In her classic book “Drawing on the Right Side of the Brain,” Betty Edwards explains that all of us have the ability to draw if we learn to incorporate the creative right side of our brains with the logical left side, forming a holistic environment.

In a similar way, digital images used  by newspapers have both a visual and a technical personality.

 

The visual portion — the image — is easy to see.

The technical side, however, is equally compelling, containing a wealth of information that details a file’s basic elements — from camera settings to the conditions surrounding the image capture (see Figure 1).

Fig. 1: A display of the technical metadata as generated by Adobe Bridge, and the actual image reflected by that metadata, below.

Mining this information, known as technical metadata, can benefit newspapers planning their workflow, management and archiving operations.

But before newspapers can exploit the value these metadata can provide, they first have to understand how they can be used in a news environment.

And to do that, they need to understand the gap between the visual and technical aspects of newspaper digital images.

To that end, the Rochester Institute of Technology launched a research study to examine how five large newspapers used technical metadata in their workflows.

The five evaluated were the Chicago Tribune, the New York Times, USA Today, the Washington Post and the Rochester (N.Y.) Democrat and Chronicle.

The study looked at the digital imaging workflow of each organization and attempted to answer two main questions: First, how do these newspapers currently use technical metadata in their digital imaging workflow? Second, where could technical metadata be used in the future to improve the creative and editorial processes used within these organizations?

XML and RDF

The storage and transfer of technical metadata is underpinned by two protocols: XML and RDF.

XML, or eXtensible Markup Language, stores information by means of user-defined tags that permit the exchange of data between disparate applications and systems.

Importantly, XML enables computer systems not only to exchange data, but it also allows systems to extract meaning from data in much the same way humans extract meaning from text.

Humans rely on where each word sits in a sentence to deduce meaning. Likewise, computer applications need to have their “words” arranged in a certain way, according to a specific syntax.

That’s where RDF, or Resource Description Framework, comes in. Simply, RDF provides the syntax for creating XML statements that systems can understand (see Table 1, next page).

Recently, several major software designers have embraced XML and RDF as a means to embed metadata in image files.

Standards

Once a method exists for systems to exchange information in a meaningful way, the logical next step is to determine the data to be embedded in the image file.

This issue has been addressed by several organizations, including the National Standards Information Organization, whose technical metadata standard was approved in December 2006.

NISO’s Technical MetaData for Still Images standard includes attributes that describe both the device settings as well as the conditions surrounding image capture. The group identified 111 separate fields of technical metadata, organized into four distinct categories (see http://www.niso.org/standards/index.html).
 

They are:

•Basic image parameters - This includes technical metadata that is considered to be fundamental to the reconstruction of the image. These would cover such elements as file type, file size, image resolution, image compression and embedded ICC profiles.

•Image creation metadata - This includes data surrounding image capture and the capture system used. Image creation metadata contains fields such as a unique image identifier, the creator of the image, details of the capture device and its settings, subject distance, scene illumination, and GPS location.

•Image performance assessment - Image performance assessment consists of metrics that access the accuracy of current image reproduction as well as the preservation of images during migration to a new file format. These include high-level spatial metrics such as sampling frequency and the source X, Y dimension, as well as more familiar lower-level metrics, such as information about test targets that serve as a reference for image analysis.

• Change history - This represents the cumulative record of image processing steps that have been made since the capture of an image.

The five newspapers were asked to rate the importance of each category of technical metadata.

As expected, metadata describing basic image parameters was judged most important.

That’s not surprising, since fields such as file size, file type and level of compression are essential to digital newspaper workflows.

However, the use of image creation metadata were not as widespread, except for fields that established date and time of capture or assigned each image a unique identifier.

Additionally, image performance assessment metadata, such as test targets, were not widely used. One exception: the Democrat and Chronicle, which printed test targets on a quarterly basis and made adjustments based on target analysis.

Change history metadata fell victim to the short cycle times required for publication. Again, that isn’t surprising, especially when you take into consideration that a history of change is automatically kept as images are saved throughout the digital workflow.

The future of technical metadata

As newspapers publish more of their content on multiple media platforms, the use of technical metadata could yield newspapers great benefits.

For example, a paper could use metadata to pinpoint the time when a particular image was captured. It could also use metadata to automate cross-media workflows, leveraging the fact that the process of repurposing the same image for different media is largely repetitive.

These practices are gaining steam even as verifying the details of image capture becomes more complex.

This complexity occurs, in part, because large newspapers tend to create their own image databases for use within their own organization.

Second, the growing practice among newspapers to extract still images from video makes it more difficult to verify each image’s authenticity.

Given the fact that computer storage space is inexpensive, it would be practical for newspapers to include technical metadata currently not now considered necessary. This procedure could prove valuable in the future as the use of technical metadata increases.

Unlocking metadata

Lowering the barriers blocking the increased use of technical metadata has the potential to provide newspapers with a significant improvement in their digital workflows.

The first step is to ensure that technical metadata is embedded in image files.

Whenever possible, technical metadata that was not included at the time of image capture should be embedded automatically. The digital image cycle time in the newspaper industry is too short to trust the inclusion of technical metadata to manual entry and newspapers should avoid relying on workers to manually populate technical metadata fields.

Second, metadata must work seamlessly with metadata embedded for other purposes, such as accounting, reproduction or archiving.

Third, metadata must be preserved and retrievable regardless of how many images are placed in a single document.

Because technical metadata represents the conditions and settings of image capture, it is understandable that in order to be useful metadata must be presented to the user by means of a well-designed graphic user interface.

Therefore, it is expected that the increased use of technical metadata in the newspaper digital imaging workflow will only occur through the development of intuitive GUIs that allow users to visually adjust images using technical metadata that resides transparently in the background.

Barriers to adoption

The study found that the newspaper industry defined image quality in ways that went beyond creating pleasing images.

In fact, all of the newspapers studied said image quality had to include data that verified the authenticity of the original image capture, the consistency between visual and reported information and the difference between previous image captures.

To address these matters, the five newspapers stressed clear communication between departments engaged in a cooperative photo editing process.

In fact, this process, which can be called editorial integrity, was found to be the common denominator in the newspaper industry’s definition of image quality.

The need to examine each image to determine its editorial integrity is in some instances at cross-purposes with the ability to use technical metadata in automating newspapers’ digital workflows.

Based on the results of this study, however, there are several steps that could be taken to bridge this gap and to increase the usefulness of technical metadata.

The use of technical metadata offers newspapers the opportunity to improve their digital imaging workflow while maintaining the image integrity that has become the hallmark of the newspaper industry.

How quickly newspapers will embrace technical metadata, however, will depend on how successfully software designers respond to newspapers’ multifaceted definitions of image quality.

Like a fine work of art, the underlying structure of technical metadata must be presented in a holistic fashion.

Howard Vogl is a visiting professor in the School of Print Media at the Rochester Institute of Technology. He can be reached at hevppr@rit.edu. Franziska Frey is the McGhee Professor at the School of Print Media at the Rochester Institute of Technology and is also an extended faculty member in RIT’s Center for Imaging Science.