This viewer is an XSLT solution for the display of multimedia files and text transcriptions of a digital object serialized into an xml-encoded METS document. The viewer can be adapted to present audio and text with links to a "slide show" of image files as well. In this viewer, METS files are transformed to SMIL files for display, and TEI files of transcribed text are transformed to text files with embedded timecode.
In order to control the handling of the display across different browsers and platforms, the viewer utilizes QuickTime to display SMIL files. The accompanying transcriptions are formatted as text with QuickTime descriptors. Because one cannot point into a location within the QuickTime text file, the text file is launched at the beginning of the SMIL file, along with the media file to which it is synchronized. This combination of QuickTime, SMIL, media file and synchronized text file is sufficient to meet the needs of a large number of multimedia viewer applications.
Version 4 of the multimedia viewer is a frames-based viewer that utilizes SMIL files that were transformed from METS files at the command line, along with text files that were transformed from TEI files at the command line. Because the SMIL-to-METS conversion uses a key, Saxon was used for the XSLT transformation. Saxon was also used to transform the original TEI files to intermediate TEI files that combine timecode from the METS files with the text of the TEI files. XT was then used to transform the intermediate TEI files to the non-XML QuickTime text files that are used in conjunction with the SMIL files and multimedia files.
METS to SMIL stylesheet: mets2smil.xsl
TEI to TEI stylesheet: tei2tei.xsl
TEI to QuickTime text stylesheet: tei2qtdescript.xsl
Sample METS document: track01METS.xml
Sample TEI document (go to View -> Source in IE to display): track01TEI.xml
Future versions of the METS multimedia viewer will be able to more generally match the par and seq elements of a SMIL file with those of a METS file. Next steps also include servlet-based approaches that will search a database to return an XML file with a list of SMIL files, which will then generate the display frameset. The SMIL and QuickTime text files will then be generated on-the-fly, based on the list selection that a user makes. It is also hoped that the text transcription may be loaded into its own separate frame.
These files are distributed "as is" and without promise of support. Questions/suggestions may however be directed to Eric Stedfeld, NYU Digital Library Team.