Some people eat, sleep and chew gum, I do genealogy and write...

Thursday, May 9, 2013

Towards Genealogical Standards for Published Photographs

To adequately preserve our genealogical heritage, we need firm standards for the digitization of analog media. The proliferation of online media in less that archival quality, threatens the integrity of the historical photographic record. Allowing less than archival quality digital images to be incorporated into  a permanent record will ultimately degrade the entire value of the historical record. To the extent possible, efforts should be made to educate those submitting images to conform to archival standards as much as possible. Should lower quality images be accepted by a repository, it should be clearly explained that the collection is not intended as a substitute for properly archiving the analog media according to the best possible methods then currently accepted by the archivist community.

Here are some general guidelines which I propose should be implemented for both the analog and digital images:

1. As much as possible, properly preserve and do not modify the original analog media; printed photographs, transparencies, acetate negatives, glass negatives or in whatever other original format.
2. Do not do anything that will preclude future standards of preservation. Do not mount or modify any original photographs.
3. Use the best currently acceptable standards of both analog and digital preservation and file formats.
4. For digital files, maintain adequate backup procedures with multiple backups on different devices.
5. Document any special circumstances concerning the images produced.

From the historical standpoint, just as sources are preeminently important for facts asserted in genealogy, it is equally important to establish the provenance of any image. For an example of a detailed analysis of a daguerreotype asserted to be of a prominent ancestor and concluded to be otherwise, see TheAncestorFiles: The Tanner Family Daguerreotype and Conclusion. Also see the links contained in that article to the preceding series of articles discussing the provenance of the daguerreotype.

Most of the photos put online are not only published without provenance, they are copies of copies and asserted as images of specific individuals and families without tying them to specific times or places. This inevitably leads to a degradation of the quality of the images, as well as to misidentifications and inaccurate attribution.

The information about a specific photograph, as much as possible should accompany any posting of the article in a venue where the photograph is connected to a particular person or family. It is preferable to attach the information directly to the image by means of metadata. However, there are not any specific methods for attaching metadata and particularly, such additional attachments, if done by a separate file, may become separated from the original. At the very least, the title to the photo should be descriptive enough to identify the subject matter.

Technical standards for the Library of Congress are found at Standards at the Library of Congress. The links from this page go to many of the exiting industry standards. However, there is a need in the genealogical community for a practical and more easily understood set of standards to be applied by researchers in general.

The basis for the following suggested standards came from Melanie I Sturgeon, Director, History and Archives Division, Arizona State Library, Archives and Public Records in a document entitled Minimum Standards for Born Digital Photography. I have modified Dr. Sturgeon's proposal by updating the technical aspects, adding standards for scanning and modifying other details.  I am proposing the following minimum standards for digital acquisition and storing:

Genealogical Image Standards

Minimum Digital Camera Image Acquisition Standards:

Minimum megapixels
Digital Single-Lens Reflex (DLSR) minimum 12 megapixels
Digital “point and shoot” minimum 10 megapixels

Note: The smaller sensors on “point and shoot” cameras can affect
image quality Low resolution images and those taken with mobile phones or
disposable digital cameras do not meet minimum standards and
are not acceptable.

Settings
Always set camera at its highest possible resolution setting.
Color images should be in RGB format (not CMYK)


File Formats:
When possible, images should be made and downloaded to a computer in Camera Raw format.


Minimum Scanner Image Acquisition Standards:

Minimun optical resolution
300 dpi [Note: Most flatbed scanner claim very high resolution. However, an optical resolution standard of, at least, 300 dpi is commonly accepted as sufficient for archive purposes. See the Library of Congress:Preservation.

File Formats for both digital cameras and scanners:

When possible, images should be made and downloaded to a computer in raw image format and converted to an acceptable file format. Where capturing the images in raw format is not available, the images should be downloaded to the acceptable file format directly. The acceptable file formats include:
  • Tagged Information File Format (TIFF) 
  • Joint Photographic Experts Group 2000 (JPEG2000) 
  • Digital Negative (DNG) If the images are downloaded in raw format, they should be immediately converted to Digital Negatives (DNG). 
  • Joint Photographic Experts Group (JPEG) in some cases on a 
  • case by case basis. 
[Note: many scanners and most less expensive consumer cameras default to JPEG image format. Care should be taken when this is the case to preserve the original image downloaded without modification.]

Unacceptable File Formats:

Bitmap (BMP)
Graphics Interchange Format (GIFF)
Portable Network Graphic (PNG)
RAW - Most camera manufacturers have created their own “flavor”
of the RAW format. These files are encrypted and
proprietary. This may affect their longevity and readability.

Metadata
Metadata is descriptive information about an object. To comply with standards images must, at a minimum, have:

  • People identified
  • Place identified
  • Date image was taken
  • Event
  • Photographer
  • Copyright information

Filename:

Metadata should be in a separate tab delimited file or in DNG format
To comply with standards, images must have appropriate filenames.

[Note: Each camera brand has its own method for naming files. Rename files with meaningful filenames. Filenames should be unique to each image

Image Editing:
Images should be unedited. Do not

  • Crop
  • Color correct
  • Adjust white balance

I most certainly accept corrections, suggestions and ideas about how these proposed standards may be generally disseminated and accepted.

3 comments:

  1. While all of this is a great idea for preserving archival photos, can you tell me how cropping or color correcting or using jpg will lead to flawed genealogy? I doubt if I use my 6MP camera to copy a photograph that it will threaten my proof arguments. What am i missing?

    ReplyDelete
  2. I would add to the metadata:

    Location of the original e.g. Mabel Smith's personal collection, or an archive.

    Details of the collection it belongs in e.g. John Smith's photo album.

    The provenance or archival history e.g. inherited by a grandchild of the original owner, donated to an archive by Fred Jones.

    For official archives/repositories, the archival reference used to retrieve the item.

    ReplyDelete
  3. I would want metadata to be embedded in the image file as separate files can easily become dis-associated. The metadata should include all the information you need to properly cite the image and analyse its reliability. Missing from the list are:

    1. The location of the original e.g. personal collecttion of Mabel Smith, or an archive.
    2. Details of the collection to which the image belongs e.g. John Smith's photo album
    3. Provenance and archival history of the image e.g. inherited by Mabel Smith, granddaughter of original owner John Smith,or an extract of the archival accession data. Part of the provenance of a digital image are any manipulations done on it.
    4. Technical data from the camera/scanner used e.g. camera model, resolution.
    5. For archive images, the archival reference used to access the original.

    As far as I know, these data are not accomodated in the exif fields of Jpeg files. Do exif fields map directly to archive catalogue/accession data fields?

    Sadly, photo processing software often strips metadata from image files that are reduced in size for publication online.

    I don't agree that image files should never be manipulated, as some manipulations are desirable e.g. for web, or to enhance readability. It is important that changes are acknowledged and made honestly e.g. you don't photoshop your grandad in place he wasn't. Cropping a head of someone to incorporate in a chart is acceptable, but the source information for the cropped image should refer to the original.

    ReplyDelete