Greetings, I would also like to add to Guenter's excellent summary a few points.
First, I can't agree more with the notion that absolute (captured) resolution does not necessarily parallel with the relative quality of various capture specs, for a number of reasons, but primarily for the dueling issues of capture-path capability (the limits of your optics, your sensors, lighting, and other environmental and process factors) and the usable detail present in the actual source (whether the physical object itself, or an interim medium such as a 4x5). Not all 2000dpi large format capable scanners are created equally - in fact, the range of results can be shocking and this must be the most rigorously tested aspect of your standards - a good clean drum scan at 1000dpi can likely provide much better output than far cheaper flatbeds at 2400dpi (scanning the same material) I would, however, like to add to this discussion that another significant point to consider is the intended function of the final digital media. Is there an output component to the digitization, or are these images destined for online study purposes? Do you have anticipated output media in mind - B&W 8x10s, 35mm slides, banner prints - or are you attempting to capture all usable data in the original media in order to provide a true digital surrogate? We can assume most are working at some level for the purpose of digital preservation. However, whether you are digitizing from existing media or doing direct digital capture, there are many variables to work out. Direct digital capture of works can, in some cases, exceed the abilities of 4x5's in capturing fine detail with minimal noise/grain. In this case, the decision is one of economics: is there a perpetual value - in terms of study and output uses - in capturing a 600MB file (for example) that can be weighed against the increased cost of storing and manipulating such an image? Conversely, are the standards setup in the past for film simply to be translated into digital standards - basing your capture on the perceived need to capture a virtual 4x5, or can you assess your needs in a more direct fashion, for instance in respect to anticipated final output formats such as fine art books, posters, study prints, etc? Some may find comfort with attempting to emulate film standards, insisting on capturing no less (or more) data than can be extracted from their previous media. This approach is simple in that it provides a digital equivalency standard, but it can't always be assumed that resolution is the only factor in equivalency. Direct capture in some/many cases has the added benefit of reduced noise in the imaging, which translates into greater usable resolution (versus actual resolution), while in others cases may result in lesser dynamic range. Digitizing from existing media presents it's own questions. While it is easy to think of things in terms of maximum theoretical resolution of your film, for instance, there is an enormous range of debate on the subject. So many factors affect the usable resolution of a scan that in many cases your scanner may be capable of capturing far more resolution than there is actual detail in the film - in the end, with large files you may be wasting bits imaging grain. This is another area where I can't stress testing on your equipment, and for your workflows, enough. Beautiful slides may be created with clean files substantially smaller than you might imagine, while much larger files scanned on lesser equipment may pale next to them. In the end, you must measure your objectives, and test your assumptions - equipment and process variables in many cases are far more signficant in achieving pleasing and relevant results than theoretical standards. Certainly, I'm not able to cover the entire process here, but just raise the questions. I would be happy to discuss more specifics, including some of our own digital standards here at The Getty, but what I truly feel is most important is the process used to assess your own needs, which will hopefully result in better answers than what size files I happen to think are nice for us. Best regards, Roger Howard Digital Media Specialist The J. Paul Getty Trust 310/440-6908 --- You are currently subscribed to mcn_mcn-l as: [email protected] To unsubscribe send a blank email to [email protected]
