Penn Libraries Digitization Preferences
The goal of most digitization projects at the Penn Libraries is to represent the original physical item accurately. In order to achieve accurate representations with consistency, SCETI operates according to the preferences listed below. These preferences are defaults. Any particular project may vary.
Resolution (all 24-bit color):
Color images are preferred. There may be limited cases where the Libraries decide that grayscale images will suffice, sacrificing color for smaller images. The resolutions chosen are in line with peer institutions, but on the higher end of that scale. Penn does not currently own equipment that can image all items at 600 ppi. In these cases, items will be imaged at the device maximum.
- Manuscripts – device maximum up to 600 ppi
- Rare Books – 400 ppi
- Modern Books – 400 ppi
- Newspapers – 400 ppi
- Oversize items – 400 ppi whenever possible
- Transparent materials, negatives, microfilm, and microfiche – PPI determined on a case-by-case basis
Every imaging session should include a reference target for color correction and size reference. This is usually made by laying the target over the first image, then removing the target and imaging again. The target can alternatively be placed within the image area if necessary. Multiple reference shots may be necessary if an item is imaged in multiple sessions.
Scholars commonly need to see the structure of manuscripts and non-modern printed books. It is generally easier to crop within the page. Because modern printed books, serials and newspapers are mass produced and that content is more commonly studied than the structure, it is generally acceptable to crop within the page.
- In general, all similar shaped parts of an object will have the same crop dimensions. Different sized portions, like foldout or spines, may be different (see relevant sections below).
- Manuscripts and non-modern printed books will be cropped outside of the page such that the book edges are visible.
- Modern printed books, serials and newspapers will be cropped inside of the page if it saves processing time. In these cases, the edges will not be visible.
Image Page Display:
In limited cases, some items will initially be shot two-up if the physical item is robust to withstand it. For most materials, these images will be split during cropping so that users can view, manipulate or reorder individual pages. Indic materials do not have this same requirement and can be displayed two pages at a time.
- Indic items will be displayed two-up (two pages in one image).
- All other items will be displayed one-up (single page per image).
It can be difficult for a camera operator to know what shots are important and what shots can be skipped. It is more timely to them to shoot all iterations than to accidently miss one and have to reshoot it later. The flexibility to stitch foldouts will achieve consistent resolution for all images in an object.
- All iterations of foldouts will be imaged regardless of whether there is content on them or not. It is not required that they are all displayed. These shots may include:
- The foldout while folded (recto and verso)
- The foldout unfolded (recto and verso)
- The crop size may be different from the rest of the item.
- Large foldouts may be imaged in sections that are digitally stitched together. The original images will be retained, but may not be displayed by default to the user.
- The pages facing a foldout will have duplicate images made so each image has the correct page opposing it. This practice may change in the future to utilize technology that will allow the user to see the same page facing different images.
Inserted material are generally foreign to the host object. It is not necessary for a user to see it on the host item, but rather where it was found.
- Inserted material will be imaged by themselves (not on top of the object).
- The crop will be different than the rest of the item, larger or smaller depending on the size of the inserted material.
Spines and Edges:
While is less important that entire structure of modern books are captured, but the spine frequently has important information.
- All spines will be imaged.
- The head, tail and foredge of manuscripts and non-modern printed books will be imaged.
- Any painted or otherwise interesting edges of modern printed books will be imaged.
Clasps, Chains and Other Hardware:
Hardware is an essential part of the object, but negating focus problems and decreasing file size outweigh the advantages of having the hardware visible in every image. Including them in a limited amount of shots should suffice for most items.
- Clasps, chains and other hardware should be visible fully visible in the front and back images. These shots will be larger than the rest of the codex.
- Clasps should also be fully visible in the inside cover shots. These shots will be larger than the rest of the codex.
- Every attempt will be made to keep the hardware in focus in the relevant images.
- They may be cropped out all other shots.
- Attached bookmarks, generally made out of cloth, will be removed from the original location and imaged in the front and back shots close to the spine of the book.
- Tissue plate protectors will not be imaged.
- All blank pages, even large sections of them, will be shot even though they may not be presented to the user by default.
- Master images will not be touched up. This includes removing dust and scratches from outside of the content area in an image.
- Framed items will all come with instructions of whether to shoot in the frame or to have them removed from the frame.