POMPEII

Vicolo di Narciso runs along the length of Insula VI, 1 on its east side, starting at via Consolare and extending to the city fortification wall.  It provides access to the rear entrances of Insula VI, 1 and to the properties in Insula VI, 2.  As can be seen in the photograph above, it is much narrower than via Consolare, and therefore presents a greater recording challenge.  Beginning in the summer of 2004 a companion photomosaic image of the rear street was produced for the project archives utilizing the lessons learned in creating the photomosaic along via Consolare.  The plane of the facade along the vicolo di Narciso was readily identifiable, and the surface relief minimal.  The distance between the walls of the two city blocks varied but was about 3.95 meters at the narrowest point.  All photographs were taken at 3.85 meters from the plane of the facade of Insula VI, 1.

New Nikon photographic equipment was procured in order to record the frontages of the buildings along vicolo di Narciso.  Camera settings were tested and selected to maximize the angle of view (amount of area recorded in each photograph) and to maximize photographic detail:

• Date:  Summer 2004
• Camera:  Nikon D-100 six megapixel digital
• Lens:  Nikon 18-35 mm f/3.5-4.5
• Image pixel dimensions: 3008 X 2000
• Photo file size:  5.2 MB each
• Image format:  Portrait
• Image Type:  Uncompressed NEF (Nikon RAW)
• Focal length:  Zoom set to 18 mm (35 mm film camera equivalent of 27 mm focal length)
• Number of images:  209 including retakes of images of uncertain quality

Several days were devoted to planning, equipment testing and experimentation with different photographic techniques. The resolution of the new camera equipment recorded significant detail.  It was noted that the facade was only in full sun between 11:00 a.m. and noon each day because of the shadows cast by the tall buildings on the opposite side of the narrow street.  At 114 meters in length, the facade was too long and tall to be recorded in one hour. The photography was therefore accomplished in segments, at the same time each day, on four different days.

The surface of the street/sidewalk was first marked with chalk 3.85 meters from the facade.  A measuring tape was laid along the chalk marks and the photographs were taken at one-meter intervals with the camera held directly over the tape.  The camera was hand held because of the extreme irregularity of the street and sidewalk surfaces and the close proximity of the opposing facade.  Upper level photographs, when necessary, were taken from a stepladder steadied by a helper.  A bubble level mounted on the camera hot shoe was used to verify that the lens was horizontal.  Both the photographer and a helper checked camera location, orientation and the bubble level for each image.  Each photograph was checked on the camera monitor and, if necessary, retaken. The photography required a total of about one workday for two people.

All photos were downloaded to a laptop computer at the end of each photographic session and checked for quality.  Backup copies were made of all photographs onto data CD’s.  Thumbnail contact sheets were printed for all photographs for later reference.  The backup data and contact sheets were filed in a three-ring archive manual.

An alternate approach was developed for creating the control background.  The facade was divided into four sections.  Over four afternoons, a dumpy optical level was used to mount a horizontal string temporarily along the wall at a known elevation and a 30-meter measuring tape was placed along the base.  The offset distances between the horizontal string and both the base and top of the wall were measured with a hand tape every meter down the length of the facade.  The lengths were recorded on a Mylar drawing at a scale of 1:20.  Other diagnostic features such as doors, windows and large ashlar blocks were measured and drawn and supplemental notes added. The fieldwork took two people a total of about two days to complete.

The field dimensions on the Mylar drawing were later transferred to a VectorWorks computer CAD drawing which produced a scaled orthographic reference image (see figure).  The completed VectorWorks file was converted to a bitmap (raster) image to be used as a control background in Photoshop 7.0.  This version of Photoshop had a size limit of 30,000 X 30,000 pixels, which restricted the resolution at which the photographs could be utilized.  A bitmap control background was therefore created at a scale of 1:50 and at a resolution of 240 ppi with a resulting width of about 24,000 pixels. One person produced the control background drawing in about two days.  The VectorWorks and Photoshop control background files were copied to data CD’s and placed in the archive manual.

The control background was then used as a base layer in Photoshop as a guide to size and place the individual photographs.  This work proceeded from south to north, or left to right on the drawing.  The color of the first photograph was adjusted to match test photos.  The balance of the photographs were then sequentially opened, color matched to the previous image and placed over the control background.  The narrow width of the street required that the photographs be taken so near to the facade that they were much larger than required for the 1:50 scale.  The size of each image was therefore reduced using the Edit/Transform/Scale tool to match the guidelines on the background.  Adjacent images were then merged and blended as previously described.  The following summarizes the metadata for the vicolo di Narciso photomosaic:

• Date:  Fall 2004
• Computer:  Apple Power Mac G4, System 10.4
• Image editing software:  Photoshop Version 7.0
• Number of combined photographs:  149
• Number of Photoshop layers:  151
• Background pixel dimensions:  23,686 X 2,746
• Scale: 1 to 50
• Background resolution: 240 pixels per inch (94.5 pixels per cm)
• Background dimensions at 100%:  98.7” (240.7 cm) X 11.4” (29.1 cm)
• Photomosaic image file size:  264.6 MB
• Flattened image file size:  70.5 MB

One person produced the photomosaic in about ten days.  The final flattened and unflattened Photoshop files were copied to data CD’s and placed in the archive manual.

The higher quality equipment and refined methodology resulted in an improved photomosaic image.  The additional experience gained during the process indicated the following:

• The higher image resolution (3008 X 2000 pixels) and larger scale (1:50) significantly enhanced the detail and definition of the photomosaic.
• The 240 ppi resolution of the photomosaic was satisfactory, but below the 300 ppi standard for quality printing.
• The bitmap control background worked exceptionally well as a guide to sizing and placing the photographs, but was time consuming to produce.
• The colors of the photomosaic images were satisfactorily adjusted with reference photos, but the procedure was still somewhat subjective.
• The bubble level worked well as an aid for producing photos normal to the facade.
• The upper level photographs were of reasonable quality, but taking them from the top of a ladder was unsafe.
• The one-meter photographic interval provided ample overlap in order to use only the central portion of each image.
• The Nikon NEF (RAW) uncompressed photographs, although large in file size, preserved the maximum detail and allowed adjustments to be made to the color balance and contrast without affecting the image quality.
• Ideally, photographs should be taken at a distance and focal length that would closely match the scale of the photomosaic, thereby eliminating the need to dramatically change the size of each image.

Top of Page