Photo scanning workflow

Stustaff · April 6, 2009

Any help with this would be welcomed.

Im planning to scan and upload to smugmug all of my oldish photos, they are all colour and between 29 and 10 years old.
My current thoughts are...

Scan
Crop
auto Photoshop some sharpening
manually colour balance
Save as digital file

anything im missing or that will make this run quicker?

My priorities are speed then quality, im happy that the majority will not be printed and for those few that may be I would do a more detailed piece of work on.

EDIT:- oh and actually what resolution should I scan and save in?

BinaryFx · April 6, 2009

I would probably do something similar to:

1) Scan in (12, 14) 16 bpc at the max. optical resolution of the scanner, sharpening turned off. Large files often result, some may be OK with this, while others may decide to drop the resolution or the bit depth after working on the file in Photoshop, before doing the final save. Some may make use of all the scanner tone/colour controls to optimize the scan at capture - while others may scan "flat, untoned" with zeroed settings making sure not to clip the histogram, then do all the work in Photoshop.

2) The scanner would apply the scanner ICC profile for the media in question (reflective or transmissive and the media brand/type) or you would manually assign this colour profile when opening the image into Photoshop.

3) In Photoshop, convert from the device profile (scanner) to a suitable working space profile for RGB editing (ECI RGB, Adobe RGB etc).

4) Dust spotting and or other restoration may be performed (often on a separate layer).

5) RGB edits would be performed, setting endpoints, gray balance, contrast, colour etc (often using adjustment layers).

6) Some may prefer to create a sharpening layer at this point.

7) This image would then be duplicated/merged and then resampled down to a smaller print size/resolution for output, with appropriate sharpening applied. Output may be various types of print or monitor based web browser display. The duped file would be converted to the ICC profile of the intended output device from the working space profile.

Most keep a separate master layered working scan that has as much data as possible (resolution, bit depth), then they dupe/rename the file for resizing and sharpening for the intended output which may require less resolution and bit depth.

Depending on the image, cropping may or may not be performed early or at all in Photoshop (cropping may be done in say Adobe InDesign).

Regards,

Stephen Marsh

http://members.ozemail.com.au/~binaryfx/
http://prepression.blogspot.com/

Stustaff · April 6, 2009

Wow awesome thankyou,
one thing I dont understand though,

"Scan in (12, 14) 16 bpc at the max" I thought it was measure in DPI?

BinaryFx · April 6, 2009

Stustaff wrote:

Wow awesome thankyou,
one thing I dont understand though,

"Scan in (12, 14) 16 bpc at the max" I thought it was measure in DPI?

I was referring to bit depth.

8 bits per channel, RGB = 24 bit RGB image
16 bits per channel, RGB = 48 bit RGB image

Scanners often produce less than 16 bpc (and often less than the marketing specifications), however in Photoshop anything higher than 8 bpc is "sort of" handled "sort of like 16 bpc".

This is separate to the pixels per inch or "resolution" of the scan. I was referring to capturing the most data (bpc) at the highest native resolution of the scanner (maximum optical resolution, not interpolated resolution). Dots per inch is for output, however I was not referring to output, input is described as pixels per inch (or perhaps less commonly as samples per inch).

Stephen Marsh

http://members.ozemail.com.au/~binaryfx/
http://prepression.blogspot.com/

pathfinder · April 6, 2009

Stephen,

How long would you estimate each frame will take from beginning the scan to saving the final edited file?

I think lots of folks anticipate scanning like this as a 30-60 sec event, and I find I spend a lot more time than that, if I want better quality than a totally automated procedure will get for me.

Comments?

TonyCooper · April 6, 2009

You haven't mentioned software. I scanned hundreds of old snapshots using Photoshop Elements 5.0 and an Epson scanner following the Import route. This allows you to put several (usually 6) photos on the scanner bed. Elements will then separate the photos and straighten them. Much faster than scanning one photo at a time.

At the time I did the project, I didn't own CS4. I now own CS4, and note that CS4 also allows you to scan several pix on the scanner bed in one pass, and then will separate and straighten automatically.

Using the right program, and importing to that program (as opposed to using a scanner program only) will save you hours.

Thunder Rabbit · April 6, 2009

Howdy.

I'm in the middle of the same process. I have a new Smug Mug site and lots of old 25mm slides. Quite a few of my own, plus 22 slide carousels (100 slides per) of family photos I'm doing for my wife's family estate.

BinaryFix gave a great response, but I thought I would go ahead and share my workflow. I figured this out by the seat of my pants (my usual method), so it probably won't be adopted as the industry standard. But it works for me.

SCAN (Epson 4490)>DEVELOP, CROP, EXPORT TO FILE (LR2)>ADD FRAME (Bridge, batch process in Photoshop)>EXPORT TO SMUG MUG (Bridge)

Scan Settings:
(Scanning 4 at a time)
24 bit color
2400 dpi (increased resolution didn't give better results)
Unsharp mask: Low (Theoretically, this should be none, but I did tests, and low sharpening gave the best overall workflow results for me.)
Dust Removal: Medium (Tried all the settings. This gave best results for my tattered slides.)

File type: TIFF, no compression

This yields images at 1560x1047 (1.5MP). Big enough for my purposes, plus any bigger, and all those years in storage units in Utah become much too obvious.

I don't scan into Photoshop. I use the scanner software. It works great. It has all the options, file naming, file type, etc. This leaves Photoshop free for other work while I'm doing the scans. When it's done, all my slides are in one folder, appropriately named.

DEVELOP:
I import the above folder into LR2 (or just open it if you use Bridge). I sort, develop, and crop. (I have develop presets for scanned slides that are automatically applied when I import them.) The scanned TIFF's are treated as RAW files and are not changed by editing (same in Bridge, ACR). You don't want to screw up your original scanned files any more than camera captures. A capture is a capture.

EXPORT TO FILE:
Once I'm satisfied with my images, they need to be converted to JPGs before I send them to Smug Mug. I have a Smug Mug file heirarchy on my hard drive that mirrors my galleries online. I export my selected files into the appropriate folder using LR2's export function. It will do everything you need, resize, rename, watermark, etc. (You can do the same in Bridge, except maybe watermark).

Now I have the original scanned files where I scanned them in, and I have the files with RAW adjustments as jpgs, cropped and resized at 72 ppi, with watermarks, in my Smug Mug folders.

At this point, I switch to Bridge. Two reasons.

One, the export to Smug Mug function from Bridge is the bomb. It's on the menu, under Tools>Photoshop Services. One click takes you straight to your Smug Mug galleries.

Second, I usually like to add a frame and inner shadow to my images before I upload them. They look a little nicer on the web. In Bridge, with all the images selected, I go to Tools>Photoshop>Batch, and run a custom action I made to add the frame. Just takes a couple of clicks, and a minute for Ps to do the math, and now all my jpgs have frames and inner shadows.

The images are still selected in Bridge, and I just click on "Publish Photos with Smug Mug". It's like falling off a log.

It's easier to do than to explain, that's for sure. In any case, whatever your optimum scanner settings, save your scans in a non-compressed format, then treat them as raw files, just like you would a digital photo. Then export them to a folder as jpgs, resizing or whatever in the process, before you export them to Smug Mug.

I also use the color labels to help me keep track of where a batch of images is in the flow. It helps a lot when you start handling hundreds of images.

Anyway, that's all I got, and half of it's probably wrong. But it seems to work for me.

PS: For what it's worth, I use five minutes per slide as a rule of thumb. This does not include the above preparation for Smug Mug. Just scan to completed first edit. It does include key wording, star rating, trying to remember who's in the slide, where you took the slide, daydreaming about that day, time wasted trying to save a favorite slide that's too far gone...

colourbox · April 6, 2009

My workflow is a lot like Thunder Rabbit's, except I export my processed JPGs, and drag those straight from Lightroom into a Smugmug gallery uploader.

The one step I didn't see was adding keywords, captions, other metadata to the scans so that they are all cataloged and ready before they're uploaded.

In Lightroom I set up a Watched Folder so that as scans come in, they pop up in Lightroom automatically. That can also be done in Bridge just by leaving a window open to whichever folder your scanning software is dumping the finished scans.

In Lightroom I will often make a tone curve for a particular roll of film. It's usually good for most of the film of the same type shot around that time, so I save it as a preset and apply it to all the scans to get a good first pass, then tweak individual images as needed.

Also it helps to create metadata templates so that you can apply copyright, cities, states, countries, etc. to many scans at once. Again, this can be done it Bridge or other decent software.

An optional step is to use a utility to add GPS coordinates to images so that they show up on a Smugmug map. I use Geotagger, which lets you use Google Earth to set the location.

BinaryFx · April 7, 2009

pathfinder wrote:

Stephen,

How long would you estimate each frame will take from beginning the scan to saving the final edited file?

I think lots of folks anticipate scanning like this as a 30-60 sec event, and I find I spend a lot more time than that, if I want better quality than a totally automated procedure will get for me.

Comments?

With scanning, one gets out what one puts in! If you put a bit of work into the scan you will get a better result than with lesser effort. As with other aspects of life, it is all about making trade offs. It will also depend on the model and type of scanner, scanner features, single or batch scanning, the original media and it's condition etc.

Simply having dust free, clean originals and scanner saves a lot of work in Photoshop. Some scanners have an IR channel that helps with automated dust and scratch removal.

It will also depend on whether the scan is a "digital archive" or just a one off use scan for a known print size/condition, which will dictate the scan settings and post processing time devoted to the image.

The amount of scans to be performed and the general time and budget all come into this question pathfinder.

So...this can take a lot longer than a minute!

Regards,

Stephen Marsh

http://members.ozemail.com.au/~binaryfx/
http://prepression.blogspot.com/

colourbox · April 7, 2009

Scanning is pretty far from a one-click operation unless you're just reproducing snapshots as snapshots. Most of the things that slow me down have to do with adjustments between rolls of film that I am scanning:

Forgot to change film roll/frame numbering of file names
Forgot to change film type profile
Forgot to reset crop after deciding to scan a portion of specific frames
Exposure is different than the previous roll of film, need to run a couple tests to avoid clipping
Exposure is different for one frame that was over/underexposed
Scanning software starts acting funny and needs to be restarted
Frame offset needs adjusting for a filmstrip that was cut kind of funny or at an angle
Film is in poor condition so defect removal needs to be cranked up
Film is curled and is not feeding properly or focusing properly, need to scan frames one by one with the manual film holder instead of the automatic feeder

Frames that don't give any trouble can be scanned in "no time" at all, but the main delays are from a few problem frames per roll that really kick up the average processing time, like a frame that takes 20 minutes to fix. I think we all know it isn't hard to lose an hour playing with an image. If you have just one of those hour images in a 36-exposure roll, your average time per frame rises to 70 seconds per frame even if all of the other frames in the roll were done in 60 seconds each.

Time reality check: Just 20 rolls of film at 36 exposures - 720 frames - will take 12 hours if you are lucky enough to average 60 seconds per frame (scanning time and all corrections included) and do not take breaks of any kind (eating, bathroom etc.). Needless to say, I'm way behind on scanning the archives...

Salviati · May 18, 2009

Thanks to everyone for the great information already in this thread. I am getting ready to start a project that will involve scanning a few thousand photos and negatives (mostly snapshots), so I don't want to make a small mistake now, only to find out later that I really should have done it another way. So any advice at this point would be great.

I will be using a Canon 8800F, but the included software stinks (No TIFF/RAW option). By the time I start this project I will have Adobe Lightroom, but I don't believe it will handle the actual scanning of the photos - only later correction. I have tried a program called VueScan, which seems to be very popular online, but it is missing one crucial feature for me (which I have verified with the developer). It can't perform auto-scew correction or auto-file separation. Even my cheap software can let me put multiple photos in the scanner at the same time and separate them into individual straightened files. I really don't want to have to manually straighten each one individually as this would add lots of work to my workflow.

So my problem is this: What scanner software can auto-straighten and separate multiple photos, while allowing me to scan them as TIFF/RAW. I think it would be best to leave the color correction and sharpenning to Lightroom after the scan. I plan to scan the photo prints at 600 dpi, and the negatives at 2000 dpi. I want to obtain all the reasonably achievable data from these scans so that I never have to worry about whether I did the best I could somewhere down the line.