. JPEG uses a form of compression referred to as "lossy". This type of compression discards similar color range pixel information each time opened and saved. Repeated opening and saving of the image file degrades quality of the image each time. Medium compression stores approximately half of the amount of color information and discards the rest. 11 Artwork was converted from CYMK to RGB and saved as JPEG image format with Low Compression. JPEG uses a form of compression referred to as "lossy". This type of compression discards similar color range pixel information each time opened and saved. Repeated opening and saving of the image file degrades quality of the image each time. Low compression stores the minimal amount of color information and discards the rest. Low compression images tend to appear very posterized. February 21-22, 2001 Amgraf, Inc.

2 Graphical E Wish your graphics files would magically transform to meet all your application needs? Won t happen. But there are workarounds. by Steven J. Vaughn-Nichols very time you start a graphics desktop publishing (DTP) project, it seems you need to deal with a new graphics format. It wouldn t be so bad if you could just use one master graphics program, but that s usually not the case. Your scanner saves an image in TIF, your paint program in PCX, and your desktop publisher turns up its nose at anything except EPS. We won t even talk about dragging in Amiga s IFF or Macintosh s MAC format. It isn t for a lack of standards. Several American National Standards Institute (ANSI) standards have been proclaimed and put into effect. Like many laws, these have been more honored in the breach than in the observance. Simply put, every software vendor believes that only it knows best about how to handle graphics in a particular application. Vendors are paying only token attention to ANSI or any other Standard. THE RASTER FACTOR There are technical reasons for this mess as well. High-quality graphics are difficult to capture. There are two ways of taming images: raster (a.k.a. bitmapped) and vector. Naturally, they re as compatible as oil and water. Raster formats are the most common. You usually find them with scanner or paint software. An image in raster format is an exact recording within the limits of the system s graphic resolution capacity. Raster programs record each pixel each small dot that builds up into an image on your screen. It doesn t take a rocket scientist to see raster s root problem: There are thousands of pixels in even low-resolution displays. By the time you get to preserving a 1,024x768 image with 256 colors, you re taking 1MB plus. The way to lick that problem is to include a data-compression routine in the graphics program. Data compression is the chief reason for the existence of a number of graphics formats. (See the sidebar, Compression Lesson. ) There are several ways to compress data, none of them ideal for every situation. So every company making this kind of software uses its own image-compression home-brew, and we end up (surprise!) with totally incompatible files. VOILÁ VECTORS Vector formats are used primarily by computer-assisted design (CAD) and charting programs, In these formats, images are preserved by recording their points, lines, and shapes. You may recall from high school geometry that the Greeks believed they could map UNLIKE VECTOR PROGRAMS THAT STORE IMAGES IN COLLEC- TIONS OF LINES, RASTER IMAGES ARE STORED AS PIXELS THOSE SMALL DOTS THAT BUILD UP INTO AN IMAGE ON YOUR SCREEN. any natural shape using only a straightedge and compass; today, vector formats prove their point. A vector program records only information about the shape and color of an image. Instead of recording every dot on a circle, vector software recalls that the images starts at one place, ends in another, and is a circle. You could say that the vector method makes a blueprint, while the raster method makes a photograph through the fine mesh of the system graphics resolution. Both methods have their advantages. Raster is better at fine details, while vector excels at making lines. Vector produces smaller files. Raster images, on the other hand, display faster. That s because the vector method forces the CPU, or graphics chip, to draw its image instead of a oneto-one mapping. The upside of this is that vector images are far easier to manipulate than raster images. As you may have guessed, there have been attempts to marry the two techniques in one product. Few programs, however, have had some measure of success. And those that succeed often spawn new proprietary formats in the process. TRANSLATOR, PLEASE! Programs that convert vector to raster are called translators. Some file translations are relatively easy. Between the graphics-file header and a software graphics-file library, good translation programs understand how X program encodes Y pixel and how to translate that into another format. It s an entirely different game, however, when going from raster to vector, vice-versa, or even from vector to vector. Vector formats use unique mathematical models to describe graphics primitive elements, like circles and squares. This is why, for example, there are so few good, reliable translators for such desirable format translators as PostScript to Hewlett-Packard Graphic Language (HPGL). Few programs manage a raster-to-vector conversion, and most of the ones that do are higherend. CorelDraw ($595), for instance, has a module called CorelTrace that performs this. 610 OCTOBER 1994 COMPUTER SHOPPER

3 TOOLS OF THE TRADE If you can t afford the price of a high-end image editor, you ll have to remain content with graphics utility programs like the $169 Hijaak Pro for Windows and North Coast Software s $149 Conversion Artist. Both of these programs and others like them are chock full of useful miniapps for graphic editors, including graphics-format and vector-to-raster conversions. Conversion Artist is noted for a great selection of raster file formats, including formats from non-x86 operating platforms such as Amiga, Macintosh, Sun, and Unix. However, Hijaak Pro stands out from everything else due to its extensive support of both raster and vector file formats it cannot, however, perform raster-to-vector conversions. If you ve just got to switch between raster and vector formats, and can t afford a high-end conversion package, check out Deltapoint Inc. s $ Graphics Tools package. This bundle not only converts between raster and vector formats and back again, but handles color separations, screen captures, file conversion, file management, and much more. Many graphics programs have strong points that aren t duplicated in other systems. If you happen to need a particular image-editing or enhancement ability, you may have no choice but to use a foreign program or utility. If this becomes your lot, just make sure that the file format in which you re most likely to work is supported by any graphics editing package you buy. To help you make these selections, here s the lowdown on the graphics formats you re most likely to encounter. BMP Bitmapped (BMP) files are Microsoft s contribution to the graphics-format alphabet soup. The name stems from the fact that it s the default filename extension for a Windows device independent bitmap (DIB) file. DIB files specify their colors in a form independent of any method used by a display to represent color. This lets Windows display the file on any screen. Of course, the reliability of the reproduction depends on the color resolution capacity of the display. The problem with BMP s is that as DIB s, they take up a lot of room. A 1,024x768 by 16-color BMP can take up half a megabyte, and that s with BMP s 8-bit-per-pixel built-in run-length-encoded (RLE) data compression kicking in to squeeze down the image. On the other hand, just about any Windows-compliant program is capable of using this file format. CGM Computer Graphics Metafile (CGM) is a popular vector format that also can include raster bitmap encoding. The National Computer Graphics Association decided in 1988 that CGM should be the graphics-transfer format of choice. CGM is also ANSI standard ANSI X With all this official backing, CGM is often used as a transfer format by programs like Aldus Page- Maker and Harvard Graphics. That s the good news. The bad news is that while CGM was meant to be device-and operating-system-independent, in reality, vendor s have improved their own house brands to the point where CGM files are no longer portable. The biggest fly in the CGM datatransfer soup, however, is Apple. The company hasn t seen fit to support CGM, so don t get your hopes up for crossplatform graphics unification. DXF Do you use CAD programs? Do you think you ll ever want to import from or export data to another CAD system? If you ve answered, Yes to both questions, be sure your CAD product supports this vector format from the grandfather of CAD programs, AutoCAD. EPS Encapsulated PostScript (EPS) files use both bitmapped and vector techniques for their desktop publishing tricks. EPS is an attempt by Adobe Systems and Altsys to create a way to let programs deal with PostScript images. PostScript is Adobe s ASCII pagedescription language. A PostScript file is really a device-independent program that PostScript printers run with a built-in interpreter to output a finished image, usually a page. Because PostScript is a language, you can also translate it into non-postscript printer code by using a Compression Lesson O ne of your worst headaches while managing high-end graphics is going to be storage. To combat the massive size of some of these files, developers have come up with a number of data-compression schemes some optimized for graphic files. Be careful, however. While shopping for such compression utilities, don t confuse them with the rash of new full-motion graphics-compression schemes. Here s a quick glossary that should prove helpful. DVI: Digital Video Interactive (DVI) is IBM and Intel s horse in the imagecompression derby. DVI is a term that is beginning to have its meaning worn off from overuse. It refers not only to compression algorithms, but to the compression hardware technologies the two computer giants are using to bring multimedia to PCs. Two compression algorithm, Real Time Video (RTV) and Production Level Video (PLV) make up the imagecompression side of DVI. JPEG: This stand for Joint Photographic Expert Group. Certified by an ISO and CCITT international-standard committee, JPEG uses a one-pass, three-fold process to squeeze down images. First, an image s red, green, and blue (RGB) color values are preprocessed into its luminance and two chrominance (YUV) values. Here, the gray-scale values are stored in Y, or luminance, while color values are stored in Y and V. Although considered preprocessing, this stage begins compression by disregarding every other pixel s chrominance value. Next, a two-dimensional adaptive discrete cosine transform (DCT) array is determined for every 8x8-pixel block. This stage essentially averages out the YUV value for each block. PostScript interpreter program like GNU s GhostScript (freeware). You can even use PostScript to drive a display as NeXTStep does. There s little difference between ordinary PostScript and EPS. In essence, EPS provides a way for applications to manipulate images without working directly with PostScript code. Optionally, From here, JPEG uses sampling techniques to transform the two-dimensional array into a one-dimensional array. The last phase, Huffman encoding, shrinks pictures into their final compressed form. Lossless: This refers to imageand data-compression applications and algorithms, like Hoffman encoding, that reduce a picture s size without losing any data. Lossy: Methods of image compression, like DVI and JPEG, that reduce the size of an image by discarding some pictorial information are lossy. LZW: Lempel-Ziv-Welch (LZW) is a popular and robust lossless datacompression algorithm. MPEG: The Motion Picture Experts Group (MPEG) is, like JPEG, both a group and an algorithm. The algorithm deals with compressing moving images. MPEG is enjoying immense popularity among multimedia developers right now, but its position as motion-video compression ruler was recently challenged by SPEG. SPEG: This is the Studio Photographic Experts Group s format for moving image compression. This Sony-backed standard is still a work in progress. Unlike MPEG, SPEG lets users edit video images within the compressed format. With this advantage, SPEG has a real shot at capturing the moving-image market. VQ: Vector quantizations s (VQ s) claim to fame is that it uses only a single memory access per pixel in decompressing images, making it theoretically the fastest of the full-motion image-compression tools. Of course, like most of these new compression schemes, it s entirely proprietary and incompatible EPS implementations can translate Post- Script code into a fairly standard graphics format. For Macintoshes, this is in the form of PICT files. On PC s, it s usually either as a Windows MetaFile (WMF) or a TIFF image. Unfortunately, while EPS is both powerful and popular, programs that imple- (continued on page 616)

4 Graphical Voodoo (continued from page 611) ment it lack file compatibility. For example, some programs support the Desktop Color Separation (DCS) extension for 24-bit color in EPS, while other programs freeze on opening such an EPS file. One product s EPS file is another s garbage. Let the buyer beware. GIF CompuServe wanted to help its users transfer graphics, so it came up with the Graphics Image Format (GIF) a raster-based format. GIF uses the Lempel-Ziv-Welch compression algorithm to shrink files to modemmanageable sizes. GIF doesn t define so much a graphics format as it does a deviceindependent data-transfer protocol for multimedia file transfers. Today, multimedia GIF files include MDX and WAV sound information as well as traditional graphics. Few business applications use GIF. Still, it s probably the format that people who don t work with graphics know best. Despite its popularity, though GIF hasn t made much of an impact outside of the graphics-file hobbyists. HPGL Hewlett-Packard Graphics Language (HPGL) files are in a vector format. HPGL/2 is the main language of today s CAD/CAM applications. When you re in the market for a CAD/CAM program, be certain that it can work with both DXF (see above) and HPGL. If you plan on sending your chart or engineering drawing to a plotter, HPGL is still your format of choice. Why? Because Hewlett-Packard is as big a standard in the plotter business as it is in the laser printer arena. IGES Initial Graphics Exchange Standard (IGES), a vector-based arrangement, is an ANSI standard with relatively few supports. Three-dimensional CAD software often uses this format. IMG Image (IMG) is a bitmapped format used in the GEM operating environment. First made popular by GEM s Ventura Publisher, still the number-two desktop publisher, IMG is becoming an obsolete standard as GEM-based applications continue to lose market share. PCX This format originated from Zsoft s PaintBrush program, but is has spread to many other raster-based programs. While PCX is a proprietary format and not an official standard, many programs including almost every DTP program, use it. Unfortunately, the first definition of PCX hasn t proven adequate for today s more-demanding displays. The newest version, for instance, supports 24-bit color, but such images are completely unusable by older PCX programs. PIC This format is a real problem child because there s more than one of it out there. To spreadsheet fans, this is Lotus s vector format. It s crude and ugly, but it works, and most business programs understand it. Supporters of paint programs like Dr. Halo and PC Paint know PIC as a raster format. It s also used by screencapture programs like Grasp. Need I mention that these identically named formats are incompatible? I thought not. TIFF Aldus and Microsoft joined forces to produce the raster Tagged Image File Format (TIFF). It was designed to solve the technical problems of trading images from Macs to PCs and back. In particular, TIFF was meant to clear up the confusing array of scanner file formats that existed at the time. Because of this, there are actually five common types of TIFF files to deal with the different requirements for scanned images. These are: Class B black-and-white images; Class F fax images; Class G gray-scale images; Class P palette-color images that use an internal color palette to define 256 colors; and Class R ^photo-realistic images that record each pixel s red, green, and blue characteristics. Optical character recognition and DTP programs most frequently use Class B images. Class F, of course, is for faxing, The next classes, G and P, are used in DTP, while R is used for high-end color DTP. TIFF is fairly successful, but you pay for that success with very large files, In the case of Class R, these files can be positively gargantuan. TIFF also has the problem that, like CGM, there are dialects of TIFF that not every TIFF compatible program can easily understand.

5 Scanned images within Mecca B/W Line Art 50% Threshold MECCA 2000 User Group Meeting 2 B/W Line Art Diffusion Dither 3 Grayscale Photograph 4 CYMK DCS Color Photo 5 RGB TIFF Color Photo Color Photo Color Photo 8 32 Color Photo 9 Max. Compression 10 Med. Compression 11 Low Compression All Images are 150 dpi Amgraf, Inc. February 21-22, 2001