Previous Article WWWiz Home Next Article
Cover Story

 LizardTech: Cutting Web-Page Image and Document Publishing Problems Down to Size

 By Tom Bunzel (tombunzel@wwwiz.com)

 When WWWiz asked me to write an article on LizardTech and their Web graphics software, I said to myself, ``Oh boy, just what I need, another image editing program."

             So I went to the Web site, downloaded and installed the free version of LizardTech's MrSID Solo.

             I thought to myself, ``Great, a reboot–must be a plug-in or funky Active-X component."            

Then I opened MrSID.  Very sparse interface–basically just a way to import a file and ``encode." 

            ``What the hell does that mean?" I thought to myself, since I hadn't bothered to read any of the documentation.  I figured I would load in an already compressed JPEG file and see what MrSID could do–not expecting a whole lot.

             I looked at the results in My Computer and not surprisingly, it had created a proprietary file format called, not surprisingly *.SID. But what instantly intrigued me was that it was one-tenth the size of the original JPEG file.  I double-clicked on it to see what would happen.

             The image opened in my browser and looked every bit as good as the original, not bad for 10% compression of a JPEG file.  Then I noticed something else–I had a new toolbar in Internet Explorer that let me zoom and pan around the image.  I zoomed into an area and found remarkable clarity. This was pretty cool.

             The next thing I did was download the MrSID plug-in for PhotoShop, which lets me open, view and encode the proprietary format from within that program, which is still more or less the standard for high-end image processing.  The pair make a nice combo.

             Of course LizardTech is not the only company offering proprietary file compression and Web distribution. Live Picture published FlashPix a while back, even before it became part of MetaCreations.  Now the technology is owned by MGISoft, which has made it part of its ZOOM plug-in solutions suite.  This is another viewer for specialized highly compressed images, but a bit different from the LizardTech version.

             MGISoft sells a Reality Studio package that creates images that you can zoom through (panoramas) and it enables the end user to create object movies, although you still need a java applet or plug-in to view them.

             And Apple's software tools include the Quicktime VR toolkit that creates the same types of files–panoramas and object movies.

             So what is the end user or Web developer to make of all of this?

             It really depends–as always–on what the application is. If you are a home user and want to put some photos on the Web, you can use the utility program with your digital camera or a light version of Adobe Photoshop (PhotoDeluxe) or an inexpensive bundled tool (Ulead PhotoImpact). Crop your image, save it as a JPEG and you're ready to publish.

             But what if you are an e-commerce developer or someone with a specific niche application–like an interior designer, travel agency or real estate agent with a need to show large numbers of very high end images to many potential viewers with the best possible quality?

             Now which tools do you choose?  And what demands are you willing to make on your end user–in terms of plug-ins or other add-ons or enhancements–to let them take full advantage of your end product?

             We can break the current Web imaging tools into four main areas:

 Image editing and processing

  1. Image publishing
  2. Document processing and publishing
  3. Mass distribution for intranets, e-commerce and other commercial ventures

             Of course, there is a lot of gray area among these categories and one can argue about distinctions among these various fields, but they are helpful in an overall analysis.

 Image editing

             First there is the basic image editing product line: still led by Adobe Photoshop with Macromedia Fireworks nipping at its heels. Many other competitors have entries, which generally let the user employ a TWAIN compatible acquisition device (scanner or digital camera) to import an image, crop it, apply filters and other editing and manipulation processes, and finally save the image in a Web friendly format:  GIF, JPEG or PNG (Portable Network Graphics).

             The obvious advantage of these formats is that they display natively within a browser without a plug-in, Active-X control or java script adjustment of any kind.

             Basically, the first level graphics tools grab an image from the real world and process it for the Web.

 Image publishing

             Then there are the image publishing programs, such as Photoshop and Fireworks, which can slice images into subparts, create rollover effects and other enhancements that provide functionality within Web pages.

            Director and Flash are examples of publishing tools that add animation capability and compress with Shockwave. They also publish their files with a corresponding HTML template to play back the resulting file.

            Microsoft FrontPage is a publishing tool, which can use its PhotoEditor as the graphics-editing component. In terms of specialized compression solutions, Reality Studio and MrSID also fall into this rapidly growing and changing category.

            Reality Studio on the PC and Quicktime VR on the Mac convert images into panoramas and object movies.

             MrSID takes a compressed JPEG file and compresses it even further–letting the viewer pan through it and zoom into it without significant degradation of the image.  Photos can be converted from the Photo Solo version, from inside PhotoShop version 4.0 or later with MrSID Photo for PhotoShop, or in a MrSID Photo Workgroup for multiple users of high volumes of images for applications like a digital catalog for e-commerce.

             By examining the acronym that comprises MrSID, we see where these programs are headed: indexing images and documents for database retrieval, online and over networks.

 LizardTech licensed Generation I of its patented MrSID (Multiresolution Seamless Image Database) software from Los Alamos National Laboratory.  The software was originally developed by the same team that created the FBI standard for storage and transmission of its large fingerprint image library.

 In February 1997, the Library of Congress became LizardTech's first MrSID customer.   In essence the product sets up a storage strategy–like a card catalog–for massive amounts of imaging data.

 Since that time, LizardTech has established global relationships with a network of high profile strategic partners and customers including Autodesk, Adobe Systems, ESRI, UCSF-Radiology, Graphic Arts Center, Space Imaging/EOSAT, NIMA, Net Libraries, Getty Images, Boeing, U.S. Navy, Photo Science Inc. and hundreds of others. LizardTech's technology products are on more than 15 million desktops and integrated into over 200 applications.

 Recently LizardTech announced adoption of its MrSID software in a solution with other third parties for the L.A. Sheriff's Department to spot illegal gunfire, and pinpoint its origin using aerial photography.  MrSID sends officers in the field a map and actual aerial photographs of the neighborhood where the gunfire went off. Dispatchers can also use the GIS-based interface to perform spatial inquiries to quickly extract phone numbers of local residents (30,000 names and addresses per minute) within a defined area around an event.

 Document processing

 As noted above, images aren't the only Web elements that can benefit from compression technology; there is an entire field of document management that has been dominated online so far by the Adobe Acrobat (PDF) file format.

             The advantages of online document management and distribution are obvious. Applications include help files, marketing materials and a host of other files that combine text, graphics and hyperlinks to enable users to find relevant information quickly through indexing and search tools.

             Many end users have experienced the benefits of this kind of indexing with their home scanners. Software such as Visioneer PaperPort, now owned by ScanSoft, provides an iconic desktop for digitized documents. Some can be dragged to MS Word for OCR (optical character recognition) to be re-used as text documents, others go to an image editor, some to fax, some to printer, etc.

             Hewlett-Packard bundles this sort of document management tool with many of its scanners and printers; it also has a proprietary Document Manager.

 It should be noted that a little more than a year ago, ScanSoft acquired Caere, a major player in the desktop OCR/scanner software market.  Recently there have been announcements from the company about agreements with Microsoft that would have the ScanSoft document imaging technology integrated into the next version of Microsoft Office.

             ScanSoft is promoting a technology known a TIFF-FX as an emerging standard for Web document distributions. It will employ three new compression technologies: Mixed Raster Content (MRC), JBIG2 and JPEG2000.

 The Mixed Raster Content (MRC) method breaks the original image into multiple constituent images and compresses each constituent image in a way optimized for its particular nature.

 This is illustrated below (from www.scansoft.com).

 MRC is a compression technology for Color Fax, while JBIG2 is a method of representing bi-level (i.e., black and white) images. It has been adopted as an international standard by the International Standards Organization (ISO) and the International Telecommunications Union (ITU) (for info on JPEG2, see http://www.jpeg.org/public/jbigpt2.htm).

 For scanned images involving text, JBIG2 is typically three to five times smaller than compression methods currently used by fax machines.  JPEG2000 is a new standard for encoding pictures that gives better picture quality for a given size than does JPEG, especially at high compression ratios.

 LizardTech's DjVu Solo 3.1 is another intriguing online version of this type of software that enables highly compressed documents to be shared on the Web or over an intranet, at download times up to a 100 times faster than PDF. (LizardTech acquired its DjVu technology from AT&T Labs for the commercialization of products in the private sector).

 Like Visioneer PaperPort and the other ScanSoft tools, DjVu Solo works with any ``TWAIN compatible" scanner, meaning one that Windows recognizes and interfaces with.  In DjVu as in Acrobat Publisher hyperlinks, annotations and other enhancements can be added to any document.

             The commercial version of DjVu Solo also creates OCR-enabled documents that can support keyword searches, and the format has the same high level of image clarity as MrSID–along with the capability of zooming, panning and enlarging the images.

             The commercial version runs $299; the personal edition is free for non-commercial use but does not include the OCR-text search feature.

             Both require the free DjVu browser plug-in. The target markets include small businesses that need to organize their work product, the medical and legal industries that have massive paper management issues, financial institutions and any business that needs to distribute documents online or over an internal network efficiently and with minimum bandwidth and overhead.

             Corporate customers will use the DjVu Enterprise edition 3.1 to encode scanned paper documents into the program's proprietary, searchable format.  In addition to OCR, the enterprise edition features a command-line interface and batch processing to integrate into a company's workflow and efficiently convert large numbers of documents.

             Collaborative efforts are supported by DjVu Workgroup, which lets five simultaneous users customize a set of DjVu documents.  This version supports drag and drop functionality, so that files can be selected from a folder, their features customized and then dropped into the proper folders for conversion and subsequent viewing online or over a local network. The key obviously is that again the browser plug-in is present for all viewers.

             In the adjacent diagram [djvu_add_hyperlnk] a document has been scanned from an HP scanner into the Solo version, saved as a proprietary form, and an area selected as a rectangular hyperlink.  In the dialog boxes the choices include going to another page, or a Web URL.  Notice that the thumbnail view is also available in the pane on the left.

Mass-market images

             Now we come to the mass-market web image and document servers.  Again Live Picture started this area with FlashPix, which required a plug-in that let users zoom and pan through images.

             The idea of FlashPix was that from a large source image, only the portion requested by the browser is downloaded piecemeal.  Then the plug-in intelligently queried the image database for more information as the user requested it.

             LizardTech software is also making inroads in this space.

 MrSID's server component is part of the Web site for the U.S. Geological Survey's Energy Resources Program.  The Energy Resources Program has developed the GEOData-Explorer that retrieves information from a database or databases, organizes the information spatially, and allows the user to more effectively manage a situation that is dependent on such data.

 MrSID imagery in the form of digitized aerial images has been incorporated into the display of the Powder River Basin in the Montana/Wyoming project area. Users can zoom into the outlines in order to view the multi-resolution MrSID images.  The Web site is http://dss1.er.usgs.gov

             LizardTech's Content Server 3.0, with a per-CPU license fee of $4,500, enables IT professionals to deliver information to customers and business partners over the Web.

             The growing customer base for these applications includes corporate digital asset collections (Madison Avenue and Hollywood), real estate, online catalogs, auction sites and retailers, libraries, medical sites and the aforementioned geospatial imagery and government agency applications.

             While in most cases the Content Server will deliver DjVu or MrSID format versions of the images requested by the browser, there is also the option to send image content for viewing without a browser plug-in. ``Content Server wraps it as a JPEG on the fly," says Joe Tradii, LizardTech product manager, ``ensuring that every user, even those with older Web browsers, can view the images."

             What's neat about Content Server is the ability to have ``one size fits all"–the host does not need a separate thumbnail and different resolutions for each image–and the one source image is accommodated by all client applications and just the right size and part of the image is sent down the pipe.  The source image can be the size of a poster, and yet it can serve up the proper dimensions and size for a Windows CE client or even a handheld PDA.

             Corporate customers for the DjVu technology include Corporate Reports, based near Leeds, England, which since 1997 has provided quick and easy access on the Internet to the latest financial information on UK listed companies.  The company sought to maximize its return on the effort required to scan documents for online distribution, and determined that the DjVu solution was more economical than Acrobat PDF files.  The firm estimated the savings at $3 million.

             ShopeCatalogs, as its name implies, puts electronic versions of catalogs online, including the Sharper Image.

             Obviously, preserving rich colors and details are key, but since catalogs often contain numerous pages, file size and download speeds become critical factors. Full-color documents scanned at 300 dpi (a level required for color pages) often become at least 20MB in raw TIFF format. If the image is saved as a JPEG, the common image format for the Internet, the file size drops but results in dramatic image degradation, especially when zooming in. Creating catalogs in the familiar Adobe Portable Document Format (PDF) results in a single page several megabytes in size, much too large when working with multiple pages of color catalogs.

             ShopeCatalogs chose the DjVu format for its ability, among other things, to strip the text layer from the image layer, and enable text searches and hyperlinks with image clarity and no loss of quality.

             Along similar lines, Cobb County, Ga., is now distributing its public documents with LizardTech solutions, and McLeodUSA distributes its scanned network construction prints with MrSID.

 Where is this technology headed?

As broadband becomes more widely available, one might think that the appetite for compression of images and documents would abate.  But dial-up is still how most Web surfers get their data.

             And don't forget bandwidth is still a major concern in the wireless device arena, where highly compressed images are crucial.

             According to LizardTech, ``XML- and XSL-based stylesheets provide the tools needed to dynamically display content for any viewing environment or device–so that a growing set of devices including PDAs and cell phones will be able to access these image libraries."

 At a recent technology show in Seattle, Lizard Tech showed how it optimizes its MrSID and Content Server product lines for Intel's XEON processors to provide wireless device manufacturers with advanced imaging capabilities. 

 The Intel XScale microarchitecture, introduced in August, maintains low power consumption for handheld devices.  Diane Wortsmann, marketing manager in Intel's Solutions Enabling Group explains,  ``By utilizing the advanced features of servers based on Intel's Pentium III XEON processor to encode and serve large images, LizardTech will be able to provide customers with outstanding capabilities as they develop their wireless and e-commerce products and services."

 LizardTech is privately held and recently obtained a $25-million round of financing from Mitsubishi.  In 1998, the company was voted the Washington Software Alliance's most promising new company. LizardTech's investors and shareholders include Oak Investment Partners, AT&T, Encompass Ventures, SeaPoint Ventures, Kirlan Ventures, Summit Ventures, Shurtleff Group and Staenberg Private Capital. Additional information about LizardTech and its full line of software solutions can be found at www.lizardtech.com.

             This is an admittedly brief overview of the ever-evolving image/document processing and publishing field.  Other programs and software publishers have competing solutions, but bear in mind that the overall functionality of what is being offered should be comparable.

             Use the information as a reference point for your own research into the various solutions for your own personal or business needs.

 Tom Bunzel works as ``Painless PC," a consulting and training facility in West Los Angeles, specializing in business, presentation and Web-authoring applications.  He can be reached at (310) 286-0969 or tom.bunzel@painlesspc.com

 They'll Help With Your Image

 A Partial List of Image and Document Processing Software Companies and their Web sites:

Adobe

PhotoShop

PhotoDeluxe (bundled with many peripherals)

Acrobat Publisher (PDF)

http://www.adobe.com/

Apple

Quicktime VR Toolkit

http://www.apple.com/

 

LizardTech

MrSID Solo

DjVu

Content Server

Enterprise Server

PhotoShop plug-in

http://www.lizardtech.com/

 

MGISoft

Reality Studio

ZOOM Plug IN

http://www.mgisoft.com/

 

Microsoft

FrontPage

PhotoEditor

http://www.microsoft.com/

 

ScanSoft

Visioneer PaperPort and Desktop

Caere OCR

http://www.scansoft.com/

 

Ulead

PhotoImpact (downloaded and bundled with many peripherals)

Album

http://www.ulead.com/

 

horiline

Copyright (C) 1998 WWWiz Corporation - All Rights Reserved
Phone: 714.848.9600 FAX: 714.375.2493
WWWiz Web site developed and maintained by GRAFX Digital Studio

Previous Article Next Article
WWWiz Home