Format Specifications for DL Image Production Delivery

Page Images

Books

The MAX size is either a MrSID encoded file or a bitonal CCITTgroup4 TIFF file.

The SCREEN delivery is a JPEG (color or grayscale) or GIF (bitonal) with a specification of a range of 650-850 pixels wide, as appropriate to meet the goal of readability of the SCREEN size without having to go to the MAX size

The PREVIEW is a JPEG (color or grayscale) or GIF (bitonal) that fits within a 120 x 120 pixel area.

Current Etext Center practice is to reduce original scans by 30% for SCREEN images, and then create a file that is 10% the size of that file for the PREVIEW.  This create page images that are consistent for each book, and creates collections that accurately show the differences in scale in related books.  70% of the SCREEN pages are readable using that rubric.

Bitonal page images

Source:  These files are currently generated internally at DLPS.  Scanning specifications are 400dpi tiff resampled up to 600dpi and compressed.

• MAX:  These files go into the central repository as CCITT group IV fax compression TIFFs and are placed in the “max” directory.
• SCREEN:  The SCREEN GIFs are resized to a range of 650-850 pixels wide and are placed in the "screen" directory.
• PREVIEW:  The PREVIEW GIFs fit within a 120 x 120 pixel area and are placed in the "preview" directory.

Grayscale page image Illustrations

Source:  These files are currently generated internally at DLPS.  Scanning specifications are 600dpi TIFFs, grayscale or color as appropriate, descreened, oriented for normal viewing. 

• MAX:  These files go into the central repository as MrSID encoded derivatives and reside in the “max” directory.
• SCREEN:  These JPEG files should be resized The SCREEN delivery has a specification of a range of 650-850 pixels wide, as appropriate to meet the goal of readability of the SCREEN size without having to go to the MAX size, and placed in the “screen” directory.
• PREVIEW:  These JPEG files are reduced to fit within a 120 x 120 pixel area and are placed in the "preview" directory.

Color Page Images

Source:  These files are currently generated at Special Collections as well as external sources.  Special Collections items are scanned in RGB color and ranged between 400 dpi and 600dpi.   Images scanned at Special Collections use the srgb icc color profile.

• MAX: These files go into the central repository as MrSID encoded derivatives and reside in the “max” directory.
• SCREEN:  These JPEG files are treated as a group, reduced uniformly by a percentage to fit within a range of 650-850 pixels wide, as appropriate to meet the goal of readability of the SCREEN size without having to go to the MAX size, and placed in the “screen” directory.  Currently, this percentage is determined manually and not programmatically.
• PREVIEW: These JPEG files are reduced to fit within a 120 x 120 pixel area and are placed in the "preview" directory.

There is a potential need to additionally support the model developed for the Early American Fiction project, consisting of PREVIEW (120x120 JPEG), SCREEN1 (650-850 wide JPEG), SCREEN2 (850-1150 wide JPEG), and MAX (MrSid).

Odd-sized book scans

Source:  These files have various origins, internal and external, and are typified by the vast range in file size from one page image to the next one.

• MAX: These files go into the central repository as MrSID encoded derivatives and reside in the “max” directory.
• SCREEN:  These JPEG files are treated as individual scans and are constrained uniformly by a percentage to a range of 650-850 pixels wide, as appropriate to meet the goal of readability of the SCREEN size without having to go to the MAX size, and placed in the “screen” directory.  Currently, this percentage is determined manually and not programmatically.
• PREVIEW: These JPEG files are reduced to fit within a 120 x 120 pixel area and are placed in the "preview" directory.

MrSID encoded derivatives

Decision (1/29/03): At this time, we will use one set of MrSID scripts for encoding page images, images, and maps.  There is no need to create separate and unique MrSID scripts for each media or content type.  This decision will be periodically reviewed.

These numbers are:

• From 0 to 800  pixels will contain 2 MrSID levels.
• From 800 to 1600  pixels will contain 3 MrSID levels.
• From 1600 to 3200  pixels will contain 4 MrSID levels.
• From 3200 to 7000  pixels will contain 5 MrSID levels.
• From 7000 to 10,000 pixels will contain 6 MrSID levels
• From 10,000 to 15,000 pixels will contain 7 MrSID levels
• From 15,000 to 20,000 pixels will contain 8 MrSID levels
• From 20,000 to 25,000 pixels will contain 9 MrSID levels
• Above 25,000 pixels will contain 10 MrSID levels

Documentary Images (Art & Architecture, Ethnographic, etc.)

Scanning specification:  300-400-600dpi TIFF – 3000 pixels on long side, for slides or print copy work or photoprints (~20-27 Mb per file), depending upon the size of the original image or media.

• MAX: These files go into the central repository as MrSID encoded derivatives and reside in the “max” directory.
• SCREEN:  These JPEG files areconstrained uniformly to fit within a 800x600 pixel aspect ratio, and placed in the “screen” directory. 
• PREVIEW: These JPEG files are reduced to fit within a 120 x 120 pixel area and are placed in the "preview" directory.