Journal: The Charleston Advisor
Author of Column: Larry R. Sheret, Program Coordinator Learning Resource Center, Central Arizona College, Aravaipa Campus
AccuNet/AP Multimedia Archive
Date of Review: 8/20/01
The Multimedia Archive is the most extensive source of photographs and sound-clips available online. Photos, sound-clips, news text and graphics are stored in four separate databases that are searchable from a single WebPage. The photo and sound archives are the most important components of the site since they contain files that reach back to circa 1840 and the 1920's, respectively. The photographic database contains 750,000 images and the audio database 500,000 audio clips. There are also 800,000 news text files that date from January 1997, and 13,760 graphics from January 1999.
Advanced searching (Boolean, truncation, wildcard, proximity, et.al.), as well as sorting of hits by relevancy or chronologically, are offered for each database individually. Simultaneous database searching is under development but is not yet available. All databases present a common user interface.
The cost is variable according to which of the databases are included in a contract. Subscribing to the photo and sound archives together provides the greatest value. Substantial discounts are given for site licenses. Small libraries are encouraged to contact the sales department to work out special pricing.
Institutional (based on FTE), site, district (user pools based on the population of the service area). Consortia may negotiate with the vendor using district contracts as a model. Everything from single user to remote access for unlimited users is available.
The AccuNet/AP Multimedia Archive consists of four separate databases with different formats: photo, sound, text, and graphics. The photo and sound archives may be subscribed to individually at the base price. For an additional 10% each, the text and graphics databases may be added. The cost of the entire suite comes to roughly two-thirds that of subscribing to each database individually.
Depending upon the type of library and the size of the user population, site licenses run between two to nine times the cost of a single user subscription. All site licenses include remote access.
The Associated Press (AP) was formed in 1848 when six daily newspapers in New York City collaborated on the installation of a telegraphic relay to transmit foreign news brought by ships docking in Boston harbor over 200 miles away. AP began adding audio content to its news services in the early 1940s. Leased wires delivered sound clips to radio stations throughout the country. Today, AP is the world's oldest and largest newsgathering organization, serving 15,000 news outlets in 71 countries.
Working in conjunction with AccuWeather, the world's largest corporate supplier of weather information and images, AP has made large portions of its photo, sound, text and graphics archives available to schools (K-12), public, and academic libraries. AccuWeather's mission is to supply copyright-cleared access to AP's professional news databases to subscribers for educational purposes over the internet, using an interface that is easily and intuitively understood by the general public. Four databases contain the various formats that comprise the multimedia collection.
This service, created for the public, students and educators, is available only to schools, public and academic libraries. Files of each kind may be downloaded, copied, or saved to a disk or hard drive. Although no specific interlibrary loan agreements are written into the subscription contracts, limited ILL is permitted since files may be downloaded and then e-mailed as attachments to qualified third parties. Students and educators may use files for papers or multimedia presentations. Educators may produce copies of files for students.
The Multimedia Archive claims an "up time" of 99 percent, and is available 24/7/365. During a hurricane, it managed to remain online by switching to a T3 connection. Although AccuNet is responsible for the development and management of the Multimedia Archive for libraries and schools, AP supplies the content, server storage and I-Net connections for the service. The only downloading limitations are on the user's end.
The PhotoArchive is divided into two parts. One contains all the news photos (approximately 800) that move across AP's spot picture system each day, including those photos that never make it into print or onto TV. After being held in the current collection for one year, about 75% of the photos are removed. Most of these are duplicates; the remaining 25% are indexed and cataloged for retention in the historical collection. In addition to these, roughly 200 photos from AP's historical archive undergo retro-conversion into digital format each day. Images in the historical archive date back to a circa-1840 photograph of Alexander Twilight, the first African-American to earn a college degree. Of the 50 million images in the AP archive, about 750,000 of the best ones are now available online.
At present, approximately 2% of AP's image library is available. The extent of the image database is evident when comparing it with other products. For example, a single volume of Britannica contains 1,000 pages. With roughly 750,000 images currently available online, it would take 750 volumes the size of Britannica to put these images into print form. Every librarian can appreciate what it means to have such an extensive collection available without the need to maintain hardcopy picture files. Moreover, the subject matter represented in the image database is as diverse as our user populations.
Users may choose to have image hits displayed in one of three ways. When the 4 thumbnails per screen option is selected, a full caption, written much like an abstract, appears next to each image. When the 12 thumbnails per screen option is selected, no caption appears. However, a hotlink below each thumbnail allows for instant display of the caption. When the index option is used, a listing of 36 photos is displayed. Although no thumbnails or captions appear, clicking on an appropriate hotlink next to each title causes a thumbnail or caption to pop up, or a photo to load.
The Audio Database is an archive of audio files of AP news clips that are at least 48 hours old. (Breaking news is only available to news outlets through another service called AP PrimeCuts). Launched four years ago as a technologically superior way to transmit broadcast quality news clips to news outlets worldwide, this same database is now available to libraries. The Audio Database was the final component to be added to the AP Multimedia Archive on June 28, 2001--the launch date of the complete set of resources.
The Audio Database is a collection of sound clips. It is radio in the raw. Most of these monaural audio clips are short--less than thirty seconds in length. To gain a sense of what is available on this database, imagine listening to a five-minute radio news spot on the hour or half-hour. About half of the allotted time is usually reserved for commercials. In the remaining time, five to ten news clips are aired. Each clip may be as little as ten seconds long, and rarely as long as sixty seconds. The collection has over 500,000 sound bites. This number grows daily as three-day-old news is added, and as the retro-conversion of older analog clips proceeds at the rate of three hundred per day. All audio files are in MPEG format and are played on RealPlayer.
When accessing audio of Roy Rogers and Dale Evans singing "Happy Trails to You," a clip about twenty seconds long is all that is available. When searching for the speeches of Dr. Martin Luther King Jr., the entire speech cannot be found. Rather, one hears the most memorable parts of his speeches--the short sections that aired on radio news.
The Text Database contains 800,000 articles written by AP staffers since 1997. Only recently was a central repository created for the primary documents that AP journalists have generated from around the world. As with the Audio Database, text is added forty-eight hours after it hits the news wire so as to give news outlets an opportunity to break news stories first. Fulltext searching is complemented by highlighted search terms as they appear within the body of the story.
The Graphics Database consists of an impressively well-produced collection of almost 14,000 maps, graphs, charts, logos, flags, illustrations, etc., that provide excellent visuals for multimedia presentations. Graphics files (as with photos) are available through the Multimedia Archive as soon as AP posts them, without the 48-hour delay of the text and audio files. These graphics are of the same high quality found in college science textbooks. Cutaway drawings of the Space Shuttle, with internal parts labeled and described, plate tectonics, volcanoes, and stock market charts are among the kinds of useful displayable information found here.
PCs: Windows 95, 128MHz CPU, 16MB RAM,
Macs: OS 8.1, 200MHz CPU, 32MB RAM. Computers must possess a modem (28.8 or faster), soundcard and speakers, and run Netscape 4.0.4 or IE 4.0.1, RealPlayer 5.0 and Adobe Acrobat Reader. Since the archive uses Java 1.2, which IE does not fully support, Netscape is the browser of choice.
The Multimedia Archive is a resource including original images, audio, graphics, and text, which makes it an excellent educational tool. Unlike "clip art," everything in the Archive is presented within a context. There is a date, place and event associated with every item--very important for users doing research. It is tantamount to finding Indian pottery in situ, as opposed to observing it on someone's fireplace mantle. Everything the searcher finds is in its original state, so information is presented from the point of view of the person who first recorded or produced it.
The Multimedia Archive was not originally created in multimedia format. It is a collection of news in databases with disparate formats; however, it provides the raw material with which multimedia presentations may be created.
Although simple search techniques may be used for locating files in each database, advanced searching requires special modifications to the search strategy when switching from database to database.
Images may be searched by specifying any of 10 colors or hues such as white, black, blue, and/or by 71 concepts including concentration, destruction, dry, excitement, and isolation. It is also possible to narrow searches by specifying the type of photograph, such as portrait, landscape, interior, exterior, formal, action, aerial, etc. Such flexibility is of particular value to art students, historians and those interested in the social sciences.
Keyword searches for audio clips produce results based upon the abstract, header, and dateline. Fulltext keyword searching of transcripted audio is not available because the clips have only been abstracted. Clips may be searched by category (Flashback, Features, Obituary, Entertainment, etc.), or by subject (e.g., Beatles, Space Station Alpha). Audio clips may be sampled before downloading so as to speed the search process. Clip sizes range from about 250-500KB. Each second of audio is about 20KB.
The Audio Database takes longer to search than the other formats. Unlike the thumbnailed PhotoArchive, which can be visually scanned at a glance, or the listing of headlines from news text that is quickly read, audio files must be opened and heard, clip by clip, and this takes time.
User stats are available weekly or monthly upon request.
The indexers who created the fields that provide points of access into each database did so with a view to optimizing the efficient use of each individual database. No thought was originally given to search concurrently across databases and formats. How to successfully effect integrated searching is a perplexing challenge for the software engineers at AccuWeather and AP. It would be useful, no doubt, for the AP to begin indexing and archiving its varied news formats from a multimedia perspective, and adding special indexing fields to the ones already being used.
An option to search all the databases simultaneously is under development, but still a ways off. To understand the complexities of enabling global searching, it is necessary to understand that the various formats in which AP created and broadcast the news were not developed with any coordination. Journalists, photographers, radio broadcasters, and illustrators gathered news in a variety of locations under a variety of managers and deadlines. AP has long been an international organization, and much of what the thousands of affiliates produced was never archived or placed into a single repository until recently.
No one could have foreseen that all of AP's news products would be placed on the internet or that there would be a benefit to indexing sound, text, graphics and photographs according to common standards and unified procedures. AccuWeather has done a remarkable job in simplifying and facilitating the process of finding raw news for multimedia presentations by placing the gateway into all of the formats on a single Web page and in standardizing the look and operation of the search and results page for all formats.
Although it takes longer to search databases individually rather than simultaneously, there are advantages to doing so. For example, if one were looking for an aerial photo of Disneyland, the search term "aerial" would be confounding when applied to the Audio Archive. Even when simultaneous searching becomes an option, many users will prefer to search each database in turn so that queries may be tailored to have the greatest effectiveness.
Searching databases with source material has certain challenges. For example, searching for ╬Hindenberg Disaster' yields no hits. However, a search for "Hindenberg" is productive because, at the time of the event, it was not known to journalists or the public as the Hindenberg Disaster. This is a subject heading that has been given to the event by historians.
The Multimedia Archive contains undigested information--the kind sought by people who want to view history through the eyes of those who witnessed it. Each database retains the immediacy in which the news was originally generated and reported. As a result, it is necessary to determine which descriptors would have been used by people who lived while an event was unfolding.
Alternative Sources for photos, audio, text, and graphics
The photo and audio portions of the archive are its most robust features. Simply put, they blow the competition out of the water. The best alternative collection to the PhotoArchive on the Net is the Library of Congress, but it tends to be of lower quality and subject material is much more difficult to find.
Alternatives to the Audio Database include the Recorded Sound Reference Center of the Library of Congress, home to perhaps the largest audio collection available to library patrons (it includes the sound files of ABC News). But that resource is not available online. The Stanford Archive of Recorded Sound contains 200,000 recordings, but these are largely uncatalogued and are not available online either.
The oldest text files are less than five years old, and the graphics archive goes back only three years. These databases face stiff online competition from NewsBank, Electric Library and Facts.com. Nevertheless, since the cost of adding the Graphics and Text archives is relatively low, most libraries will probably wish to subscribe to the entire package.
The 48-hour delay in adding audio and text to the archive gives news outlets the first opportunity to disseminate breaking news. This does not seriously weaken the effectiveness of the Multimedia Archive since late-breaking news audio and text are available free from several online news services such as those found on the splash page of Real.com .
I would encourage everyone to take a test drive of the Multimedia Archive. Also check out the FAQ, which is well written and jam-packed with useful information in a relatively short read.
Lesson plans are provided to teachers (K-12) for a variety of subject areas to assist them with student assignments. These lesson plans are not being created by the staff, but are submissions by classroom teachers who share their best lesson plans with AccuWeather/AP.
Educators who wish to involve students in creating multimedia class presentations and high-tech portfolio projects would find the Multimedia Archive to be perhaps the best resource currently available.
Suggestions for Improvement
Certain enhancements ought to be made to the site. It is perplexing that "OR" is the default Boolean operator! This generates enormous amounts of useless hits unless queries are placed in ╬single' quotation marks. "AND" is the standard used by most search engines and by librarians. Its incorporation into this site would greatly improve user friendliness. I was told by the staff at AccuWeather that a switch in the default operator to "AND" is likely to be made.
After entering search terms, it is a bit annoying to have to move the mouse pointer down the left column to the lower part of the screen to select a database, then up the screen to click the search button. The query bar, database menu and search button should be placed next to each other at the top of the screen to minimize mouse travel and to speed searching. It is also annoying that the databases in the selection window are not all displayed at once. It is cumbersome to scroll through the tiny menu window to select between formats.
The "Image Archives" menu along the left column should be changed to "Archives," since it also contains text and sound.
The Lightbox feature, which allows files to be saved for further perusal and selection, does not discriminate between multiple users entering through a proxy server. This is most unfortunate since it allows one user to delete the files of another user! It would be a good idea to allow subscribers to deselect the Lightbox as a contract option.
Some search options available to the professional news outlets are not accessible to libraries. For example, the SoundBank, available to professional news outlets offers a Similar button, which provides an easy way to expand a search by introducing new descriptors, thereby increasing the amount of serendipity in the results. Unfortunately, this option is not available to libraries at this time. To provide a consistent interface, the Similar button should be applied across all databases.
Audio files use an extension that is not recognizable in PowerPoint. This is a major problem, but it is being seriously addressed.
In fairness to AccuWeather/AP, the Multimedia Archive is in a never-ending state of improvement. The problems mentioned here are likely to be corrected, perhaps even by the time this review goes to print. Apart from the default Boolean operator, the site works remarkably well and is very well constructed.
Libraries that provide remote online access to databases may authenticate patrons who possess valid library cards by username and password. However, the primary form of authentication is IP filtering, with remote access granted to users who enter through a proxy server with a secure, trusted URL. If access is attempted from an unrecognized IP address, an alternative homepage window loads, which prompts the user for name and password. Libraries will be contacted if an unusually high volume of hits is generated during the contract period.
Author Selected References
No author references provided.
Advisor Additional References
"Archive provides access to historical multimedia.", T.H.E. Journal, 28, no. 11 (Jun 2001): 38..
LaGuardia, Cheryl. . "The Accunet/AP Photo Archive.", Library Journal, 126, no. 7 (Apr 15, 2001): 44..
Machovec, George S. "Accunet/AP multimedia database now available.", Information Intelligence Online Libraries and Microcomputers, 19, no. 5 (May 2001): 8..
O'Leary, Mick. . "The BUDDIE goes multimedia.", Information Today, 18, no. 5 (May 2001): 13..
Tallent, Ed . "Accunet/AP Multimedia Archive.", Library Journal, 126, no. 11 (Jun 15, 2001): 112..