MULTISPECTRAL FILM IMAGING

Current Imaging Methods

Until recently, technology has limited the amount of data we can capture and store digitally from an analog format (film). Our focus as archivists has been to make the moving image usable and easily viewed using current technology. However, with imaging technology advancing at an exponential rate, we find ourselves constantly re-scanning for new resolutions and technical specifications.

Film is both reflective and transmissive, which means its native colors are detected by capturing the light reflected from it, and different colors are detected by shining light through it. Current archival and preservation technologies use RGB scanning to assess changes in color depth and resolution. However, traditional RGB scanning, the result of shining a light through the film, only captures three colors that are widely spaced in a two-dimensional raster. Additionally, the current technologies rely on access to physical copies to allow for changes in resolution and color depth. In situations where original film is lost or deteriorated beyond repair, we rely solely on any digital copies we have preserved. As time passes, this reliance becomes riskier, and the artifacts introduced by reprinting film lead to generational fading.

Film, as we know, has an "expiration date". Continuous reprinting is no longer the most technologically sound or economical way to preserve these works. A digital solution is not only necessary, but with new technologies, it becomes capable of eclipsing current preservation methods.

What is MULTIspectral Imaging?

  Courtesy Wikimedia Commons User:Arbeck

Courtesy Wikimedia Commons User:Arbeck

Multispectral scan data is called a datacube because of its many layers of scans at a variety of narrow spectral light frequencies - ranging from infrared to ultraviolet. These dozens of narrow layers provide a far more accurate representation of the color information in film. Instead of an averaged RGB pixel value, multispectral scanning allows for each pixel to have a specific spectral response curve. By using narrow-band imaging technologies, obtaining the spectrum for each pixel in an image will enable a much richer, informative, and accurate data representation than ever before.

 

Applications for Multispectral data

FilmicPresentation_AMIA16.pptx.jpg

Capture of image data carried on film at the grain level, with multispectral color depth, allows for not just digitization to the point of what the human eye perceives as acceptable – but faithfully encodes the richness of data that film is capable of capturing. Coupled with other technologies, multispectral imaging provides a solid foundation for the extensible, data-centric model proposed by FILMIC for rich preservation of film.