1
\$\begingroup\$

TL,DR: What should I look for in a scanner for digitizing photographs from 10 to 80 years old, and various other media as secondary importance? Do you have any theoretical, or practical advice, or references to offer? If any further details are required, please let me know.

Details of media

I have a small collection (approximately 200) of photographs which I believe would be of value to those interested in the history, and architecture of Johannesburg. They were taken by my grandfather, and grandmother between 1950, and 1970. There are several other photographs which were taken prior to that, I think the oldest is around the mid 1940's, and a few which may have been from the late 1930's, photographer's likely unknown. These are what I am most interested in digitizing, and getting online. Many of the buildings no longer exist, and were taken for the explicit reason of documenting the architecture, and buildings, as well as land which has now been developed.

I have also got various photographs which are much newer, but mainly of personal interest. There are also various documents, books, news paper clippings, and other printed media which is out of copyright but currently unavailable online. Some of the books contain hand painted water colour. There are also water colour paintings themselves.

Unfortunately, I believe we no longer have the negatives, nor many of the slides which were also produced, but if any turn up, I would like to digitally archive those as well.

Copyright

Everything which is out of copyright, or which is within our legal rights, we are planning to upload to a relevant image sharing, or collection site. (Barring those of purely personal interest.)

Personal/budget/time trade off limitations

I am not really concerned about the time, or manual involvement with the scanning process, my main concern is accuracy and quality of the scans. Technically, there is no budget available for this, so we are aiming to keep costs low but do not wish to completely skimp, and go into this blindly. Properly archiving the originals is highly unlikely to be an option, and some of it is to be sent off to other family members.

I have spent a little time researching my options, but this is all far out of my area of any theoretical, or practical expertise and gaining the theoretical knowledge is not something I will have time to do for a while.

I have no image editing skills, nor anything to manage colour calibration. It is highly preferable for the scanner to have good Linux support, and for any software to be available for Linux as well.

Storage

The storage of the resulting scans which are either still in copyright, or of only personal interest is something which is well within my skill set, file size is not too much of a concern at this point. (For those interested in the details, I am currently basing storage on an HP N40L microserver running Ubuntu 14.04, a long run UPS, and ZFS as the filesystem in either a mirror, or RAIDZ configuration. For local browsing and display, I am undecided, though might roll my own. Long term I do wish to have tape storage.)

\$\endgroup\$
7
  • \$\begingroup\$ Long run: Blu Ray in a closed case in the dark. For the HP N40L go RAIDZ2 or mirror, don't trust RAIDZ. I leave the rest to the experts in the other matters. \$\endgroup\$
    – FarO
    Oct 16, 2015 at 11:37
  • \$\begingroup\$ As odd as it may sound, I'm still a little hesitant to rely on optical media. It's not out of the question though. This article is exactly why I wouldn't run either plain RAIDZ, nor RAID4/5. I won't be running a uniform batch of disks either. I keep an eye on SMART, and 'stress' test drives between any form of re-purposing. I will likely mirror, depending on the outcome of processor overhead, and disk, or controller availability. \$\endgroup\$
    – Phizes
    Oct 16, 2015 at 12:25
  • \$\begingroup\$ I have readable optical CDs and DVDs from >15 years ago... Buy Verbatim and keep them in the dark, in sealed or in carton boxes, they will last a lot. And BD should last more than the other types, since the wavelength is shorter and a way smaller percentage of the light you find around would be able to damage them. And the manufacturing has been updated. BD > DVD > CD. \$\endgroup\$
    – FarO
    Oct 16, 2015 at 13:05
  • \$\begingroup\$ Concerning the HDDs, you can simply stagger them: buy one, leave it running one month, add a second one, leave them one month and so on. When you add the last one, format the ZFS pool and start putting data on them. And use smartd.conf to send yourself emails when specific attributes change. \$\endgroup\$
    – FarO
    Oct 16, 2015 at 13:06
  • \$\begingroup\$ Scanner support on Linux is a tricky business involving binary blobs for most recent scanners, or was as of a few years ago, anyway. Before buying one based on it being mentioned as "supported" somewhere, you might want to search for others' experience with that model on the specific flavor you plan to use it with... packaging of proprietary drivers is something distros have widely varying approaches to. \$\endgroup\$ Oct 16, 2015 at 17:12

1 Answer 1

1
\$\begingroup\$

Any scanner will do provided it can scan transparencies as well as photographs. I have not run Linux since the days of the very early Red Hat distributions. I cannot help with scanning software. In the non-Linux world I would suggest something professional in the way of scanning software such as Silverfast. Vuescan is also an excellent piece of software and version 9.4.32 should run in Ubuntu.

Look here: http://sysads.co.uk/2014/05/install-vuescan-9-4-32-ubuntu-linux-mint/

You should find a list of supported scanners and you can pick the best of them by looking at their native resolution and whether or not they provide a transparency scanner hood with the device.

In terms of scanning resolution you will not need more than 300ppi if you want to print the photographic images at their original size and web display will not require more than 100ppi. Save the files in a format that will be transferable and useable later. Avoid .jpg files because each save will destroy a little more of the file and they will deteriorate beyond useable after several saves.

Suggest a scanner which can scan natively up to 2400 ppi. More is nice but it creates enormous files which you do not need. Another storage format you may consider is PNG. Alpha channel (transparent) support and lossless 16bit files are possible. Does not support CMYK so file printing can be problematic.

My preference would be to save files with as much information as possible so that would probably mean a lossless file format such as .tif and a 16bit file format so that reworking of the file would be as painless as possible. It would mean more file storage space is used though LZW compression is available. Consider a scanner which can save RAW file format because there would be no loss of information and the opportunity to rework the file would always exist.

As for RAID, I suspect it is far better to avoid it. If you use a single disk, its failure is less likely than if you use multiple disks because you are effectively increasing the risk of failure with every disk added to the storage solution. I would opt for a single disk, copy it and store the copy off site. I once had used a very large and expensive tape drive of 20GiB capacity. It was slow, sequential and when it failed, I was unable to retrieve any of my data.

I forgot to add that most of your file transfer post scanning in Linux would probably be via the GIMP software (the Photoshop simulacrum). I have not used that for 10 years and the last time I looked it could not use or process 16bit files. That may have changed but it is worth bearing in mind if you are going to keep the maximum detail in your scanned images.

Hope this helps

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.