Page 1 of 1
Recommended parameters for archiving photographic images
PostPosted: Tue Jun 17, 2003 12:50 am
by dpatterson
What are some recommended set scan parameters in SilverFast Ai 6 when archiving photographic images today for use with future imaging technologies? Current publications recommend an LPI setting and other scan parameters consistent with the resolution and other limitations of the output device/s in current use. When a different output device with higher resolution becomes available at a future date, the scan data in the current file will be less than optimal for use with the future output device, and it will become necessary to re-scan the original photograph to record the maximum detail available in the original image. In my example, however, it will no longer be possible in the future to re-scan the original photographic image and record the greater detail needed by the latest output devices. The original photographs will no longer be available for re-scanning. Consequently, all photographic image detail needed for use with future software applications and output hardware must be captured and archived by today?s scans with SilverFast Ai 6. What guidelines are there, if any, for using SilverFast Ai 6 in one or more scans to capture the maximum detail possible for archiving the image data for use with future versions of Adobe Photoshop or other image editors and various output devices? Should the selected Scan Type for such examples as century old black and white photographs be 48-bit colour, 48-bit colour HDR, or other selections? Should the photographic image be archived at more than one resolution, if so, what other selected resolutions or LPI selections may provide the greatest opportunity for minimizing resampling artifacts? How can SilverFast Ai 6 be used to capture the highest resolutions to an archival data file without degrading the details of the image with aberrations introduced by high resolution scanning?
PostPosted: Fri Jul 25, 2003 5:06 pm
by Gregory C
I personally would use the highest hardware resolution your scanner can handle and store the original scan data. with SF Ai, this would mean using the HDR image types; 48-bit Color HDR, etc. whatever the future devices are capable of, you can only scan as much as your current device allows. there's no point in trying to exceed this real limitation.
if you are scanning negatives, the HDR image types are probably still your best bet. even if SF is not around in say 10 years time, other new improved technologies will be available which will be able to read your negative raw data and convert it. in this case though, your chances of success would probably be better if you save a note with the HDR files describing the brand and type of film used. it would be nice if SF could save this data inside the HDR files as exif information but that's not currently possible.
saving the files as tiff or jpg will not make much of a difference to the result if you use the highest quality jpg compression. jpg compression is lossy but the result file size is much smaller. so the final decision will be yours; do you want the best quality image file or a very good quality image file which requires much less disk/disc space.
I have examined saving rolls of 35mm negative film to DVD discs. if I scan the maximum available image area and save it in the HDR raw tiff format, and if I add screen size auto-corrected images for preview/perusal purposes, the 4.7GB DVD is actually not quite large enough. (for the record, 35mm images are much larger than APS images.)
I have requested that SF's new LT feature develops a function that would allow us to preview negative images. if LT would allow us to specify the film brand and type for each image, and save those specifications within an LT archive file saved with the original images (in the contained folder; one archive file per folder?), and if LT could use that information to automatically revert and color-correct the negative HDR files on disk for preview, we wouldn't need to save screen size corrected images to disk.
any seconders for this request? any better suggestions?
to my limited knowledge, none of the currently available photo archiving applications support automatic reversion and color-correction of negative image files. there seems to be an opening for any application that could do this well. SilverFast LT?
as your question regarding 'aberrations' seen at high resolution, I would suggest that this is a hardware problem (quite normal for most high resolution devices at this time) and that SF doesn't have the ability to correct for these aberrations. therefore, you best bet would still be to store the raw data in HDR files and hope that a future technology will be able to see and correct the aberrations in the image files.
regards
Gregory
aberrations in scanner sensor
PostPosted: Thu Oct 07, 2004 6:16 pm
by jackbechtel
You have hit a touchy spot with remarks about high resolution scanner aberrations. A scanner has a row of individual sensors, each one of which has its own "profile." Thus a high resolution scan of a solid color can have "stripes" of variation all the way across it. Image processing software often exacerbates this phenomenon. I noticed it when scanning a very faded ektachrome image and trying to "bring the image out of the ether." I also noticed it in days of yore when scanning slides on an Epson flatbed equipped with a light attachment. Finally, it becomes increasingly apparent in an otherwise good scan as you zoom in on a 3600dpi image. You know - there's always that ONE pixel sensor that makes a nasty line straight down the enlarged image, usually in a dark part of the image. Ideally, these aberrations should be less noticeable than the individual grains on the film itself. In order to cancel out this striping, it would be necessary to re-orient and rescan the slide several times, then electronically overlay the resulting images on top of each other to "average" each image point's values. (You are doing this mechanically when you copy the pixel columns next door and paste them over the offending line.) Correct me if I'm wrong, but I assume this happens to a degree when using multiple-pass scanner settings. Now the point: Does or will SilverFast have the capability to create a profile that compensates for these aberrations in a single pass scan?
PostPosted: Fri Oct 08, 2004 2:12 am
by degrub
If you use a single pass multisampling scanner like the Nikons (cs4000, cs5000) you can reduce the reandom noise, but not a faulty sensor.
Frank
Archiving photo
PostPosted: Tue Aug 16, 2005 8:23 pm
by schrubs
Does anyone know a place (books, web sites, forums) where I can find more informations about archiving photographic scans?
Thanks
PostPosted: Wed Aug 17, 2005 1:30 am
by degrub