Image File Backup

By Carey Brown

Table of Contents

Passé. 1

CD / DVD.. 1

External USB Hard Drives. 2

Image File Formats. 3

I’ve seen so many forum postings on image file backup and archiving schemes that I thought it might be a good time to jot down some suggestions based on my experience with trying to maintain backups of (at this point) over 9 years of digital images. Most of this applies to non-image files as well. I’m currently set up on a PC running Vista but most of this should apply to other operating systems.

Media

Passé

Tape and removable hard disks (such as the Iomega Jazz disks) had their day (and their problems), but have now been replaced by newer technology.

CD / DVD

CDs (DVDs are also included in this category) are a cheap way to backup files.
CD drives with the ability to read CDs are ubiquitous. Only the oldest of machines would be lacking a CD drive. Drives with the ability to read DVDs have been standard fare on machines purchased within the last four years. The drives on older machines may not have the ability to write to CDs or DVDs though. External drives to read and write to CDs and DVDs are readily available at low cost.
Always burn more than one copy and keep one copy off site in case of fire.
They are only “mostly” reliable. There have been numerous reports of “bit rot” after as little as 2 years of sitting on the shelf. What makes “bit rot” particularly insidious is that there is no way to know when one of the files on the disk has been ruined by “bit rot” until it’s too late.
They are write-once/read-many devices. This has the advantage that you know the contents haven’t been modified since you burned the disk. A problem that I ran into with this is that I spent some time updating the keywords and description in my image files and had to re-burn copies of my CDs, and then I either had to get rid of my previous CDs or somehow keep track of which one contained the updated images.
There are re-writable CDs but my early experiences with them showed them to be unreliable. They (and the drives) may have improved but my distrust lingers. I do use them for temporary work such as creating a DVD slide show so that I can review it and make changes prior to the final burn on to standard DVDs.
Because of their “limited” capacity (funny, this used to be considered a lot of capacity) CDs are not usually suitable for system backups. For data backups you may find that you have to pre-group your data into directories that will fit on a CD (or DVD for that matter). Note that there are some backup programs with the ability to span multiple disks.
Keeping track of what files are on which disks is problematic, especially when you may have different versions of the same file on different disks.

External USB Hard Drives

(This also applies to Fire Wire drives. See ‘Feedback’)

The price per giga-byte of these devices makes them Ideal choices for backing up.
Make sure to have more than one and keep one off site in case of fire. Hard drives are not so much prone to “bit rot”, but to catastrophic failure, hence the need for multiple backups.
All new USB drives use USB version 2.0. If you have an older computer it probably only has USB 1.0. This will still work but the transfer rate is twice as slow.
They are also available with large capacities. I have drives that will hold my entire 9 year collection with room to spare. This is ideal for when I want to use a program that will search entire directory trees for images containing a specified keyword and/or description.

Raid

Raid drives configured for redundancy are a good choice for your desk top computer but should not be considered a substitute for backups.
It does not give you the ability to have off site backups.
Raid provides a sort of continuous back up. This is not my ideal choice because it provides no safety net for an “oops”, such as mistyping an edit on one of your files. For that type of file recovery you want a backup system that runs on command (or schedule) as opposed to continuously. Of course, it’s then up to you to remember to run the backup program.

Software

There are two basic categories of on-command backup software: system, and data.

System Backups

System backup software usually comes bundled with your operating system.
Designed to backup entire systems including program and system files that are typically inaccessible to other software.
The file formats used are proprietary and the contents apply only to the machine on which the backup is made or an exact hardware duplicate. You may not, for instance, do a system backup on a Windows XP machine and expect to restore it to a Windows Vista machine.
System backup is a very important step and provides a way to restore your machine if the hard drive needs to be replaced or re-formatted. Usually your system files do not change as often as your data so a system backup need not be run as often as data backups.
Full system backups may take several hours depending on the size of your hard drives and how full they are.
System backup software often comes with the ability to do incremental backups. Incremental backups only backup those files that have changed since the last running of the system backup software; therefore they are much quicker than a full system backup. A major issue to watch out for when using incremental backups is that they don’t track files that have been deleted. When you restore from a system backup and subsequent incremental backups you may find files that you had deleted magically reappearing.

Data Backups

Data backup software does not typically come bundled with the operating system. It must be purchased or downloaded separately.
Data backups do not include directories that contain system or program files.
The backed up files and directories should be copied in a non-proprietary format. There are software products that may use proprietary formats, but I would avoid these, they may be useless if you want to restore to a different machine. A non-proprietary approach will copy directories as directories and files as files such that you can use standard programs to list, view, and access them. One format that I include in the non-proprietary category is the ZIP file. ZIP has become so ubiquitous as to qualify for non-proprietary status.
Some data backup software has the ability to do incremental backups. Some include the ability to track deleted files so that they don’t reappear when you restore them.
Because data changes much more often that system files, data backups should be done much more frequently.
A quick-and-dirty form of data backup is to use Windows Explorer to drag-and-drop data directories from your local hard drive to your external media.

Image File Formats

If you backup copies of your image files in either TIF (TIFF) or JPG (JPEG) formats you can rest assured that your files will still be readable for many years to come. For any other image file format, this may not be the case. This is especially true for RAW formats and formats specific to your editing programs, such as Photoshop and Paint Shop Pro. These are proprietary and are subject to obsolescence.

When I shoot using the RAW format (which is most of the time) I always run a batch process to generate JPG copies. I prefer JPG over TIF because the JPGs files sizes are so much smaller than TIF. JPGs have gotten a bad rap because they use a “lossy” compression technique. I have found that the losses are imperceptible as long as you use the highest quality setting that your editing software allows (that’s 12 for those of you with Photoshop). I always keep the original RAW file around in case I want to start over with the editing process; the RAW files contain more data. Likewise, when I create a Photoshop file that is multi-layered, I always save a non-layered version in JPG format.

Bit Rot

Bit rot is the unintentional modification of a single bit of data, usually as a result of passing time or environmental conditions.

None of the backup software that I’ve come across has support for validating the integrity of the backup files after time has passed. Some programs allow you to verify the copies at the time of backup, but that’s not the same thing. Some devices store the data in chunks that have a number stored with them that is derived from the data (typically called CRC or Cyclic Redundancy Check) that allows the detection and possible correction of minor instances of bit rot. Unfortunately, the devices do not warn you that bit rot has been detected and as the damage increases will eventually return corrupted data, again, without warning.

I would like to see a backup program that will store some sort of check number for each file on the backup and allow you to go back, say, 6 months later and validate the backup files against the check number, and tell you about any corruption it found. I usually have more than one backup copy so if I’m made aware of the corruption I can retrieve the data from the other copy. This would also let me know when it’s time to throw the device/media away and not rely on it.

There are some utilities that will compute and verify files against a check value (MD5 is one) but one that is integrated into the backup software and is aware of incremental backups seems like an obvious and necessary enhancement. If anyone knows of such a backup program, please let me know.

Miscellaneous

Fire Safes

Some people choose to use fire safes instead of keeping a backup off site. Be aware that most fire safes are only rated to protect paper not computer media. If you intend on going this route, look for a safe that is “media rated”. These are much harder to come by and you won’t find them in your typical office supply store.

Resources

One data backup program that I liked is called “Karen’s Replicator”. It is available for free at:

http://www.karenware.com/powertools/ptreplicator.asp

The Replicator has an option for mirroring file deletions as well as additions and modifications. The Replicator only copies the files that have been changed, making for a fairly quick backup.

Feedback

David:

The writer mentions external USB hard drives but did not mention Fire Wire drives. We have two external FW drives and they fly. Depending on your current computer and future computers I would suggest a drive that works with both FW 400 & FW 800. There are drives that have ports for USB, FW 400 & FW 800.

Karl:

Since you don't mention Win or Mac, I'm assuming Winblows. I've been using a tool called SyncBack for quite a while and love it. You can grab a free version at
http://www.2brightsparks.com/freeware/freeware-hub.html
Each Profile (as they call it) can be set as a full or partial BU, you can set a date range, include/exclude folders and files, use FTP, and lots more. You can even add a Group profile that runs 2 or more other
profiles. A nice easy way to do various folders.

Glossary

Proprietary –Unique to a specific manufacturer and may also be unique to a specific product and possibly the version of the product. The specifications are subject to change based on the whims of the manufacturer. Proprietary data formats, typically, cannot be read by software from other manufacturers.

Bit Rot – The unintentional modification of a single bit of data, usually as a result of passing time or environmental conditions.