Slashdot Mirror
Digitizing a Large Amount of Photos?
mcj0422 asks: "With what seems like the many increasing disasters, and also the freak accidents that can happen, there are certain non valuables that people end up losing, the main one being pictures that are printed on film. I know my mom has several thousand photos in our basement, which could be wiped out by water damage in one heavy rain season. Are there any scanners designed to take loads of pictures and turn them into digital files? Is there a service that does this, if so which ones would you recommend?"
Using a standard scanner and some scripts that used image magick to do some sorting and converting. I don't know of any bulk scanners other then network copiers. You could try using those. I presume you will be doing this in a linux enviroment.
Charles Wyble System Engineer
I'd go with the usual "teenager next door with too much time on his/her hands" approach. Five bucks an hour and all the lemonade he/she can drink.
Unless said photos are pornographic.. Then you might have a problem :)
"Better to be vulgar than non-existent" -Bev Henson
BritePix
Dig My Pics
Digital Memories Online
Digital Pickle
Photo Max
Slide Converter
I'm sure there are more services, but I'll leave the job of going to page 2 up to you.
Karma: SELECT `karma` FROM `users` WHERE `userid`=138474;
Most ADFs can feed photos too. Hp even made a scanner (HP Scanjet 5500c) Just for this purpose. Of course image management gets tricky, but picasa could probably be a good starting place.
I'd do something interesting, but my server can't handle a slashdotting.
We stack the pictures in, face down, they get fed through to a flatbed scanner. But I doubt you would be willing to pay what we did to get the device.
A GOOD digital photo store should have a similar setup.
Whether they will charge you a reasonable price with a discount for bulk is another matter.
excitingthingstodo.blogspot.com
Seriously if you are talking about a system that you can just pop a stack of photos on and have the process automated you'd be talkin' $$$$$.
Personally I really like epson scanners if yuo get a USB 2.0 or Firewire compatible one they are fast. Store the photos in a lossless Jpeg, RAW, or my preference in TIF. Back up to a RAID 5 or Mirror and then archive with RAR and recovery segments onto DVD-R.
--
So who is hotter? Ali or Ali's sister?
For over 1,000 photos, I suggest you buy a high end scanner for over $500 and pay a teenager $20/hour to do it.
excitingthingstodo.blogspot.com
If the images are 4x6 or smaller, you could scan them at least 3 or more at a time.
Just scan them into one large TIFF and do cleanup later.
That way you scan a pile or box at a time.
Cleanup is a bitch but worth it.
Just crop the images out of the TIFF, color correct and remove dust.
if you steal from one source, that is plagiarism, if you steal from many, well, that's just research.
Grab yourself a $40 USB scanner. Stick it on an old PC, and install your favorite GNU/Linux distro. Then use SANE and scanbuttond to bulk-scan your photos. Here's how I am doing just about the same thing:
I inherited my grandfather's QSO cards (W3FFZ) from the 50's. I figured I'd scan 'em and put 'em on the web. For the scanning process, I have scanbuttond run my script to scan a postcard size from the scanner, and toss it into a directory. So what I do now, is I go over to the scanner, put a QSO card in, close the lid, and press a button. The scanner scans the card and I can then flip it over and press the button again.
It is difficult to bulk-scan things in general. You really need to apply meta-data to your images, whether you populate any comment fields, the way you name your files, etc. I find it best to go through and rename my images as I take them out of my camera/phone.
Zhrodague.net - I do projects and stuff too.
Easiest? Pay someone.
Not too many people make photo quality digitizers that are affordable for the average joe. The teenager approach might work but you've got definate food-put issues to deal with- greasy fingers, etc.
You might find it easiest to make up a 4x6 template, some bright lights, a glass coversheet, 2x polarizers and shoot thru the glass with a macro lense and capture a 6mp image of the 4x6. You'll have dye mis-match issues for colour balancing but that would probably be your fastest route. (If you have a vacuum plate/table that would work better than the polarizer/glass method).
By far the hardest, costliest, riskyest, and most time consuming part of this process will be arranging a several-thousand photo collection to be scanned. If you are going to take that step, I'd recommend you arrange to wind-up with both a digitized copy and an old-fashioned one.
We have a good understanding of what it takes to preserve photos, with almost 200 years to learn from our screw-ups. We don't have the same experience with digital artifacts, and the experiences we do have says we're abysmal at it. Physical objects can survive thousands (millions?) of years by accident while we've all experienced the loss of digital ones which were important just seconds ago.
If these photos are important,
Also be aware that making a digital copy of some things (like a photo) can introduce threats which were not there before. A machine jam while scanning or improper handling of unstable photos can cause irreparable loss. I'd hate to see your precious photo collection lost completely to a freak minor auto accident or random theft. Also beware that digitizing a photo is a lossy process: no matter how high a resolution you have a photo scanned at, there will always be some information which cannot be recovered from the digitized version, should the original be lost.
And finally, understand that the simple act of making a digital backup of something like a photo makes the original a tempting target for disposal in the name of 'efficiency'. If everyone in the family has a digital copy of every photo in the box, it might be a lot easier to justify leaving the box in the basement for the termites. And once the box is gone, will you really care about your copy on your crashed hard disk, when you're sure you can get another copy from anyone else at the next reunion. Until you find-out everyone else was counting on getting a new copy from you...
The thing about things we don't know is we often don't know we don't know them.
I know this is over-budget for pratically everyone, but I just have to share.
My workplace recently replaced our venerable Fujitsu 4097D scanner. We ran hundreds of thousands of sheets through that thing, and it never needed service beyond my unskilled labor and Fujitsu's ScanAid consumable kits. But when the lease ran out, we chose to replace it with a color model.
Since the 4097D worked out so well, we looked at two of the current Fujitsu models. Both of these scan up to 600 dpi x 24 bit color (optical) and have hi-speed USB2 and SCSI interfaces. Both have flatbed capability in addition to the ADF.
The successor to the 4097D is the fi-5750C. It's roughly $6,000 and has a duty cycle of 8,000 pages per day. (They call that a "light duty" scanner, which cracks me up.) It also has a clever rotating 200-sheet 57 PPM ADF unit that makes it easy to use for both right- and left-handers. It can scan up to 12"x18".
The model fi-4340C is a bit more reasonable, going for about $3500. It can handle a slightly less huge variety of paper, and has a duty cycle of a mere 3,000 pages per day. It has a fixed 100-sheet 40 PPM ADF. It can scan up to 8.5"x14".
We purchased the fi-5750C. The hardest part of the installation was getting it upstairs... it's bulky and almost 80 pounds. Once I had it running, I took a small stack of mixed-size photos and dropped them in the ADF... it handled them wonderfully. Obviously a 600dpi 24-bit scan doesn't run at 57 PPM, but it's still pretty quick and it produced very nice-looking scans. Most importantly, the ADF didn't damage the photos.
One of these weekends I'm going to bring in a portable hard drive and a box of photos, and see how many gigabytes I can fill up in a day.
---
On a more realistic level, here's a couple things to keep in mind. First, scanning a photo print is making a copy of a copy. If you have access to negatives, try to scan them instead. I have no idea what equipment does that well, but I expect it's very expensive. It's probably best to work through a service for that.
Second, digitizing is the easy part... indexing is the hard part.
With reasonable men I will reason; with humane men I will plead; but to tyrants I will give no quarter. -- William Lloyd
i'm actually doing this right now
don't get the document feeders like the others say, they're made for documents, not photos.
get yourself an epson perfection 3490 or 3590. THe difference betwee the two is about 50 dollars, one has an automatic FILM FEEDER. I didn't think i needed a film feeder so i went with the 3490. Both can take a "Multi Photo / Business Card Feeder" ~$150. But epson doesn't want you to know the 3490 takes the multi photo, just so you can order the more expensive one. But it's in the manual on their support site.
The automatic photo feeder holds about 25. it does jam once in a while, but usually it's because the photos don't line up correctly. I scan my photos at 300dpi, each takes about 35 secs. The only annoying thing is it comes out reverse, so you might want to sort it backwards.
all in all it's pretty decent. the only bad thing is the dust problem. For some reason they don't make the higher end scanners with feeders. I think these higher end use some system to detect dust and remove it from the picture. So in the end, I occasionally remove dust from the flat bed and any noticeable ones from the photos.
HP had a similar solution but it seems to be off the market now and they rather you buy some very expensive solution instead.
the other thing is, you can also use the flatbed for multiple photos, it autocrops the pictures. throw 3-5 photos on the flatbed and it'll automatically find the pictures. I had some issues with it cropping too much, but it's still quicker than 1 by 1.
-khang
If you already have a good digital camera and you want to digitize prints, my advice is to photograph them. There are special rigs specifically for photographing documents, but it's actually fairly easy to set up yourself. Get a tripod that allows you to reverse the central stand, i.e. so it points down between the legs. Then place is on a desk, do some tests so that you have it manually focused correctly on the desk. Take photographs of graph paper to make sure everything is level. Also, get some good lights - it's just the bulb that is important - you want ones that are "full spectrum". Diffuse the light through something, or bounce it. If you do some tests you should be able to get it so the photo is very evenly lit.
The advantage with this setup is that once it is all correctly set up, you can photograph a lot of pictures very quickly. If you have a Mac, you can plug your camera into it and use the Automator to trigger the camera shutter so you don't even have to touch the camera and risk knocking it. You can even get the Automator to automatically crop/thumbnail/whatever the images.
The Nikon Coolscan line appears well reviewed. The best of the line, the 9000, runs ~$1700 on eBay, or ~$1900 new. If you don't need to do any medium format film scanning, consider the 5000, which operates faster. Once you've scanned everything you have, resell it on eBay. With luck, the only thing you'll lose is your time.
I'm planning on doing this in a couple of months.
As a photo lab tech, I occasionally get giant batches of photos or slides to digitze. We have an Epson 4870, which has incredible resolution and Digital ICE built in. Most of the time we just put as many photos as will fit on the scanner and set the selection to scan them all individually at sufficient resolution to get a good 4x6. You typically need a 5mb file to get a 300ppi 4x6, so you'll need to set your resolution to a level that will yield at least that size. Smaller photos will require a higher scanning resolution. Digital ICE will clean up some of the dust and scratches, but not all of them, and can sometimes double your scanning time. Once I feel like I have a good amount of photos to work with in Photoshop, I crop them all, usually manually, and then I create a few actions for color correction, exposure/contrast, dust and scratch cleanup, and sharpening. Usually Auto-Color Correction and Auto Levels will work just fine, but sometimes you need to do some fine tuning. You can either use these actions on each of the photos individually, or you can batch-process all of them at once. While I would love to have an Automatic Document Feeder, I don't think we get enough of that business to justify the expense.
There might be some freeware that does it as well, but Photoshop Elements does a lot of this. You can scan several photos in one pass and it will find, straighten, and save each photo individually. it has the color correction features of photoshop, but quicker for most uses. It has a nice sorting/cataloging setup and lets you group, add keywords, and archive to CD/DVD. (full disclosure: i work at adobe, but have nothing to do with the photoshop stuff.)
The big lesson was that forsight is required. The hard drive would have been best to run with, but it's fragile, so pack them well. A CD book good, but heavy unless you move to DVD. As usual, having multiple live copies is the easiest solution.
Everyone's pictures are important, so digitize them soon. My digitized pictures are outlasting the ink in my physical versions. Even older silver based black and white images are going away. Digitize as quickly as possible and store the originals as well as you can - correct humidity, acid free backing and all that. Real dissasters can and will take your physical copies. Give gift CDs to friends and family of the images you think are most important. That will protect you against fires in a way that is too expensive and time consuming with physical coppies.
I'd recommend you arrange to wind-up with both a digitized copy and an old-fashioned one.
Is there a way to end up with less?
Friends don't help friends install M$ junk.
I will promise that data associated with the photo will be lost. The only SURE way to keep the data with the scanned photo is to attach the photo to a sheet of white paper, type the data on it (names, dates, locations) and scan the whole thing. This makes for a very large file, but on the pluss side, the white paper gives you an accurate whitepoint in the photo.
Here will be an old abusing of God's patience and the king's English.
A scanner can take more time, but it's worth the effort. Kooka works as well as the best Windoze software with them and you can scan in several photos at once. The quality, at all resolutions, is better than the camera method. It takes time to split them up, but you can save that task for later. With proper equipment layout, this can be almost as fast as the tripod way. You can find tough old HP scsi scanners at used computer shops and buy them for a song. Many USB scanners also have good sane backends. See the list of sane devices before you buy.
The easiest thing to do, is to use a photo scanner like this. I'd rather use kooka than a script but scripts are flexible and powerful.
Friends don't help friends install M$ junk.
I've been working on this project off and on for months. I bought one for around $500 from eBay, and plan to sell it once I'm done. I finally finished scanning recent history, and am as far back as college. After that, I'm taking on dad's slide collection. At this rate I should be done in about 6 months.
:)
This is not an urgent thing for me, but I'm pretty happy with the scans I'm getting. I settled on jpeg and decided to live with scans that end up ~10MB/photo. In raw format, at the highest bit depth, I'd be at over 140MB/photo. Disk is cheap, but not *that* cheap. What I'm getting is darn good, and will more than suffice. I'm not a professional, tho.
Whenever I walk past the computer, I swap out the negatives and queue up four more to be scanned. It takes a while, but my life is not consumed with this effort, either. Thankfully I switched to a digital camera about 4 years ago, so all new pictures are already digital. I know that the scope is limited to the one big container of photos and negatives I've got left to scan.
As I finish a group of pictures, I move the scans to a directory with the rough date of when the photos were taken. I then run a batch utility to update the EXIF data with the "originally taken" date. iPhoto works VERY well with that.
When I'm done I plan to sell the scanner on eBay. We'll see how that works out.
But Herr Heisenberg, how does the electron know when I'm looking?
I've been considering this for a while now. My recomendation is to scan the negatives. Wet drum scanners are the best, but it may be prohibitively expensive, and possibly overkill. You do have your negatives right? The big problem I ran into looking at film scanners is that the vast majority of the good ones are 35mm only. My parents have a lot of photos in a variety of formats. 110, 117 roll or medium format, along with 35mm. All those need scanned, which means for me I'd have to go with a flatbed, and use a jig/matte for differnent formats. You'll also need a flatbed to scan photos that you don't have the negatives to. Color correct software, especially software that can correct chroma shift in old color prints. You'll need automatic dust removal software as well. Check out photo.net. That site has all sorts of info, but it is heavily tilted towards pros.
The other thing I've been wanting is photo organization software. You know query for who, where, and when. I guess I'd have to make a custom DB system for that.
OK, this may not be popular, but realistically don't even bother. Nobody really wants to see all those photos. Maybe you do, and your brother and sisters - but I can guarantee none of your kids will. Not 1000s. Maybe keep a couple of good ones around, give one each to the kids - that is all you need. You think these things will be cherished for generations to come - that is just kidding yourself.
Don't mean to be cruel, but this is realistic. Before you take on such a task, ask yourself is it really worth it. The most cherished these photos will ever be is by your mum, and then less so for each generation after that.
Whatever you do, you're probably not thinking enough. Archival is hard.
Here is what I'm doing, and the rationale for why
First, I bought an Epson Perfection 4990. It does 48-bit negative scans with Digital ICE for dust removal. The scans of a 24 exposure roll take 180 minutes, but trust me, its worth the time to not have to spend the time in Photoshop removing the dust by hand.
You are scanning negatives right? Photos are great, but they are a small dynamic range snapshot of what the camera actually recorded. Scan negatives, and scan them in high bit depth, otherwise you're not really archiving digitally, you're making a lossy copy. You want to be able to make large prints, not just 4x6's forever.
I use the bundled scanning software, but other packages are probably as good or better. Each picture is numbered sequentially, and the negatives are moved after scanning into an archival binder with non-PVC negative protector sheets, and each sheet is labeled with the range of image numbers. This is important, as you *will* need to go back and rescan a few images for some reason at some time, and the negatives themselves will last longer than the digital media if neglected.
Now, as for the format, I'm encoding to JPEG 2000, which preserves the lossless, 48-bit image, at 1/3 the size of a tiff. However, no software really uses it, so each DVD includes a Windows and Linux statically linked build of the converter.
Each image group is burnt onto two DVDs, one DVD-R, and one DVD+R, from two different reputable manufacturers. DVD reliability is all over the map, and you don't want bitrot taking out one brand. Burning in different formats mitigates the risk that one format stops being as readable in the future. Each DVD also includes parity (PAR2) files, about 5-10% of the disk depending on how full they get. This allows you to verify the disk is intact, a step you should do to all the DVDs once every couple of years. If the disk is starting to fail, you can copy both DVDs to harddrive, recover from parity, and burn anew.
Each DVD set is a mix of half DVD-R, half DVD+R (eg Disk 1=-R, Disk 2=+R, etc), and a set is sent to my parents for safe keeping, and one set stays here. I've sent the negatives home too, since they live in a safer climate.
Finally, the useless master DVDs with JPEG2000's are nice, but people really want to *see* these images. Here, metadata is key. Make sure each image is at least tagged with basic metadata in the Dublin Core set, like date, subject, location. I'm doing that as a baseline, and adding Flikr style tags to all the images, and Getty TGF tags for locations. The images are then converted to JPG at HDTV resolution for viewing, and I'm writing a viewer application for searching, and all of this will go online.
Its not an easy project, but its really rewarding to come across old photos and know that they won't sit in a photo album or shoebox unseen, and future generations will have something to look back at. Have fun!