I’m periodically asked the best way to archive old photos. I’ve seen the Canon CanoScan 8800F recommended by a few different public librarians, so we just picked one up for a community center. I started experimenting with it yesterday on some of our old family photos. Based on some research online, scanning film (negatives and slides) will provide superior results to scanning the photos because you are scanning the original source as opposed to corrected versions, you scan the entire region instead of a cropped region, the media has generally been handled less often, and you can scan at higher resolutions. But to do so, you need a flatbed scanner and software that can properly identify each image on a film strip and can turn the negative into a positive. The 8800F comes with the necessary holders and light sources to do just this for up to 12 frames of 35mm film strip or 4 frames of 35mm slides at resolutions up to 4800 dpi (dots per inch).
What does this mean? Check the resource link at the bottom for references to learn more, but the essentials are that scanning a frame of 35mm film strip at 3000 dpi is roughly equivalent to a 12 megapixel camera and can readily produce a great quality 9×14 print. Scanning at the full 4800 dpi would be equivalent to a 33 megapixel camera. I scanned at 3000 dpi and saved the image as a full quality jpeg using the advanced tab on the supplied PhotoStudio software, and also applying the medium level of dust and scratch removal. The image at the right is an example of such a scan. (It wasn’t obvious at first that I needed to use the advanced tab. The basic tab used a lower dpi count — it didn’t actually specify what the scan level was, just what the target output printer dpi was — and/or more compression in jpeg, so the files were much smaller.)
When previewing images, each frame on the film strip will show up as a thumbnail. I then selected which were to be scanned. I’ve selected as many as 7 images so far, and then had to wait around 40 minutes for the scanning to complete. So it’s not a fast process, but it’s really nice to be able to do batch scanning of a bunch of photos, and then check back later for the results. Each frame shrinks down to under 2 MB, so I’m definitely loosing some quality using jpeg compression, but it’s still pretty decent for basic family photos. Plus it’s immediately ready to use online or in photo frames. I’d recommend saving as TIF files if this is something that will see considerable editing later or if it’s of significant historical value, but prepared to end up with much larger file sizes to avoid loosing data, somewhere closer to 30MB per scanned image. (A simple way to think of compression is 1.2222222 can be stored raw as that number using 9 characters — the number 1, the period, and 7 number 2’s, compressed without loosing data as 1.2[x7] using 7 characters, compressed loosing some data as 1.222 using 5 characters, or compressed loosing more data as 1.2 using 3 characters. TIFF may or may not compress and if compressing can do so without loosing data, while jpeg always compresses and does so loosing some or much data depending on jpeg quality chosen.)
The documentation provided was pretty straight forward to do the simple scans of photos and film. What was missing was whether it was better to scan film or photos, what dpi meant in reality for archiving when scanning film, and details about file formats. My references for figuring those issues out were: