By Ctein
You should read last week's Part I before reading this one if you haven't already. This picks up directly where that left off.
For scanning negatives, whether black-and-white or color, Adobe RGB is the best choice for output space from the scanner. It can easily capture the full range of such negatives, and it's fairly similar in gamut to what a printer can render. For scanning color slides, though, you will need to use a larger color space such as Wide-Gamut RGB or ProPhoto RGB. If you limit yourself to the Adobe RGB color space, you're likely to see clipping in one or more channels in highly saturated colors. You don't want that! The ideal histogram from a scan contains no data at absolute black or absolute white in any of the channels. You can push tones to pure blacks and whites later when you decide how you want to print them.
Black-and-white film can be especially tricky to scan, because the silver particles that comprise the image scatter light very effectively. To the scanner, with its highly collimated light source, a black-and-white negative appears to be much denser and more contrasty than you'd expect. Scanning black-and-white film so as to capture the full density range of the film is probably the most difficult thing to do. But, having said that, I would note that even a medium-quality film scanner captures a density range from black-and-white film that's equivalent to printing on Grade 1 or even Grade 0 paper. That may not include every bit of detail that's in the negative, but it's more than most people usually printed in the darkroom.
Many people believe that the orange mask layer in color negative films makes them difficult to scan and work with. In fact, just the opposite is true. That orange mask layer may look peculiar to the human eye, but what it's doing is linearizing the characteristic curves for the image-forming dye layers in the film and eliminating unwanted spectral absorptions that cause crosstalk between the color channels. It makes scans (and prints) better, with cleaner tones and purer colors. Restoring visually-normal black-and-white points to a negative scan is a trivial matter in Photoshop; just create a Curves layer and use the black and white eyedropper tools on those portions of the image that you want come out as black and white. You will frequently get a credible rendition of color just by doing that. If not, use the middle gray eyedropper to click on an area that should be neutral, and, nine times out of ten, you'll have a color balance that's good enough to be a decent starting point for doing serious work. Be sure to set the eyedropper radius to 5 pixels or so. You don't want the dropper selecting a single film grain/pixel as the reference point; you want the average of a bunch of adjacent grains.
This brings me to my two final points on scanning:
Your scanning should be done at the highest resolution that you can feasibly work with, not only to capture maximum detail in the film but to minimize film grain and artifacts. Scanning tends to increase the appearance of grain. There are several reasons for this, but the most basic one is that in a digital file a "grain" can never be smaller than one pixel in size. Even in relatively coarse-grained films, the actual grains (or dye clouds in the case of color images) are still much smaller than a low-resolution-scan pixel. The higher the scan resolution you can use, the smaller the minimum-size pixel-grain can be.
1. and 2. Wet scanning markedly improves grain and noise in film scans. The lower enlargement is from a wet scan. The one above it, a normal dry scan. The scans are, in all cases, unretouched and just as they came from the scanner.
Wet (liquid immersion) film scanning will, in many cases, substantially reduce the appearance of grain as well as help suppress dust and scratches (above). It adds preparation time to doing scans, but these days I do 100% of my scans that way. The improvement in image quality is more than worth it. You don't need a special scanner for this; companies like ScanScience (below) sell kits that work with many different models of film scanners.
ScanScience makes nice, easy to use wet-scan adapters
for many film scanners.
The digital file is only the starting point. Really, most of the work lies in how you manipulate the image in Photoshop. This is the custom printing part of the business, and as such is the hardest to explain. I'm afraid there is really no substitute for lots of practice, lots of reading, and possibly attending expert workshops. Do expect to have to climb a substantially long learning curve; it is reasonable for it to take many months to years to become a good printer. It is also reasonable to expect to continue to get better year after year and to find yourself looking back on earlier efforts that you previously thought were wonderful and find it difficult to imagine how you could ever have let the world see results that you now perceive as well below your standards. This is not a bad thing; it just means you're getting better and better.
Next time I'll wrap this up with some tips on image manipulation and for printing the best of all possible prints from your efforts.
Ctein
Ctein's regular weekly column appears on TOP on Thursday mornings.
Send this post to a friend
Note: Links in this post may be to our affiliates; sales through affiliate links may benefit this site. More...
Original contents copyright 2010 by Michael C. Johnston and/or the bylined author. All Rights Reserved.
Ctein - tremendously useful information - tx to you and Mike.
I believe I have asked this before, but now I am _begging_ you ;-)
Will you offer a scanning workshop? I will be the first to enroll.
Posted by: Michael C | Thursday, 14 October 2010 at 11:44 AM
Ctein, another great, helpful post, thanks. As an editing starting point (colour neg), I set black and white points using Enhance Per Channel Contrast, Snap Neutral Midtones, and clipping percentages in a levels adjustment layer. Do you have any thoughts on this technique? I don't know a better, and it works well most of the time.
I'd love a book from you on film to digital.
Posted by: James W. | Thursday, 14 October 2010 at 12:10 PM
Thanks for this series, though my heart sinks a bit at the thought that now I have to go back and re-scan my slides. Or actually re-re-scan. Why do I feel like this might not be the last time?
Posted by: HD | Thursday, 14 October 2010 at 12:15 PM
Another great color space for scanning is Beta RGB by Bruce Lindbloom which is free and "fits" most transparency material...
http://www.brucelindbloom.com/BetaRGB.html
Also while there check out the information about working spaces as it provides a great visual reference...
http://www.brucelindbloom.com/WorkingSpaceInfo.html
Cheers-
Stephen Schaub
Posted by: Stephen Schaub | Thursday, 14 October 2010 at 12:50 PM
Hmmm. Rather a succinct article for wide-ranging subject-matter although I appreciate the understanding of grain(iness) and rounding-up to a pixel-size.
But really, what happened with those two wet/dry scan image-samples? No mention of how many dpi/ppi they were made at, or what format film was used, but sheesh, from those samples I'd just choose the dry one with prior application of shirt-sleeve to remove the dust, because the edge-contrast is *so* superior.
Posted by: Tim | Thursday, 14 October 2010 at 01:06 PM
Thanks for the column. It's fascinating to see what can be accomplished with the proper equipment. Being a very cheap person (in another word, an indigent student), I use a DIY film scanner. I find it adequate for occasional use, although somewhat limited in definition...
Is there any particular risk of negative damage with wet scanning?
Posted by: Damien | Thursday, 14 October 2010 at 01:21 PM
I've been looking at the scanning backlog (and re-scanning; your point about getting better over time certainly conforms with my experience!), and finding the arithmetic totally daunting. Clearly I can't even vaguely consider the level of deep delving I'd like to do, I'm not going to live that long (by current actuarial tables; I do hope for technological improvements still).
Then you point me at people who make a wet-mount system I can use in my Coolscan 5000. That's certainly not going to speed anything up. (And I knew they existed; but hadn't looked them up previously.)
I'm very worried about cleaning up the negatives after that. My old negatives, which I think of as having been reasonably carefully handled, have a lot of scratches along them for example, and I'm getting paranoid about doing anything that touches the negs at all. How finicky is the process of cleaning up after wet scanning?
Posted by: David Dyer-Bennet | Thursday, 14 October 2010 at 01:22 PM
Thanks, Ctein.
So it seems unnecessary to profile or calibrate the scanner, assuming one is dealing with images one at a time. Or am I skipping ahead?
Posted by: robert e | Thursday, 14 October 2010 at 01:31 PM
I would suggest another article describing your wet scanning process with ScanScience products, including lots-o-pictures.
Posted by: Kenneth Voigt | Thursday, 14 October 2010 at 01:56 PM
Hey Ctein - great to see you're still committed to film after your previous post which seemed to indicate that the digital muse had overcome you!
Posted by: Richard | Thursday, 14 October 2010 at 02:04 PM
Does the wet scan affect sharpness? It seems a touch less sharp on the screen but I don't trust my eyes for something like this?
Posted by: Dave Gess | Thursday, 14 October 2010 at 02:10 PM
It hit me the first time, and it has hit me again with the second part of this article.
The word 'perfunctory' seems so at odds with the article. It is the kind of word that I do not think I would want to use to describe my own work. Certainly nothing I would want to admit to.
I am British and in British English I I associate the word 'perfunctory' with - casual, inattentive, going through the motions, not really wanting to give attention to the subject, giving the minimum.
Is the American English connotation of the word different?
Posted by: David Bennett | Thursday, 14 October 2010 at 02:31 PM
Ctein- Not doubting or contesting any of your advice or info- but am I the only one unacquainted with wet scanning who is more than just a tad queasy about ending up with some ruined negatives with this whole immersion procedure (I went to the site and didn't see this addressed)?
Posted by: Stan B. | Thursday, 14 October 2010 at 02:46 PM
Dear HD & DDB,
In many cases I am on my third round of scanning the same film, over a 15 year period. This time everything gets wet scanned, scanned at maximum resolution, into Wide Gamut RGB space, to insure that I won't have to go back and rescan them in the future when my demands and needs change.
Yeah, that's what I thought the last two times I did it, too! [sigh]
Lemme know if you ever get to that "I've really, truly scanned it for the last time" point. We'll throw a party.
~~~~~~~~
Dear Tim and Dave,
Wet scanning does not compromise sharpness (at least up to 4800 ppi, 4000 resolved, which is as good as I can go). If you think the wet scan looks less sharp than the dry, it's a trick of your eyes, an artifact of preparing illustrations for publication (we all know better than to pixel peep an illo), and the illusion caused by grain and noise being on the same scale as the fine detail). As Mike and I have said innumerable times, believe the words, not the illustration.
The ppi I used and the film format are irrelevant. And both negs were cleaned as thoroughly as possible before both scans. Honestly, Tim, do you expect I'd do anything less?
~~~~~~~~
Dear Damien & DDB,
The ScanScience fluid is solvent, not oil based. So it dries pretty residue free. I used Pec Pads and Pec 12 for cleaning the film before wet scanning and for wiping off any dried residue afterwards, because that's the least abrasive combination I know.
I do not believe that there are any long-term deleterious effects from wet scanning. Yeah, there's always a risk of physical damage when you handle film, clean film, use No-Scratch (in the darkroom), etc? So whaddayagonnado? Not print your negs the best way possible because they're too valuable? They ain't that valuable unprinted.
~~~~~~~~
Dear robert e,
Never had good luck color-managing my scanner. Either I don't know how to do it right, or working with a multitude of films and subject lighting conditions just confuses things too much. Either's distinctly possible.
pax / Ctein
==========================================
-- Ctein's Online Gallery http://ctein.com
-- Digital Restorations http://photo-repair.com
==========================================
Posted by: ctein | Thursday, 14 October 2010 at 03:29 PM
Some great, useful info here. Still, I'm a bit baffled by:
"Restoring visually-normal black-and-white points to a negative scan is a trivial matter in Photoshop; just create a Curves layer and use the black and white eyedropper tools on those portions of the image that you want come out as black and white. You will frequently get a credible rendition of color just by doing that. If not, use the middle gray eyedropper to click on an area that should be neutral, and, nine times out of ten, you'll have a color balance that's good enough to be a decent starting point for doing serious work."
I do a lot of landscapes, flowers, etc. Most of them have no areas clearly identifiable at black, white or neutral gray. I do use the dropper technique on some images, but find it far from trivial with many others. Setting white and black points using Levels on a scan where the output doesn't quite reach the ends of the histogram is indeed trivial. Neutral WB is another matter.
I find that creating and using ICC profiles results in scan images that are much closer to correct color and require much less work in post. None of these images but the IT8 target has suitable neutral reference points. Even the "white" hydrant paint isn't really white. All are nicely corrected by the
ICC profile.
One may also apply the profiles later, in post, rather than as part of the scanning process.
I can't create profiles for old films, but I've found that profiles from similar newer films from the same manufacturer often give a better starting point for post than the scanner alone.
Moose
Posted by: Moose | Thursday, 14 October 2010 at 03:32 PM
I can't think of a good way to phrase this Ctein: I can't afford a good scanner, so should I proceed directly to paying for drum scans? Or is there an intermediate stop along the way for hiring scanning done?
Will
Posted by: Will Frostmill | Thursday, 14 October 2010 at 03:33 PM
David Bennett my almost-name-mate -- "perfunctory" certainly can include casual, shallow, brief.
It doesn't seem unreasonable to me to apply it to these short articles here; particularly in the context of Ctein as somebody who writes entire books on photographic topics. These articles aren't a detailed treatment of a small topic, they're a rather quick overview of a huge topic.
(I think of myself as American; but in fact my father was born in England, so I may have a bit of a bias towards English usages in some contexts.)
Posted by: David Dyer-Bennet | Thursday, 14 October 2010 at 03:37 PM
Dear Richard,
Nope, you're dead wrong. I am completely UN-comitted to film. It is a dead end, a dead horse, so last century. I no longer own a film camera and will never buy another one. I have given away a freezer full of film and darkroom paper to educational institutions that will make use of it.
What I do have are hundreds of portfolio-worthy photographs I've never even gotten around to printing yet (not to mention all the stuff I haven't printed digitally, that I could print BETTER that way). So what should I do? Just discard all those great photos?
Frankly, it's a pain in the ass. Between cleaning, prepping, scanning and spotting, it takes me at least an hour to digitize a single frame of film. If it's an especially cranky one, it might go to two hours. That's before I even start on that whole "fine printing" business-- that just gets me a good digital "negative."
But, like I said, not much choice, really.
Well, at least it's a heck of a lot faster than making a dye transfer print
~~~~~~~~
Dear David,
No, it's not much different a word in American-- "Adjective: (of an action or gesture) Carried out with a minimum of effort or reflection."
It's a bit of self-mockery, for my own amusement.
It also has considerable truth. Any explication of the very complicated craft of producing a digital print from film that only runs 3,000 words will, in truth, be perfunctory, no matter what the author's aspirations.
The pertinent Chapter 4 in my book, DIGITAL RESTORATION, is not perfunctory-- it runs 10,000 words and has 38 illustrations. And that's just about scanning, alone.
Posted by: ctein | Thursday, 14 October 2010 at 03:41 PM
Will - for what it is worth, and Ctein might have another opinion - but Ken Rockwell swears by these folks:
http://www.northcoastphoto.com/
As for color correcting the negatives, why not take a photo of a color chart or Whibal? You can simply invert it and have any number of whites, greys or blacks to set the balance. Try it on a few rolls to see if it is consistent and you might not need to do it all the time.
Posted by: KeithB | Thursday, 14 October 2010 at 04:18 PM
Hiya, Moose!
As I'm sure you suspect, relatively few of my photos have a useful midtone point in them, just like yours. But my experience with other photographers is that most of their photos do, so it's a useful trick for them.
I don't worry about fine-tuning things *too* much in the scan. The software isn't all that sophisticated, and 16-bits-per-channel hides a multitude of sins [g].
Hey, got some profiling questions for you! I've run into two (related) problems. The first is that when I profile in the standard approved fashion, the scanner is happy to produce for me results that look a lot what's on your web site, with true whites and true blacks... which I don't want! Invariably there is some clipping in the extremes. I don't want that my scans; I'll decide what the final black and white points are myself, thank you very much Mr. Scanner. I've tried messing with profiler settings to do something like set the black point to 25 and the white point to 235, but it never worked right. Dunno if that's a basically wrong approach, or just my ineptness.
A separate problem is that 99% of my work is negative. In uncontrolled lighting situations, I end up capturing eposure ranges in the negs that are far outside what the IT8 target assumes. Mondo clipping ensues when I scan such negs using the IT8 profile. It's useless.
Consequently, as I said to robert, I haven't been able to get consistently good results out of color-managed scans.
Any thoughts? Helpful suggestions?
pax / Ctein
==========================================
-- Ctein's Online Gallery http://ctein.com
-- Digital Restorations http://photo-repair.com
==========================================
Posted by: ctein | Thursday, 14 October 2010 at 05:17 PM
Dear Will & KeithB,
I'm afraid I've yet to find a decent scanning service that cost less than dollars per frame, not dollars per roll. (Hey, DDB-- didn't you point me at a potential candidate several years back? Got a pointer to that?) I know some pretty good ones that charge only (ahem) $10-$20/frame; they're not quite as good as me, but they're close.
The big problem for me with what these folks are offering is that they're JPEGs, which means 8 bits per channel. One *can* do really good printing from 8-bit scans; I did it for years. It's harder work and takes more skill and I really can't recommend it any longer.
They also don't mention what color space they're scanning into. If it's sRGB, just forget it. Even AdobeRGB's gonna produce some clipping when scanning slides. Problem is that with only 8 bits per channel, you can't use really large color spaces like ProPhoto or Wide Gamut RGB-- you'll get contouring.
The resolution on the medium format scans is definitely inadequate for serious work.
If Ken likes these folks, they're probably OK for routine stuff, but it's below my acceptance threshold.
pax / Ctein
Posted by: ctein | Thursday, 14 October 2010 at 05:30 PM
Thanks for this, Ctein.
Just this past week I jumped on getting a mint, lightly used Nikon 9000ED scanner. I already own a 5000ED but the 9000 is a multi-format unit with a stellar rep and is no longer in production. (See a hint in that last part?)
I also recently picked up a lovely Fujifilm 670 folding medium-format (6x6 & 6x7) gizmo. Great fun.
So I will continue to shoot and scan a little film just for the fun of it (and the fact that I still have several terrific film bodies)
But it really is such an utter waste of time. Really. Scanning, re-scanning, re-re-scanning ultimately to get an image that's measurably poorer than I can capture digitally.
That aside, Ctein I appreciate the link to the wet stuff supply site. I wonder if, before you get to printing, you could show your process for wet scanning? I've not yet tried it myself but am now tempted. (Hey, what's the point in getting a little pregnant?)
Thanks!
Posted by: Ken Tanaka | Thursday, 14 October 2010 at 05:47 PM
I contacted ScanScience and they had some great options for my Minolta 5400 II. They also said to tell you hello!
Posted by: Keith I. | Thursday, 14 October 2010 at 07:37 PM
I just thought I would point out that www.betterscanning.com has excellent wet and dry scanning solutions targeted mainly to Epson flatbed scanners, though other flat bed scanners are supported.
I'm not affiliated with betterscanning.com just a happy customer.
Posted by: Don Bryant | Thursday, 14 October 2010 at 08:46 PM
Ctein, I think I just had a revelation regarding wet scanning. In restrospect the explanation seems obvious but it's one of those things that's ONLY obvious AFTER the event. As are most really useful snippets.
Looking forward to your follow up photoshop article but in the meantime this is the most useful factiod I have accumulated in, oh, a long time. Thanks!!!
Posted by: Steve Jacob | Thursday, 14 October 2010 at 09:07 PM
Will,
I've had pretty good results with an Epson 4870 I got off ebay and a Better Scanning dry/wet mount holder. I have not wet mounted as of yet. I'm sure for any number of reasons this is a pretty mediocre setup but it's what I've got and I have developed some tricks to make it work for me. The scanner is useless without the betterscanning.com holder, and pretty useful with it. I think I have $225 into the whole rig.
It's less good for 35mm, more good for 4x5 and 5x7. But I've made good 35mm scans and some pretty good 8x10 prints from those scans. It has digital ICE, which was absurdly helpful when I scanned in a couple thousand family slides. Really did an amazing job on dust.
I can say that my skills as a BW photographer and film-developr are not eclipsing the power of this scanner yet.
You have to make do with what you've got. Or can get.
If you can't afford a film scanner, you probably can't afford to have someone do drum scans for you. It's expensive.
In retrospect, I'd probably try to find an Epson 4990 used rather than the 4870, but this setup is working fine for me.
All the recent black and white work I've done on my flickr page is done with this scanner.
Good luck!
Paul
Posted by: Paul | Thursday, 14 October 2010 at 09:52 PM
Dear Ken,
For medium format scanning, I follow ScanScience's instructions to the letter (that manual is downloadable from their website).
35mm is tricky, sometimes. If the film has developed too much of a curl, the capillary attraction from the fluid between the flexible overlay and the glass plate isn't enough to keep the film from springing the sandwich apart and letting air bubbles in.
I worked out the following procedure for scanning Jim Marshall's chromes on my Minolta Dimage Multi Pro by combining the ScanScience overlays and mounting fluid with the glass carrier for the scanner. It may have some helpful tricks. (No, I can't tell anyone exactly how to adapt this method to a different make or model of scanner.)
Anyway, here's the notes I wrote to myself:
INSTRUCTIONS FOR SCANNING JIM'S KODACHROMES
Preparation
1) Use glass carrier. Tape edge of slide in place with the emulsion down.
2) Spray emulsion and carrier glass with Lumina; lower slide onto glass.
3) Spray base of slide and overlay sheet with Lumina; lower sheet on to slide.
4) Place multiformat-35 mask on top of overlay sheet and close carrier.
5) Mask sides to exclude white light.
Scanning
1) Use the following scan settings: multi-format 35mm, glassless, slide film, 4800 ppi.
2) Use the following options settings: Auto expose for slides - DON'T AUTOEXPOSE DUPE SLIDES
16 bit color depth (NOT 16 bit linear!)
multi-sample scan = 8X
output color space = Wide Gamut RGB (NOT Adobe RGB!)
3) Scan with Digital ICE^3 off. (Scan with DI^3 ON when scanning Ektachromes.)
4) Open in Photoshop: assign Wide Gamut RGB when asked about profile.
pax / Ctein
Posted by: ctein | Thursday, 14 October 2010 at 11:41 PM
There's some perceptive discussion (e.g., here ) on Colin Jago's excellent - now suspended but so far still available) blog - about the idea of saving a scanner-Raw file, and converting this as a separate stage. This seems to offer some of the same longevity as camera-Raw does: as our software (and, perhaps, understanding and experience) improve over time, we can re-convert for better results, without necessarily needing to "re-shoot" every time.
Posted by: richardplondon | Friday, 15 October 2010 at 05:11 AM
Got a low end scanner (Epson V700) and the next upgrade is your kit. Any more details of your setup e.g. what you bought ... May I ask how and what adjustment of negative height?
More importantly, may I suggest to have a paid video. I guess if you can walk us through one of the slide, b/w and color negative. Within reasonable price, it would be a tremdous help. The simplest way is to a Podcast - with a mice plus a screen catch (on a HD level video or better the same resolution as the Mac screen), it could be very effective. I do not know but I subscribe to paid podcast before for training. It works. Very unlike the LL training video (which I found video the head is not nice and I cannot follow it easily; so far the best one I can follow is -- well the video on your darkroom on doing your special print). Just screen video capture plus a mice, plus not too high a price point.
Of course, something about taking wet scan need a bit of head shoot. But for those, a simple video setting might be better. Seeing is believing.
The alternative is to have workshop but flying from Hong Kong to US take a bit time and hard to be arranged with so many other commitments like family.
Posted by: Dennis Ng | Friday, 15 October 2010 at 06:36 AM
Just wanted to add my thanks for this article, it's a very good reference. If I ever wanted to print really big, I think the cheapest way for me to do it is to buy a 2nd hand medium film camera and scan the negs. It would cost me more in time, of course, but the other alternatives, "full-frame" digital or even medium-format digital, will be out of my price range for years to come. I may never do it, but if I do, I'll know where to look for advice.
Posted by: Robert Roaldi | Friday, 15 October 2010 at 08:01 AM
Scanning vs digital capture is in most cases a flawed discussion as the variables involved in scanning, the equipment used and the experience of the operator make a huge difference. I have seen amazing scans (I do this for a living) and I have seen great images made from digital cameras of all makes (I do this as well for a living so no bias here) but just for further reading/ information check out this link / article:
http://www.imx.nl/photo/Film/page169/page169.html
Cheers-
Stephen Schaub
Posted by: Stephen Schaub | Friday, 15 October 2010 at 09:00 AM
Dear Richard,
While there's nothing wrong with the ColorNeg approach to conversion, and his plugin works well, the notion of "RAW" scanner files is a complete non-starter, in my judgment.
The inadequacies in scans come from the data that's collected in the first place, not how it gets converted. The resolution you get out is a hardware function-- tell the scanner you want a 1200 ppi scan and it's only going to sample the image every 1/1200 of a inch. You can't magically go back later and tell it to give you that maximum 4800 ppi-- not unless you scanned it that way to begin with. The RAW contains no magic. You want universality? TIFF is as universal a format as you can get (ummm, ignoring that there are about 271 different flavors of TIFF, so don't get too cute-- be as generic as possible). You need 16 bits? Well, that's what you scan at. Noise and grain too harsh? Sorry, that's also an inherent property of the physical scan. You can do things to mitigate those during the scan-- wet scanning, higher scan resolutions. But if you don't, you're stuck.
What many scan software packages call "linear" output *is* as close to unmassaged data as the scanner can produce. and, hence, as close to RAW as you can possibly get. You still get a TIFF file.
A scan is not like a digital photograph. The things that make RAW sensible for digital photography simply don't apply to scans.
I think it's a dumb idea and a waste of people's discussion time. Better that they should spend that energy learning to make good scans.
pax / Ctein
==========================================
-- Ctein's Online Gallery http://ctein.com
-- Digital Restorations http://photo-repair.com
==========================================
Posted by: ctein | Friday, 15 October 2010 at 12:14 PM
Dear Stephen,
Sorry to have to say this, 'cause I really admire Erwin and what he does, overall, but that's a put-up job and not at all useful unless photographing with ultra-high-resolution repro films is your particular thing *AND* you demand that your digital photographs come out looking like film.
If that's what you're into, it's a great comparison. For the other 99% of us, it's useless and pointless.
I've written a lot of columns talking about the ways in which digital is clearly superior to film and vice versa. I could pick any one of those columns and construct an experiment that focussed on the particular image characteristic and construction I was talking about and, voila! I would thereby have demonstrated how superior film was. Or digital was. For some extraordinarily narrow definition of "superior."
As you said, it's an almost intractably flawed discussion, and it's why I choose not to go there.
And related to that...
~~~~~~~~~~~
Dear Robert,
It's not obvious you're better off that way. A decent used medium format camera and a good medium format scanner are going to cost you in the same neighborhood as a Sony a900. I've worked with files from that camera, printing up to 24"x30". Overall (and painting with a very broad brush), it produces medium-format film quality. But, take into account my remarks to Stephen when swallowing that statement. The Sony produces digital photos. Your unnamed camera produces film ones. They will not look the same. And for various narrow measures of quality, each will be able to demonstrate superiority.
What you need to do is figure out what it is you want to do and what kind of photographs you want to make. The money will be a wash.
pax / Ctein
==========================================
-- Ctein's Online Gallery http://ctein.com
-- Digital Restorations http://photo-repair.com
=========================================
Posted by: ctein | Friday, 15 October 2010 at 12:30 PM
For a really big print, for many subjects, the cheapest way is digital mosaic -- if you shoot a 9 x 3 matrix of images at 12MP each, you've got a LOT of pixels to work from! Of course this is completely unfeasible for some subjects, and quite hard for others. But for quite a few, it's a great choice; it's a technique worth keeping in mind.
Ctein, the place I sent stuff for scanning was for fairly high-volume scanning of snapshots for my records. Although they scan by hand using Nikon 5000 and 9000 scanners, their focus is on volume and price, not exhibition quality. So they probably don't apply to this discussion. (The place was scancafe.com; they did what they advertised to do quite competently.) (They do advertise some higher-end services, up through medium format scans done wet-mounted as I recall; I didn't use the high-end services, perhaps they are in fact of exhibition quality.)
Posted by: David Dyer-Bennet | Friday, 15 October 2010 at 12:56 PM
HI Ctein - in your first article you mention that
"Scanning film, especially negative film, is very straightforward in principle. "
Can you elaborate perhaps on the distinction maybe?
Posted by: Richard | Friday, 15 October 2010 at 03:52 PM
I know some pretty good ones that charge only (ahem) $10-$20/frame;
Ouch.
Okay, I guess I'm getting off the train at this stop. No more film for me without a really good reason.* Thanks Ctein, now I understand why you don't do film: it is really, really, expensive in time and money.
Robert Roaldi, I did just what you suggested: I bought a used MF camera, and scanned the negatives. But a back of the envelope calculation gives me 135-270 scans for $2700, the price of the Nikon 9000 Ken Tanaka uses. $2700 can buy an awful lot of camera - even full frame, if you want an A-850. (Which isn't a bad point of comparison, since it's best iso range is reported to be between 100 and 400, about the same as you can do with fine grained emulsions.) Trading an hour (or 20$) for zero time to acquire the image file, and a limitless supply of film seems like a very, very good buy.
I think I'm going to flip through my negatives file and do a little math.
Will
*when in a hole, stop digging.
Posted by: Will Frostmill | Friday, 15 October 2010 at 04:14 PM
Dear Will,
Hey, remember I do flippin' dye transfer printing. The "time and money" associated with photographing on film is truly insignificant.
I just ain't interested in making film photographs any more.
pax / Ctein
Posted by: ctein | Friday, 15 October 2010 at 06:15 PM
Ctein- first you tell us that even manual cameras (no matter the age, cost or design) rarely, if ever, focus accurately and now, that we've never made a scan worth a rat's...
Why must enlightenment be so damn discouraging?
Posted by: Stan B. | Friday, 15 October 2010 at 06:22 PM
Thank you for the wet scan process outline, Ctein.
p.s. Hey, as I noted I got a great deal on the Nikon 9000ED; only $1,700 for a like-new unit. (And from the Earth-is-small department, the fellow from whom I bought it turns out to live a block from where I grew up!) Yes, that still buys a very nice digital camera. But film is not logical. Live rich and die happy.
Posted by: Ken Tanaka | Friday, 15 October 2010 at 10:54 PM
Ctein,
Yes, I do remember that. Oi! You are a patient, patient, man. I have a huge amount of respect for that.
Will
Posted by: Will Frostmill | Saturday, 16 October 2010 at 09:19 AM
Just have to say that yesterday I showed a (big name-super high end printer to some really famous people) some 8x10 digital prints I did with my crappy dying printer, scanned with my crappy 4870 and he was suitably impressed and complimentary. Not "omg that's great" but complimentary. Again, I think it's the photographer in the situation. You can learn to make do with what you have.
Posted by: Paul | Saturday, 16 October 2010 at 10:04 AM
Great series Ctein, thanx.
I have one or two questions, if I may, about scanning before embarking on scanning all my color negative films again this winter on the Minolta 5400 II.
- how does the Expose To The Right practice of digital capture translate to scanning color negatives? Around half of the imagefile data of a digicam are in the brightest zone. Can we translate this to scanning color negatives somehow and what would be best practice? Do you leave some room on the left and/or right of the histogram at the scanning stage?
- is it best to invert the negative image in the scanning software or Photoshop?;
- what it comes down to is... what would be best practice at the scanning stage in order to preserve highlight and shadow detail (with room to spare in post). Is there some set of rules after the scan preview to adjust settings for the actual scan regarding image adjustments or can a straight 'leave as it is scan' work well?
Jan
Posted by: Jan Koning | Saturday, 16 October 2010 at 10:13 AM
Ctein and Will,
Thanks. I never anticipated needing more than 1 or 2 large prints per year, and even that is probably an overestimate. But I've seen local classifieds for Mamiya 330s or Yashica 124s for less than $400 (including lens) and that's what I had been basing my wild guess on. I can see what you're saying though. Maybe renting the "full-frame" digicam for a weekend might make sense, considering how frequently I would do this. Assuming I ever need to do it.
Posted by: Robert Roaldi | Saturday, 16 October 2010 at 03:59 PM
Dear Jan,
No arcane scanning rules. Just set the level controls so that your blackest blacks appear to be 5-10% above 0% and the whitest whites 5-10% below 100%, for all three color channels.
That's to prevent clipping in case you missed a few extreme highlights/shadows.
Doesn't matter where you invert. Some scanner modes, like 16-bit linear in my Minolta, won't even let you invert before exporting the data.
pax / Ctein
Posted by: ctein | Saturday, 16 October 2010 at 06:58 PM
"I'm afraid there is really no substitute for lots of practice, lots of reading, and possibly attending expert workshops."
Any recommendation on reading material ? or workshops for that matter (Though where I am, workshops would probably be very far away...) ?
Basically, how does one gain experience with no reference point that he can hold in his hands ?
Posted by: Erez | Tuesday, 19 October 2010 at 10:40 PM
Dear Erez,
Ahh, but you can hold them in your hands. Books!
Both O'Reilly & Associates (http://oreilly.com/) and Rocky Nook (http://www.rockynook.com/books/all/index.html) are good places to look. I know that they both have several books on fine digital printing. I haven't read said books, but I've rarely seen a book from either that wasn't worth the price. They publish quality.
Other reads may have specific book recommendations for you, but if you're really at the "so early in the learning process I have nothing to grab onto" phase, grabbing any book on printing by either of those publishers will likely teach you a whole bunch of useful stuff you didn't know.
pax / Ctein
Posted by: ctein | Wednesday, 20 October 2010 at 01:25 PM
Hiya yourself!
“As I'm sure you suspect, relatively few of my photos have a useful midtone point in them, just like yours. But my experience with other photographers is that most of their photos do, so it's a useful trick for them.”
Just thought I’d clarify that for those who don’t have them. I can just see someone thinking "The master says to do it this way.", while being bemused at the wild colors that are turning up.
“I don't worry about fine-tuning things *too* much in the scan. The software isn't all that sophisticated, and 16-bits-per-channel hides a multitude of sins [g].”
Indeed it does, at least if the color space has a wide enough gamut.
“Hey, got some profiling questions for you! I've run into two (related) problems. The first is that when I profile in the standard approved fashion, the scanner is happy to produce for me results that look a lot like what's on your web site, with true whites and true blacks... which I don't want! Invariably there is some clipping in the extremes. I don't want that my scans; I'll decide what the final black and white points are myself, thank you very much Mr. Scanner. I've tried messing with profiler settings to do something like set the black point to 25 and the white point to 235, but it never worked right. Dunno if that's a basically wrong approach, or just my ineptness.”
I only come close to understanding VueScan. The CanonScan that came with the scanner hasn’t run on at least the last three versions of Windoze.
I’m generally not worried about empty space at the top and bottom of the histogram of the scan. I like the highlights to just ‘kiss’ the tops and with most images don’t mind losing a little on the bottom. The bottom couple of bits only represent 4 values anyway. But for you …
Here’s the result of using VueScan to give top and bottom room in the histogram. I’m not sure what the little thing at the bottom of the first histogram may be, but I don’t think it’s real data. It disappears in the first processing step.
You can see how I control the histogram in post, never blowing anything top or bottom. It may appear at first glance that I clip the bottom, but I have only compressed it, piling up the values. Artists choice.
The last version of the image is using no ICC profile and less effort to leave space in the histogram. The color cast is pretty easy to see; that iris is yellow, not greenish. The final choice shows the VueScan Color Tab settings for both.
“A separate problem is that 99% of my work is negative. In uncontrolled lighting situations, I end up capturing eposure ranges in the negs that are far outside what the IT8 target assumes. Mondo clipping ensues when I scan such negs using the IT8 profile. It's useless.”
The above example is a scan of a Kodak Portra 160NC negative, shot in direct, middle of the day sun. The ICC profile made by shooting an IT8 target in the same light, although at a different date and time, has worked fine.
“Consequently, as I said to robert, I haven't been able to get consistently good results out of color-managed scans.
Any thoughts? Helpful suggestions?”
I hope the above helps. It’s certainly possible to get reliable results scanning color negative film on a Canon FS4000 scanner using VueScan. I do it quite a bit as I slowly scan my old film. I've also found that using a profile made with more recent versions of a film often works quite well with older version of the same film. A big help when scanning older film that I can’t profile.
As to scanning film with subject colors that are outside the IT8 printed range, I’ve had little trouble. Sunsets direct light from various kinds of artificial light and so on seem to come out pretty well. That may be because most of them are outside the range of our color memory and papers can’t reproduce them.
Moose
Posted by: Moose | Sunday, 24 October 2010 at 02:10 AM
This is a tremendous discussion, and I noticed the wet mount part of it in particular.
At ScanCafe we have found that wet-mounted scanning is extremely helpful, too. We do it--as we do everything--manually. The twist however is our proprietary wet mounting process that we're patenting (our head of Operations, who wrote his doctoral thesis at U of Minn on image processing, holds several patents already, and quickly saw the opportunity here).
I would encourage anyone thinking about scanning B&W 35mm negatives themselves to look hard at our 69 cent/scan price (for wet mount!!) and try us out. Adding a few cents for a TIFF file could get you right to the door of exhibition quality for an almost unthinkably low cost, and an incredibly small time commitment on your part. Even if you look at the example above, there are still plenty of scratches left to work on in a B&W wet-mounted scan, and that's something you may just be better off letting us do.
More about our wet mount scanning here (it was a product of the year for Outdoor Photographer last, year btw): http://www.scancafe.com/services/black-white-scanning
Posted by: Wade Lagrone | Thursday, 11 November 2010 at 03:00 PM