|Input/Output/Workflow Scanning, printing, color management, and discussing best practices for control and repeatability|
| ||Thread Tools|
I finally wore out my old Epson 2450. Blue channel suddenly started appearing offset about 1/4 inch and blurry, so all my scans looked like those 3D photos that require fancy paper glasses. Replaced with a refurb Epson 3170 ($124 from Epson) because I am cheap. No digital ICE and no Silverfast, but at that price I can go through a couple of them before getting to the $400 price tag of the 4170. Also managed to find an Acer Scanwit 2720S film scanner, with SCSI card, for about $99 on eBay last year.
Anyway, I figure it will take the better part of a year to get through this set. I am going to work on scanning one album a week, fitting the negatives and framed shots in when spare time appears. Putting my left brain to work and combining it with my innate sense of laziness, I have come up with the following workflow and would love comments on it.
.. maximum future print size around 8x10
.. backup medium to be checked and/or rewritten annually
.. Epson Perfection 3170 flatbed, Epson Scan in pro mode
.. Acer/Benq Scanwit 2720S film scanner, MiraPhoto software
.. Photoshop 7.0
.. lots of disk space and a dual-layer DVD burner
Workflow for each photograph:
1. scan settings:
...... 24 bit color
...... perform any major color or gamma corrections on the scanner
............in order to get a solid "digital negative"
...... size with short side close to 2400 pixels (i.e. 35mm => 2400x3200)
...... no unsharp mask, dust removal, or descreening at this stage
2. postprocessing theory:
...... as little as possible
...... major dust/scratch removal
...... still no sharpening before saving
...... any edits done on layers, preserving original scan
...... print only the really good ones at this point
...... store flat, one album per folder
...... save as TIF RGB, LZW compression
...... ICC profile: sRGB saved with file
...... filenames according to page & position on page, like this:
........0405.tif means page 4, position 5 out of 9 cardinal positions
........on page from upper left to lower right, 3 rows x 3 cols, rows first
............1 2 3
............4 5 6
............7 8 9
...... record in spreadsheet with the following data points:
.......... album name (or "box" if unorganized)
.......... filename, see above
.......... short description
.......... list of known people in photo
.......... date taken (guess is ok)
.......... date faith (range 1-5, 1=date is known, 5=date is raw guess)
.......... original (35mm neg, 4x6 print, etc)
...... print spreadsheet list on plain paper
...... contact sheets, thumbs about 1x1.5, 6 rows x 5 cols, caption=filename
........with album name in upper outside corner
...... two local copies of entire set on active drives
...... copy one album on a CDR, put in safe in garage
...... copy 7 albums per DVD, probably 3-4 DVDs total, put in deposit box
...... make copies for family members and send out as requested
Guessing at 150 photos in each album and about 4MB compressed per photo, that makes around 600MB per album. If there are 2 albums worth of negatives and another in framed photos, that makes 28 albums for a total of about 17GB worth of well-organized information, which will be a vast improvement over the 10GB of disorganized photos I currently have that also need to be gone through at some point.
The next project after this is the boxes of 8 years of 35mm family photos. I am already happily printing contact sheets of our digital family photos---we made the move to digital about a year ago and haven't looked back much. I will post a new thread about contact sheets soon.
This has gotten quite long, but I wanted to cover everything. I think it is very valuable to share workflows like this, and I look forward to seeing those used by others.
Last edited by kaulike; 01-28-2005 at 12:26 PM.
About the only things I'd suggest would be to skip the processing in-scanner (Photoshop has much higher quality tools), and a move to Mitsui gold discs. I might also skip the LZW compression.
Are you digitizing with the assumption that the digital files will outlast the paper originals? They probably won't, assuming the originals are moved to archival storage. Optical discs are not archival, even the Mitsui gold. If that's not the idea, nevermind
The main goal is to get organized and to be able to both restore and reproduce these images on demand, particularly in case of fire, flood, tsunami, earthquake, or whatnot (I live on the California coast, we have lots of potential for whatnot).
The paper and/or film originals are for the most part in good shape, although some of them were kept in decidedly non-archival storage for many decades, and some that we are not in control of continue to be kept in family members' garages, attics, etc. I have dragged scanners clear across the country in pursuit of archival safety. The ones we have are kept in relative temperature and humidity control, but I don't expect them to last forever.
It seems to me that the best method for preserving these things is not in choosing one medium over another, but by duplication and dispersal. Having multiple copies along with the ability to duplicate the entire set with very little effort seems to me to be the best security method.
That being said, the gold disks sound interesting. I'm doing some more research on them. I am already well-stocked on good-quality DVD-Rs thanks to an optimistic purchase last year, but I have almost no CD-Rs left.
Why skip the compression? It is fairly effortless, lossless, ubiquitous among image software, and compresses a 28MB file down to about 4MB.
In-scanner corrections, vs. postprocessing
Forgot to reply about in-scanner color manipulation vs. postprocessing.
I should note at first that this mostly applies to color scanning and my scans are almost all in black and white, so for my purposes it doesn't appreciably matter. That being said...
This is a subject of some debate and opinion. Normally I prefer postprocessing for almost everything. However, unless the scanner can provide the 48-bit raw data it sees for itself, there is some postprocessing done no matter what, and it makes sense to take control of it. The 24-bit gamut is fairly small compared to the number of possible colors or shades of gray. The Epson scanner software makes a 48->24 bit conversion before shipping the info off to a file or to Photoshop, and it is my impression that my best results will follow if I fill that 24-bit gamut with the data I am most likely to use. This follows the same logic as squeezing the histogram---with only 256 levels for each primary color, I don't want to sacrifice one end or the other to unused data.
In reality, all I normally do on the scanner is to squeeze the histogram and make any gross color balance changes, e.g. correcting a color cast.
Wayne Fulton has a great discussion about this on his web site at http://www.scantips.com, or at least used to.
Another option is to use VueScan, which for many scanners can preserve the original 48-bit raw data. I made a concerted effort to use and keep up with VueScan for about a year, but in the end I found I preferred the native tools that came with the scanner.
Thanks again for the reply
|Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)|
|Thread||Thread Starter||Forum||Replies||Last Post|
|Music from the '60's --The 3D's Poetry Album||CJ Swartz||Salon||33||03-14-2016 08:49 AM|
|Film scanner for home use?||bodegg||Hardware||12||05-20-2007 11:40 AM|
|Scanning workflow||Doug Nelson||Input/Output/Workflow||3||04-24-2007 05:03 PM|
|Scanning Workflow||Jon Foster||Input/Output/Workflow||7||04-04-2006 03:21 PM|
|High Bit Scanning?||Doug Nelson||Input/Output/Workflow||1||08-18-2001 02:06 PM|