>We have been scanning our records into our new EDRMS and finding that the
>files are of considerable size.  Does anyone know of a tool that is used
>at the scanning stage that will compress the files down in size when
>sending the images to our EDRMS.  Someone recommended JPEG2000, but it is
>my understanding that this is used for photo's or GIS type documents. We
>have been storing the documents as TIF's and want to continue storing them
>this way.  Any solutions?

John-

You can use JPEG for office documents, but it has some limitations in terms
of legibility.  TIFF typically uses Group 4 compression which is a lossless
algorithm - it preserves every pixel.  TIFF's compression algorithm works
best on horizontal lines which is one reason why tools that deskew the
image prior to compression can get greater compression than on images that
are deskewed.  JPEG is a more complex algorithm designed for color, but the
compression algorithm works on just about anything.  The key behind JPEG is
that it is a lossy algorithm:  it discards pixels in order to make the file
sizes small.  It isn't really a problem in most photographs, but the
algorithms can create distortions in text in certain circumstances.  We
generally do not recommend JPEG for office documents that are primarily
comprised of text.

There are other algorithms that use "wavelet" technology such as DjVu or
generalized chen transforms.  Some of these algorithms can be lossless and
work very well, but they are significantly hindered by the fact that they
require specialized viewers (with the decompression technology built in)
and have very low acceptance rates on the desktop.  But these tools are
ideal to reduce the amount of network traffic.  In that situation, you can
scan to TIFF in a remote location, convert to DjVu, send it across the
network to the storage place, decompress from DjVu, and the recompress back
to TIFF.

Hope that helps a little.

-Dan


Dan Elam
Vice President
804-342-7400
Visit www.eVisory.com for unbiased information regarding document and
records management technology.

List archives at http://lists.ufl.edu/archives/recmgmt-l.html
Contact [log in to unmask] for assistance