If you found this page useful perhaps you might like to join WSRS, click here to find out more

Digital Backup strategies

Many sound recordings are held as computer sound files, typically uncompressed wav audio data. Unlike analogue tape recordings, these files are ephemeral and the reliability of the storage media is often poor. This is because the high data density of digital storage makes small physical blemishes wipe out a lot of the recording, and because the data is often stored on highly complex active systems like hard disks and memory cards. Hard disks consist of many active components all of which must work correctly in order to retrieve the data. A power surge can destroy the disk controller rendering your data inaccessible, even though it is still on the magnetic medium. Very expensive data recovery companies can scan the disk platters, but the price tag is high.

Fortunately, digital data can be copied without generation loss, and the theory of effective backup is simple. Duplicate and tripicate your data, minimising common elements in the data carriers. That means using different technologies, eg hard drive and optical disc.

The theory is fine, but the devil is in the detail. Backup is boring and uninteresting, until you find you've lost your precious one-and-only recording. It then becomes fascinating... What we need is something that is relatively easy to do to create a duplicate of your data on a day-to-day basis, followed by a less-frequent process which safeguards your data against some of the less-likely risks. You can do this many ways, I propose one which is reasonably low-cost, but it is only one of the many ways.

first line of defence - USB hard drive backups

Get yourself a USB hard drive with its own power supply. Unless you are an extremely active recordist, something 160Gb or more will probably be enough for a year - equivalent to 200CDs or more than 250 hours of 44100kHz sampling rate stereo (at about 606Mb/hour)

As you download a recording to your PC from the recorder, you will have a file on the main hard drive of your PC, hopefully in a directory where you can find it again. Copy the new file or files to the USB hard drive preferably using a new directory for each session, either numbered 001, 002 etc or by date, in yymmdd format eg 070131. That way the directories list in consecutive order, and a file rec001.wav recorded today does not have a chance to overwrite rec001.wav recorded yesterday. Do not use backup or mirroring software to automate this as an image of the directory on your computer's internal hard drive. One of the threats to your data is you - accidentally erasing an original recording on your PC. You don't want to mirror that.

This duplicate safeguards you against the most common hazards to your data, which are accidental erasure/overwriting, general computer woes/crashes, and hard disk hardware failures. Because the USB hard disk is separately powered from the PC, you have some resilience to a PC power supply failure that fries the internal hard drive, or USB drive power supply fault frying the USB drive (you main PC will be the backup in that case). However, if your mains supply takes a direct lightning hit or you suffer a voltage surge you are likely to lose both hard drives. This is where you need to add a different medium -

second line of defence - offline optical storage

Every six months or so, you should then copy the data from your USB hard drive onto optical discs, CDs or DVDs. CDs have a longer track record than DVDs but store a fifth as much. If you use CDs store your data as wav files to take advantage of the extra error correction of CD-ROM compared to CD-audio. If you use DVDs you will do that anyway. Make two copies of each disc, but use different brands of disc. That way if a particular brand has a manufacturing problem that shows itself in time, you have some chance of having an unaffected copy. Store one copy in a different physical location from the other to hedge your risks of fire, flooding and theft. It goes without saying that you should use the verify feature of your CD writing software. This reads what is on the disc you have just made and compares it with the source data on the hard disk. It is better to find out this is not that case ten minutes after making the CD than two years afterwards, even if writing each CD takes longer.

Media handling and storage

The offline optical storage becomes your main backup after the transfer from the USB hard disk, which you can then clear ready for reuse.

the long term

Three issues affect whether your media store is useful in the long term.

long-term stability

You should test every two years that you can copy the files from a sample of your oldest media in service (from both media makes used). If you start having trouble, you need to copy these discs to new media. This is where having two copies of the same data is useful. You may find that utilities which mount the disc as an iso image will pull more files that the operating system file copy commands you usually use.

format longevity

For how long will you find hardware and computers that read CD-ROMS, and software that understands WAV? Although it feels now that CD-ROMS will be with us forever, less that twenty years ago I wrote my MSc thesis on an Atari 800, saving it onto a 5 1/4 inch floppy disk. The paper copy in the University library is still readable. The floppy disk is not - the hardware and software is hopelessly obsolescent. CD ad DVD will not be common data formats for ever.

indexing

How will you find your March 2004 bittern recording in March 2024, faced with 500 sequentially numbered CDs? It is not enough to have the backup - you need to be able to find it too.

prev
Editing
guide top

  1. NIST CD and DVD Care and Handling Guide
  2. MAM-A (disc manufacturer) Recommended Handling & Storage Conditons for CD-R and DVD-R Media

Valid XHTML 1.0 Transitional Valid CSS!