Is a Picture Worth a Thousand Searches?
Getting the Most out of Royalty-Free Image Libraries
By David Patschke
For student designers, royalty-free image libraries are an invaluable resource. Learning how to navigate these digital archives can radically reduce search time; students and parents on a tight budget no longer have to pay for stock photography. And perhaps most important, students can steer clear of copyright violations.
In general, these royalty-free image libraries are a huge improvement over Web sites offering free fair-use imagery, which tends to be spotty and non-specific, and of variable quality. For diverse, high-quality imagery, you have to pay for it through sources such as iStock.com, probably the most famous and largest library for this purpose. Students can buy credits towards individual images (they vary by size, type, and resolution), or they can even earn credits by selling their own images on the site.
Trolling the web for free pictures typically leads to long searches yielding uneven results, and (needless to say) most likely copyright violations. In contrast, when schools or school systems purchase an image library, students immediately have free access to DVDs with hundreds, even thousands of images. Each library commonly has a theme, such as “business” or “fast food” or “active lifestyles” with relevant items in these categories.
Multiple libraries can be purchased over time, building a large collection of these works. And, like any creative element, styles, fads and fashion all change –leading to graphics that are stuck in time while the students are fresh every year. Libraries need to be constantly added to and cultivated to sustain their usability.
A major problem emerges, however, soon after a second library is purchased. Each disc usually comes with some form of index that allows users to see what’s on each volume - a quick and easy way to find what you’re looking for. As soon as you add another disc, however, now users have to look in two places to find images. Then, as more discs are added, three places, then six, and so on. Now your student is spending more time on image-searching than on the creative process. So how does a school build a single index for multiple libraries?
One solution to this problem is found at the University of Baltimore’s Digital Design Studio, where over 500 volumes of royalty-free images are on hand for users to search through – well over 100,000 images. But searching through so many libraries is not an option for a student needing to get their project finished sometime this semester. Commercial products are available that can help users to quickly and easily search through all possible images, with Canto Cumulus emerging as the most widely used. In fact, many libraries ship with a Cumulus index and free reader for that disc for electronic searching. It is possible to purchase a Cumulus server application to store the main combined database of thumbnail images and keywords that all local users have access to, however each client station also needs a licensed copy of Cumulus running to do the searches from the server. Plus, the discs that don’t include Cumulus indexes would need to be manually entered into the system.
Given the current economic climate, though, many school systems can’t pay for multiple copies of an indexing program. Are more economical options available? At UB, rather than invest in yet another software platform, we chose to build our database on top of an existing program that was already in use: FileMaker. Designing and building our database from scratch using FileMaker allowed us to custom tailor it to our needs in our Apple Macintosh environment without additional investment. The tricks, though, are what make adapting this use to FileMaker more interesting.
While some libraries come with an aforementioned Cumulus index (which includes a thumbnail image, image info, and relevant keywords for each image), most do not. Some come with a PDF file, some with a local html page, and some with nothing at all. Various trickery is involved in getting this data into a single new platform. Knowledge of AppleScript (or a strong desire to learn while doing) as well as crafty use of text programs is a must for doing this type of work.
Canto Cumulus is Apple scriptable –meaning it can be programmed to do repeated tasks (or macros) over and over again automatically using Apple’s built-in scripting language. In this way, it is possible to have a Cumulus index open on the host system, as well as a FileMaker database, and run a script that copies each item from the desired Cumulus fields into the new FileMaker fields one at a time, switching back and forth between each application as necessary. It can take some time, but can run unattended. Libraries that do not have any index can be indexed by a single copy of Cumulus, and then similarly transferred to FileMaker (since FileMaker has no way of creating thumbnails of images).
If a PDF file or html file already exists, then similar scripting can be employed for moving this data to FileMaker– although it is usually easier to create new thumbnails using Cumulus or Adobe Bridge. The resultant records in the database end up with only a thumbnail image but no other relevant data. Other methods need to be employed to get the text (image name, location, keywords, etc.) into the appropriate FileMaker fields.
Many times the producers of image library reference files resort to giving more information than you’d ever want to use –or should use. The result is a text file full of extraneous values that do more to confuse users than draw them nearer to their image choices. A tab delimited text file can easily be converted into a table in most word processing programs. The advantage of working with text in tables is that whole columns can be selected (and subsequently deleted), allow you to easily and selectively remove whatever fields you don’t need or want. Converting back to a tab delimited field is easily done. Sometimes you may find individual fields that need combining, such as a “keyword” and a “description” field, which you may want in a single keyword field. An initial response to this challenge might be “no problem, merge the fields”. However, that action merges all fields in all records, combing all of your words into a single massive field. Instead, rearrange the columns to be next to each other, then add a new blank column in between them. Insert a variable string (such as “xxxx”) into each row of the column –very simple and quick to do. Then, after you covert back to a tab delimited file, simply do a Find and Replace where you search for a tab character, then xxxx, then another tab character, and replace it with a single space. That combines the fields on each row separately.
Once you have extrapolated the various data from the libraries and have created a single database to host the reference, an easy-to-use interface should be created that encourages simple interaction for the users. Searching is a basic function for Filemaker, so that requires minimal effort. However, formatting the search results in an effective manner should be a priority. You might consider two types of layouts: one that maximizes the number of thumbnail results the users can see in each search, and a second that includes keywords for when images may include subjects too small to comprehend in a small image.
The image sources that come distributed via disc needn’t stay on that format (check on each disc for the copyright limitations), in fact keeping the images on discs is logistically difficult and very risky. Lost, stolen, or damaged discs rob you of the resources, and handing out and collecting discs is not very manageable beyond a few discs. Using a network file server is the best choice here, one that allows for assigned privileges to prevent over-writing or deleting the images. Even still, once a user has located an image she wants, she still needs to locate and mount the source. Luckily, FileMaker is also scriptable, and scripts can be written that take the information from the record and use that data to automatically mount the file server and the exact directory that hosts the particular image, which can be revealed to the user so that she can copy it to her local drive for editing.
Extrapolating all of the data from each new resource you obtain can be a painstaking process, but once it is finished, you’re left with a result that far exceeds the input effort. And with each new volume acquired, techniques can be scripted and automated to streamline the process of adding new images. Further, tech-savvy students can assist IT with the endeavor, making the indexing process itself educational. Ultimately you’re left with a single supply of imagery that is far more useful to students than a collection of individual pieces.
David Patschke is the Director of Technologies at the University of Baltimore in Maryland.