One consequence of the move to cross-media publishing is that still photographers are now also shooting video. In the past the International Press Telecommunications Council (IPTC) has developed metadata standards for tagging still images with essential data to ease the exchange of images between news organizations. IPTC photo metadata standards include essential information as to the source and copyright as well as information about the subject of the image.
As news photographers start to shoot video, they need the same facility to add similar data about the clip as are used to adding to still images.
At the recent Photo Metadata Conference organized by IPTC and held in conjunction with the CEPIC European image library Congress in Istanbul, Turkey, on May 20, 2011, the IPTC challenged vendors across the media industry to create the conditions for interoperability for metadata embedded in media files.
Following a 2006 metadata manifesto issued by the Stock Artists Alliance, the IPTC has created a new updated document the Embedded Metadata Manifesto to cover all media types, including stills and video.
The manifesto outlines a set of five principles:
- Metadata is essential to identify and track digital media.
- Media file formats should provide the means to embed metadata in ways that can be read by all.
- Metadata values and semantics should remain the same across media formats.
- Ownership metadata should never be removed.
- Other metadata should only be removed with permission of the copyright holder.
"Metadata is a key driver for business" says Michael Steidl, managing director of the IPTC. "With photographers now shooting video and agencies looking to market cross-media and multimedia products, there is a compelling need for metadata which can be used across stills, video, audio and even text. There are challenges in the video sector, with multiple formats in use, and no widely shared way of embedding data in the media file itself.
Nevertheless, we believe the challenge must be taken up by users and vendors in the media business, so that solutions can be found. We will work together with the media industries to build on the work we have done with still images, so all businesses can benefit from clear, unambiguous data."
Still image metadata is carried in the extensible metadata platform (XMP) and can be embedded in JPEG and TIFF image files. In a JPEG file the XMP data is stored in a marker segment in the file header. XMP can be embedded is several video clip containers, including MPEG-4, QuickTime MOV, as well as MXF-based media formats like P2 and XDCAM-EX.
The issue here is not whether the IPTC data in a XMP format can be embedded in a video clip, but whether manufacturers support the format. Most video equipment vendors have gone down the SMPTE MXF route.
What we are seeing here is one set of users (still photographers) on the same learning curve that the broadcast community went through when the SMPTE drew up standards for metadata dictionary structures and MXF data structure. As users found out, it was a long time before they saw vendor support and support that allowed equipment from different vendors to plug and play.
I think the broadcasters could say to the photographers “we’ve been there, and done that.”
The photographic community already has tools for tagging in the field; Photo Mechanic from Camera Bits is one widely used logging tool, and Adobe Bridge is another. The destination of photo files is often Adobe applications, so there are no problems with XMP support. However, when you move to audiovisual news content, there are far more vendors to deal with: NRCS’, nonlinear editors, WAN acceleration distribution networks, and that is just television news. There is video news carried on websites, and another set of platforms and products.
Wrapping any future IPTC Audiovisual metadata, perhaps in XMP format, with audiovisual clips has two possible routes forward. Use the MXF platform, and take advantage of all the work that has been done so far, or start persuading vendors to develop IPTC and XMP metadata support.
We all know that vendors will only develop products if there is a business case. Product release cycles can be measured in years, so support for new specifications will not be immediate. Standards are acknowledged to be a good thing, but the vagaries of implementing support for standards are well known.
The usual workaround is for third-party vendors to make transwrapping utilities that sit between the image source and the processing applications. These work, but it is an added layer of complexity and cost. There is nothing like primary vendor support for the native standard.
Interoperability continues to dog the creative industries, especially since the move to file-based workflows. Everyone involved knows these issues cannot be fixed overnight; it takes much work behind the scenes to cajole vendors to meet the needs of their customers, without impacting on their commercial goals.
I have no doubt that some standard will eventually be reached, but support and acceptance across the board may take a while.
