CarLibrary.org - Metadata Demonstration

July 14, 2020

This webpage uses "before" and "after" photos taken at the Frazer Nash factory in 1952 to show that metadata can be embedded in these photos to preserve date, location, car ID and other data for historical purposes and to ease creation of  traditional inventories, Collections Management Systems (CMS) and digital archives.

The open-source ExifTool software and the open-source add-on program, the ExifToolGUI (Graphical User Interface) are used to improve the identification - and eventual classification - of these digital photos.  The steps are the same for  documents scanned to image formats and PDF files.  "Embedded metadata" is the goal.

What is Embedded Metadata?

Definition: Metadata, embedded

Embedded metadata is a component of a digital file that exists alongside the content (usually binary data) within the file, making the digital file self-describing. In the realm of still images, embedded metadata may conform to specifications like the following: Exif, IPTC Metadata, TIFF Header and XMP. Other elements or chunks found in non-still-image file formats play similar roles.

Resource: Sue Chastain article on metadata in graphic images - http://graphicssoft.about.com/od/glossary/f/metadata.htm

Information from the Metadata Working Group - http://www.metadataworkinggroup.org

Software

Metadata can be added to photos in many photo editing programs, such as Photoshop.  This webpage describes using the open-source program, the ExifTool.

The ExifTool can be used:

  1. To read metadata from an entire folder of images or documents- or any set of folders and subfolders of digital files - into a CSV (Comma Separated Value) file. 

  2. This CSV file is then easily opened in Excel or other spreadsheet program.  

  3. Incorrect or missing metadata of each file/record/image can be corrected or added in the Excel (spreadsheet) file.

  4. The Excel file can be used to import this "inventory of digital assets" into a database or a Collections Management System (CMS).  

  5. The Excel file can also be used to "write" the updated metadata back into the image (or other) digital files. 

Although the ExifTool must be used from Windows (or Linux) through the "command line", which may be daunting for users accustomed to a Windows or Mac graphical interface, its power to create archive/collections inventories - and files with useful embedded metadata - should justify the effort necessary to master its use.

A separate  CarLibrary.org webpage explains the ExifTool and the ExifToolGUI in great detail, including instructions for downloading and installing the ExifTool:

CarLibrary.org/CarLibrary-Metadata.htm#ExifTool

There is an older introductory video guide on YouTube for the downloading, setup and basic operations of the ExifTool and ExifToolGUI (also: http://youtu.be/CB6F19zRqSw). 

Another older video on YouTube explains the steps for reading the metadata to an Excel file from a directory with hundreds of photos. adding metadata in the Excel file and then writing it back to the same group of 650+ photos (also: http://youtu.be/T4gTQEwUwwY).

Phil Harvey's ExifTool website also has a link to download the ExifToolGUI. Highly recommended if you wish to avoid the command prompt!

The Demo Files

This is one of 53 photos sent to the FrazerNash-USA website in 2009:

These photos, with 9 others later sent by Alec Rait, were put on a webpage of the FrazerNash-USA.com website:

http://frazernash-usa.com/AFN-Factory-orig.htm

The right panel of this screenshot of the ExifToolGUI shows the metadata categories are mostly empty:

This is the same photo after image editing and adding metadata:

The right panel of this ExifToolGUI screenshot shows the metadata categories are filled:

To see the status of metadata in the AFN factory photos, this was done:

1.  Press the "Windows" key and "R".  This returns the Windows command prompt.

2.  Change to the directory where the photos are located: "f:/afn/afnpics/AFN-Factory"

3.  Then run this at the command prompt: 

exiftool -csv -r -Contributor -Copyright -Coverage -Creator -Date -DateTimeOriginal -Description -Directory -FileName -FileSize -FileSource -FileType -Format -GPSLatitude -GPSLongitude -Identifier -Keywords -Location -ObjectName -Relation -Source -Subject -Title f:\afn\afnpics\AFN-Factory > afn-factory.csv

This is the screen after running the ExifTool command:

This is the resulting CSV file, opened in Excel:

This shows there is very little metadata any file.

5.  Because the files have the same data in common, selected metadata was added directly to multiple files with the ExifToolGUI   For example, John Fearn was the photographer.  The location of the A.F.N. factory was London Road, Isleworth, England. James Trigwell, author of "The Post-War Frazer Nash" provided the "captions" - title, description, subject - for each photo.  This screen image shows all files selected with such metadata ready to add:

6.  The ExifToolGUI was further used to fill in missing metadata.  If several files required identical data to be added, those files were highlighted and such data was added.

7.  Embedded metadata can be verbose!  In the photos sent by John Fearn, this was added to the DC:Source category:

John Fearn: "I have some 40 plus photographs of various cars taken during my time at Isleworth - (I must admit, I find it hard now to accept that I took so much time off taking so many photos!) My apprenticeship was a 'year out-of-college' - Faraday House Electrical Engineering College in Holborn - on work experience. I was so fortunate to be accepted by Frazer Nash - I have never lost my love of 'real' motor cars, and the knowledge of motor engineering acquired during those months have stood me in good stead ever since."

This embedded data is 96 words and 527 characters!

8.  To conform to "best museum practice", an accession number  was easily added as a column in the Excel table (as above) as "Identifier" (unique ID).  A basic Excel function will create a series of numbers, such as 1951.3.1, 1952.3.2, etc. thus avoiding tedious individual additions.

9.  An Excel "input" file was created to update two columns (categories) of metadata: "Identifier" and "Date", to match the "DateTimeOriginal".

10.  The Excel files was "saved as" a CSV file and then used to update all images.  

11.  The ExifTool command to write this updated metadata to the files in the folder was:
exiftool -csv=afn-factory-input.csv-ext jpg f:/afn/afnpics/afn-factory/

12.This is the spreadsheet that shows the full extent of the metadata:

A.F.N. Ltd Factory Photos, 1952

Original webpage:  http://frazernash-usa.com/AFN-Factory-orig.htm

Current webpage:  http://frazernash-usa.com/AFN-Factory.htm

ExifTool and ExifToolGUI: Summary and Recommendations

1.  The ExifTool can be VERY useful because it can add and/or replace specific metadata in images (and other file types) with new data from a CSV file (a standard file type exported from Excel) .  

2.  The ExifTool and ExifToolGUI are a powerful combination to improve embedded metadata of digital images. I use these tools for my collections, to make archives with the Greenstone digital library software and the PastPerfect Collections Management System.

3.  Although there are standards for metadata categories (EXIF, XMP, IPTC) and "authorities" for certain metadata (person names, places, cultural objects), no standards have been found specific to vehicles - manufacturers, divisions, models, etc.

4.  The factory images were improved using Adobe Photoshop Elements.  James Trigwell also enhanced the original photos more skillfully using these steps.

5.  Although the metadata can be usefully collected in a Excel/CSV file and an Excel file is an inventory of these digital assets, importing the Excel data to a database provides a more secure home for the data.  The metadata of these factory photos was further imported from Excel to the open-source LibreOffice Base, as a "Table".  Forms for searching and reporting from Base can also be made.   However, if the archive/collection assets include physical media - prints, slides, negatives - a dedicated database for photos should be considered,  Suggestions welcome!

Therefore:

1.  Use the ExifTool from the command line to read entire folders/subfolders of photos.

2.  Review the resulting Excel file ("saved as" from the CSV output file) to determine which metadata categories will help organize these collection assets. Using Dublin Core categories should be a high priority.

3.  Use the Excel copy/paste functions to fill in missing metadata.  Use the Excel "data fill" command to create accession numbers in the DC:Identifier category.

4.  Use the ExifTool from the command line to write the Excel file ("save as") to the CSV format and then back to the entire set of photo folders/subfolders.

5.  The ExifTool may be used again at any time from the command line to read the metadata-updated photo folders/subfolders, creating "data-rich" Excel files for import into a collections management, content management or digital library (e.g. Greenstone) software program.

Email me with any comments, suggestions or questions: Bob Schmitt, rgschmitt@gmail.com