Report of the Library Technology Officer

January 2003

Vanderbilt Television News Archive

Following two very slow months, Website activity and videotape requests increased significantly in January. In addition to the increased activity, staff made headway on a variety of projects, and significant progress has been made on the NSF digitizing project.

Off-air recording

Web server access

1,084 new customers registered on the website
69,621 total entries in activity log
13,271 views of the home page
5,927 views of the search page
7,607 searches executed
3,332 Calenadars viewed
19,381individual records viewed
4,974 program listings viewed

Abstracting and Database Maintenance

Visitors

Loan Requests filled

View the Cumulative Table of Statistics for the Archive's activities.

Publicity

The Television News Archive was mentioned in the Vanderbilt Hustler, suggesting some of the free resources available to Vanderbilt students. (Clancy, Megan. "Many opportunities for fun and freebies on campus" January 28, 2003)

Other Activities and Projects

Special projects underway include an effort to catalog some of the backlog of Nightline broadcasts and continued clean-up of entries on the calendar system. In addition to regular staff, students, and Cheryl Carpenter are assisting with these projects. With their help, the backlog of Nightline broadcasts is being reduced at a rate of about 10 programs per week.

In recent months, Archive staff have begun making more use of the calendars to provide to users as much information as possible regarding shows that are still in the process of being abstracted or those that are missing from the collection. This enhanced use of the calendar system now replaces the daily logs that were previously kept as a separate up-to-date record of programs recorded for the collection.

Cheryl has also been working on helping us develop a database record structure for the presidential state of the union addresses. We need to develop a way of creating records for these events both to accurately describe the material that we have on tape that includes both the speech and network commentary, as well as the sub-set of digitized files which include only the speech itself.

Website Development

A minor enhancement that Marshall developed this month involved creating a way for users to easily navigate from a monthly calendar page to a program listing and then return back to the same calendar page.

NSF Grant Report

Segmentation of Video. A key issue for the digitization of the videotape collection of the Television News Archive involves dividing the news programs into short segments. In a video streaming environment, it is much more desirable to offer short clips of 1-3 minutes rather than complete programs of a half hour or more. Further, we envision a system where each of the records in the online database of abstracts corresponds to a single video clip. The TV-NewsSearch database follows a structure that divides each broadcast in 25-30 segments according to topic covered. Each block of commercials also forms a segment.

In our earlier planning, we had planned on physically dividing the files, a labor-intensive process, but one we thought necessary. To physically divide the files, the whole program would be digitized and the individual segments would be divided using a video editor.

More recently, we have developed ways to avoid physically dividing the video files, relying on a software control mechanism to present short video clips out of complete program files, relying on metadata in the TV-NewsSearch database. This virtual segmentation will drastically reduce the amount of labor involved in the digitizing process.

The procedure that we have developed involves digitizing each half-hour program on ¾-inch videotape into a single MPEG-2 file, which becomes our digital preservation copy. Using the Helix Producer Plus program, we transcode the MPEG-2 file into RealMedia format, which is suitable for streaming on a video server. When transcoding the MPEG-2 files, we also reduce the screen size by half to reduce the size of the file and the bandwidth necessary to view it.

The presentation of short video clips is accomplished through the use of programming within the TV-NewsSearch interface. When a program has been digitized, the user will see a button for viewing the video. The button launches an intermediate Web window, which includes scripts to calculate the beginning and ending offset of the segment requested relative to the complete program. It constructs a URL that links to the video server, selects the appropriate file, and specifies the start and stop times for the video clip. When the user views the video, only the material corresponding to the segment requested plays, just as if the file had been physically divided. The metadata in the database record is accurate only within 10 seconds. We compensate for this level of accuracy by adjusting the start and stop times accordingly.

A prototype of this streaming video delivery system has been constructed. At this time we restrict access to the streaming video is available to authorized staff, while we continue to work with copyright and other legal issue.

Steve Davis continues various activities related to digitization. Some of his current projects involve digitizing presidential speeches into MPEG-2 and transcoding them into RealMedia format. We have been experiencing some technical problems with conflicts between the digitizing and transcoding software. He was eventually able to make the two applications co-exist.

Steve has also been working on a series of benchmark tests for each of the MPEG-2 encoding cards we are testing for the project. For each card, we encode the same video clip at a range of different bitrates, resolutions, and audio sampling settings. So far, we have obtained and tested three different cards in this way, the Expert Archiver Plus, the Canopus Amber Archiver, and the Canopus DVR-1000.

This month we also obtained a card that receives and encodes broadcast television signals. A drop from the Campus TeleVU system was installed in our workarea in order to be able to begin developing a process for digitally recording the news. A key part of the NSF grant involves developing an approach for migrating from the current videotaping operation to one based on digital recording. The initial card that we are testing for receiving and encoding news broadcasts was the Hauppauge WinTV-PVR-350. After some successful initial testing, the card eventually stopped working and had to be sent back for replacement. We expect to have the replacement soon so that we can continue this part of the project.

DVD Storage. As we create an increasing amount of digitized video, the storage of these files is a significant problem. In order to avoid consuming all available disk storage, we have been making copies of the files on DVD-R. We obtained a Pioneer A-105 for this purpose, and have learned how to make both a data DVD that stores the MPEG-2 files in their native MPEG-2 program steam format, and a DVD that will play with standard consumer equipment in the same way as DVD movies.

Long-term storage options.Our plans involve relying on the Library of Congress to provide long-term archival storage of the MPEG-2 files. We do need to have the files available locally also, but not necessarily on-line. We have been making copies of the digital files on DVD.

Automated MPEG-2 to RealMedia Transcoding. Marshall worked on establishing an automatic procedure for the MPEG-2 to RealMedia transcoding. He wrote a series of Perl scripts that can be run in conjunction with the Windows 2000 scheduler to allow this process to be offloaded onto a remote system, with the resulting file created directly on the media server.

Digitized material. While working on developing methodologies and procedures, we have also been accumulating an increasing set of digitized content. To date we have about 25 hours of evening news digitized as well as about 31 presidential speeches.

Library Technology Officer Activities

The majority of Marshall's time this month was spent involved in the TV News digitizing activities described above.

Marshall attended the American Library Association Midwinter Meeting in Philadelphia, PA. Some of his activities included the Editorial Board meeting of the LITA publication Technology Electronic Reviews; participated on the LITA Top Technology Trends panel, and convened the SIRSI Large Sites Interest Group. Marshall also attended other LITA meetings and met with representatives of all the major library automation companies.

This month Marshall participated in meetings of the Heard Library Web Task Force, its Search Engine sub-group, and the new Digital Collections Committee.

Extra-curricular Activities

Marshall's regular Systems Librarian column was published in the December 2002 issue of Computers In Libraries.