Bryan Lee
Expat Street Photographer
The way I visualize the process lately is thinking about it this way. The scanner is just a big digital camera. Aperture, Exposure, Flash, and every other variable is ajustable but the object in the image will remain unchanged. Making it simple has helped me relax and try to undertand the technical aspects better. Im currently at the level of a kidergarten student with the big crayons, Just graduated from finger paints.
back alley
IMAGES
i know that kindergarten feeling too well.
but i also like to keep things simple and so far it seems to be working.
joe
but i also like to keep things simple and so far it seems to be working.
joe
schmoozit
Schmoozit good...
Joe, perhaps it's just a way of saying, "This photo is so technically poor that I don't know why you would think of posting it in public. Hmm... must'a been a bad scan!"
I hope not... I have a lot of "bad scans"!
I hope not... I have a lot of "bad scans"!
aad
Not so new now.
I had odd scans until I figured out the whole destination size/screen appearance relationship-otherwise my first attempts wouldn't fit in the post limits. They all looked like "South Park" Canadian photos.
ChrisN
Striving
I've just got inspired reading the above posts, and spent a couple of hours scanning a single negative with many different settings, on two different scanners. I thought I was making progress, then realised that the same image looks totally different when viewed in 1. Windows Picture and Fax Viewer; 2. Internet Explorer; and 3. Photoshop Elements. It will probably look different again when printed, on my cheap printer, then on a shop printer. How do you go about determining a benchmark?
schmoozit
Schmoozit good...
Chris: Google monitor/ printer/ scanner calibration, and hold on to your hat!
PaulN
Monkey
ChrisN said:
I think that may have to do with the colorspace, ie sRGB, Adobe98, etc. Some browsers ignore the color space, while others don't. That is why an image may look good on FireFox but not on IE, etc. Here's a highlevel overview of browsers & sRGB/Adobe98 via Smugmug:
http://www.smugmug.com/help/srgb-versus-adobe-rgb-1998
aad
Not so new now.
ChrisN, I so far have had my negs printed, or I get slides. I benchmark against the print/slide. So far, PS Elements, Preview, iPhoto and Safari look the same. B&W needs no tweak, slides often need a little saturation boost, and color negs need the most work.
I also notice new Ektachrome (within 10 years) scans much better than older versions.
I also notice new Ektachrome (within 10 years) scans much better than older versions.
jlw
Rangefinder camera pedant
Mister Moderator: I just had a flick through your gallery, and I don't think there's anything at all wrong with your scans except that occasionally the highlight areas lack detail relative to the shadows -- i.e., either the highlights have detail and the shadows are inky black, or the midtones and shadows have detail and the highlights are dead white. (Of course, maybe this is the effect you're trying to create.)
If you're scanning directly from negatives rather than from prints -- and if you're using a Canon 2710, I assume you are -- then this can be difficult, if you're using conventional silver-based b&w films. The highlight areas of these films are much denser than those of chromogenic (C-41 and slide) films, and the scanner's default settings may not punch enough light through those dense highlights to record a good level of detail. If the highlight detail isn't in the original scan, then nothing you do afterward in Photoshop is going to retrieve it -- you'll wind up having to tweak the gamma (midtone contrast) to get a reasonably normal-looking distribution of highlight tones, and that will pitch the darker midtones and shadows down into near-blackness.
In order to get around this, you have to override the scanner's automatic exposure setting and manually choose a setting that will produce detailed highlights.
I'm not sure what controls are in the scanner plug-in that Canon supplied for the 2710, but on the one for my FS-4000, you can get at this setting under the Settings>Exposure Settings menu. Choosing this brings up a little dialog box with an "Auto Exposure" checkbox and a five-step slider, from -2 to 0 to +2 (+2 what? They don't say.) What you have to do is uncheck the "Auto Exposure" box, then choose a slider value. The + settings give more exposure, like leaving the enlarger lamp on longer, so you'll get darker overall values, including darker highlights. The - settings give less exposure, yielding the opposite results.
(You can do the same kinds of things in VueScan, but the settings have completely different names and are scattered around on various panes of the interface, some of them not visible until you change the Options pop-up to Standard or Advanced. So be prepared to do some experimenting.)
Whichever software you're using, though, the principle is the same: set the scanner controls to yield scans with the highlight and shadow detail you need, and then use your Photoshop adjustments to get the distribution of tones you want in the image. This is an oversimplification, but to put it in darkroom terms, your scanner exposure settings are like getting a printing time that yields good highlight exposure; once you've done that, the Photoshop part is like choosing a variable-contrast filter that puts the rest of the tones where you want them.
Okay, now comes the kicker: Most scanners don't have enough density range to capture the full range of tones in a conventional b&w negative. So, your exposure setting that yields good highlight detail probably will leave the dark midtones and shadows too black to recover with Photoshop. You can lighten them up, but they'll lose the correct linearity and look murky and smokey.
So what do you do? Well, here's what I do, although I'll admit it makes the process slower and more complicated, and lots of people have told me they can get perfect results without going to all this bother, but it's what works for me:
-- I scan the negative at least twice: Once with a scanner exposure setting that yields good highlight separation, letting the dark tones fall where they may; and again with a scanner exposure setting that produces well-separated dark tones, even though this usually leaves the highlights blown out.
-- Then I put the scans together as separate layers in a single Photoshop document. (To get them to line up perfectly, hold down the Shift key as you drag one onto the other.) I like to put the light scan (the one with good shadow detail) on the bottom and the dark one (the one with good highlight detail) on the top.
-- I set the top (dark) layer to Multiply mode, so it combines with the one underneath it rather than just replacing it. Then I adjust the opacity of the top layer (the one with highlight detail) until it multiplies just enough of its highlights onto the bottom layer (the one with shadow detail) to give a full range of tones. If the whole thing looks too dark, I add a Levels adjustment layer over the bottom (shadow-detail) layer, and tweak its settings until I have the desired balance of highlights, midtones, and shadows.
This may sound confusing, but it's really only a three-layer "sandwich" -- the dark scan that has good highlight detail on the top (with its mode set to Multiply); the Levels adjustment layer in the middle; and the light scan that has good shadow detail on the bottom. Adjusting the opacity of the top layer and the level control on the adjustment layer lets you balance the tones, and everything's completely undoable and adjustable -- if you change something and don't like the result, just change it back.
This usually gets the job done for average negatives -- it's the darkroom equivalent of a "straight print." If you've got a negative that requires the digital equivalent of burning-in, you can add a second copy of the dark scan on top of the first. Again, you set its mode to Multiply, but this time you add a layer mask that hides its effect. (Layers>Add Layer Mask>Hide All) As you paint on the layer mask, it'll let some of the "burn-in layer" become visible and add more detail to the highlights beneath. What's cool about this vs. traditional darkroom technique is that if you screw up, you can "un-burn" the highlight by inverting your paint color (just type X) and painting out the mask -- again, all reversible, undoable, and re-doable.
Once you understand the concept of layer masks, you also can add one to the bottom (shadow detail) layer if necessary, to hold back highlights. This is all very advanced stuff, though, and you should seldom need this whole bag of tricks if your negatives are reasonably well-exposed and you start with correct scanner exposure settings for your highlight and shadow scans; in other words, usually the three-layer sandwich is the most you'll need.
I didn't really intend for this to turn into such an extensive spew on my unnecessarily complicated scanning techniques -- it's just that earlier this evening I scanned another old negative to add to my gallery collection, and this one was a real bear! Because my original film exposure had been so poor, I wound up needing THREE scans (for highlights, midtones, and shadows) and two layer masks (on the highlight and midtone layers, so I could balance the tones in various areas of the scene the way I wanted them.) I'm attaching it so you can see what I mean, and also because I want to get a little extra mileage out of the scan after all that work!
As I said, other people probably have techniques that work just as well without nearly as much bother, and all this stuff shouldn't be necessary when scanning chromogenic negatives. But as I said, it's what works for me with the wacky negatives I produce. (I remember describing this technique on another forum and noting that it's like pulling teeth, and another member who happened to be a dentist posted back, "No, actually, pulling teeth is a lot easier than that!")
If you're scanning directly from negatives rather than from prints -- and if you're using a Canon 2710, I assume you are -- then this can be difficult, if you're using conventional silver-based b&w films. The highlight areas of these films are much denser than those of chromogenic (C-41 and slide) films, and the scanner's default settings may not punch enough light through those dense highlights to record a good level of detail. If the highlight detail isn't in the original scan, then nothing you do afterward in Photoshop is going to retrieve it -- you'll wind up having to tweak the gamma (midtone contrast) to get a reasonably normal-looking distribution of highlight tones, and that will pitch the darker midtones and shadows down into near-blackness.
In order to get around this, you have to override the scanner's automatic exposure setting and manually choose a setting that will produce detailed highlights.
I'm not sure what controls are in the scanner plug-in that Canon supplied for the 2710, but on the one for my FS-4000, you can get at this setting under the Settings>Exposure Settings menu. Choosing this brings up a little dialog box with an "Auto Exposure" checkbox and a five-step slider, from -2 to 0 to +2 (+2 what? They don't say.) What you have to do is uncheck the "Auto Exposure" box, then choose a slider value. The + settings give more exposure, like leaving the enlarger lamp on longer, so you'll get darker overall values, including darker highlights. The - settings give less exposure, yielding the opposite results.
(You can do the same kinds of things in VueScan, but the settings have completely different names and are scattered around on various panes of the interface, some of them not visible until you change the Options pop-up to Standard or Advanced. So be prepared to do some experimenting.)
Whichever software you're using, though, the principle is the same: set the scanner controls to yield scans with the highlight and shadow detail you need, and then use your Photoshop adjustments to get the distribution of tones you want in the image. This is an oversimplification, but to put it in darkroom terms, your scanner exposure settings are like getting a printing time that yields good highlight exposure; once you've done that, the Photoshop part is like choosing a variable-contrast filter that puts the rest of the tones where you want them.
Okay, now comes the kicker: Most scanners don't have enough density range to capture the full range of tones in a conventional b&w negative. So, your exposure setting that yields good highlight detail probably will leave the dark midtones and shadows too black to recover with Photoshop. You can lighten them up, but they'll lose the correct linearity and look murky and smokey.
So what do you do? Well, here's what I do, although I'll admit it makes the process slower and more complicated, and lots of people have told me they can get perfect results without going to all this bother, but it's what works for me:
-- I scan the negative at least twice: Once with a scanner exposure setting that yields good highlight separation, letting the dark tones fall where they may; and again with a scanner exposure setting that produces well-separated dark tones, even though this usually leaves the highlights blown out.
-- Then I put the scans together as separate layers in a single Photoshop document. (To get them to line up perfectly, hold down the Shift key as you drag one onto the other.) I like to put the light scan (the one with good shadow detail) on the bottom and the dark one (the one with good highlight detail) on the top.
-- I set the top (dark) layer to Multiply mode, so it combines with the one underneath it rather than just replacing it. Then I adjust the opacity of the top layer (the one with highlight detail) until it multiplies just enough of its highlights onto the bottom layer (the one with shadow detail) to give a full range of tones. If the whole thing looks too dark, I add a Levels adjustment layer over the bottom (shadow-detail) layer, and tweak its settings until I have the desired balance of highlights, midtones, and shadows.
This may sound confusing, but it's really only a three-layer "sandwich" -- the dark scan that has good highlight detail on the top (with its mode set to Multiply); the Levels adjustment layer in the middle; and the light scan that has good shadow detail on the bottom. Adjusting the opacity of the top layer and the level control on the adjustment layer lets you balance the tones, and everything's completely undoable and adjustable -- if you change something and don't like the result, just change it back.
This usually gets the job done for average negatives -- it's the darkroom equivalent of a "straight print." If you've got a negative that requires the digital equivalent of burning-in, you can add a second copy of the dark scan on top of the first. Again, you set its mode to Multiply, but this time you add a layer mask that hides its effect. (Layers>Add Layer Mask>Hide All) As you paint on the layer mask, it'll let some of the "burn-in layer" become visible and add more detail to the highlights beneath. What's cool about this vs. traditional darkroom technique is that if you screw up, you can "un-burn" the highlight by inverting your paint color (just type X) and painting out the mask -- again, all reversible, undoable, and re-doable.
Important fine point: The reason that this is better than using the little burn-in tool in the Photoshop toolbox is that it actually lets you add detail to highlights, just the way darkroom burning-in can do. The Photoshop burn-in tool only darkens the pixels that already are there; it can't bring in any additional tonal information that's not already present.
Once you understand the concept of layer masks, you also can add one to the bottom (shadow detail) layer if necessary, to hold back highlights. This is all very advanced stuff, though, and you should seldom need this whole bag of tricks if your negatives are reasonably well-exposed and you start with correct scanner exposure settings for your highlight and shadow scans; in other words, usually the three-layer sandwich is the most you'll need.
I didn't really intend for this to turn into such an extensive spew on my unnecessarily complicated scanning techniques -- it's just that earlier this evening I scanned another old negative to add to my gallery collection, and this one was a real bear! Because my original film exposure had been so poor, I wound up needing THREE scans (for highlights, midtones, and shadows) and two layer masks (on the highlight and midtone layers, so I could balance the tones in various areas of the scene the way I wanted them.) I'm attaching it so you can see what I mean, and also because I want to get a little extra mileage out of the scan after all that work!
As I said, other people probably have techniques that work just as well without nearly as much bother, and all this stuff shouldn't be necessary when scanning chromogenic negatives. But as I said, it's what works for me with the wacky negatives I produce. (I remember describing this technique on another forum and noting that it's like pulling teeth, and another member who happened to be a dentist posted back, "No, actually, pulling teeth is a lot easier than that!")
Bryan Lee
Expat Street Photographer
jlw, Thanx for your description. While it might sound complecated at first it totaly makes sense to me. Did you come onto this through trial and error or did you learn it from someone?
back alley
IMAGES
wow jim!!
i have never played with layers yet.
i think i understand and will have to apply this when i think i'm ready for the challenge.
and thanks for the feedback about the scans i already have.
joe
i have never played with layers yet.
i think i understand and will have to apply this when i think i'm ready for the challenge.
and thanks for the feedback about the scans i already have.
joe
ChrisN
Striving
jlw - I'll echo Joe's "Wow" and thanks. Your explanation of both the underlying principals and the procedure is a work of art - are you a technical writer? I'm going to print out this page and see if I can make it work. Is all this possible with Photoshop Elements 3.0?
jlw
Rangefinder camera pedant
Bryan Lee said:
I've since read that a lot of people use similar techniques -- but I mostly figured it out for myself, by starting with the basic theory and then trying various ways of applying it until I came up with something with which I was comfortable.
It's like a lot of stuff in both conventional and digital photography: The basic principles apply to almost everything, so once you understand them, it gets easier to devise solutions to a lot of problems. Weird but true: the basic idea behind this scanning technique isn't much different from the "bump" and "flash" techniques used to improve the tonal range of halftone negatives made with an old-fashioned stat camera!
jlw
Rangefinder camera pedant
ChrisN said:
I'm not sure; I don't know much about what features are in what version of Elements. (I've tried using Elements -- but I've been using Photoshop since version 2.0, so it was simpler for me to just stick with that and learn the new features as they came along.)
The basic capabilities you need are: to be able to convert two (or more) separate scan files into layers in the same document; control how the top layer mixes with the bottom layer ("blend modes" and opacity controls); and adjust the levels of one layer without affecting the other.
Different software packages may have these capabilities in different forms or with different names; once you understand the general idea (capture highlight and shadow detail in different scans, and then combine them) you should be able to figure out how to make it happen in the software you use.
I know there also are ways to do this kind of combining via pure mathematics, and I've read about software that does this. What I like about the layers technique is that you can work visually, in much the same way you'd make test prints to arrive at a dodging and burning plan in a wet darkroom. And because you're making only non-destructive changes (layer opacitiy, adjustment layers) you can always change your mind later without having to rescan.
N
Nikon Bob
Guest
I have noticed that images look different when viewed in different programs and use the PS Elements 2.0 view as what it really looks like. I have never played with layers like jlw but have used the photo merge option in PS Elements to try something similar as a quick and dirty approach. It looks like I will have to learn what layers are all about in the future, thanks jlw. It has been a really good thread on the subject of scanning and thanks to Joe for starting it. Always lots to learn.
Bob
Bob
Vagabond
Accomplished Malingerer
Bryan Lee said:
This should be required reading prior to using a scanner.
Thanks for the excellent link.
willie_901
Veteran
When I began scanning it seemed to me the scanner was just a large, slow digital camera and it made sense to adjust the scan exposure as if you were using a digital camera.
There are two disadvantages to this. First, you better get that scan exposure right or you have to rescan and weeks or months later this is really annoying.
The second disadvantage has to do with Bayesian siganl-analysis theory which basically says the less you muck around with the original data (the change in the film-dye chemical particles due to interactions with light waves), the more likely your model of the data (your final image) maps onto the original data in a one-to-one fashion.
Now, it is my preference to use the scanner (Canon 9950F in my case) as an image recording devide rather than an image manipulation device. I use SilverFast AI6 as my scanner driver. SF has a 16 bit HDR mode which is similar to a RAW digital camera capture. The HDR scan is subjected to minimal software intervention.
Then I use PS to adjust the colorcast and contrast, etc. It took me awhile to get my PS workflow down for the HDR scans, but I am very happy with the results (except for under-exposed color negatives).
For badly exposed (grainy) color negatives I do use SF's degraining software and scan in 16 bit color mode (tif files) without any other changes. This does soften the image a bit so you get to choose between obnoxious sahdow grain or a slightly softer image.
willie
There are two disadvantages to this. First, you better get that scan exposure right or you have to rescan and weeks or months later this is really annoying.
The second disadvantage has to do with Bayesian siganl-analysis theory which basically says the less you muck around with the original data (the change in the film-dye chemical particles due to interactions with light waves), the more likely your model of the data (your final image) maps onto the original data in a one-to-one fashion.
Now, it is my preference to use the scanner (Canon 9950F in my case) as an image recording devide rather than an image manipulation device. I use SilverFast AI6 as my scanner driver. SF has a 16 bit HDR mode which is similar to a RAW digital camera capture. The HDR scan is subjected to minimal software intervention.
Then I use PS to adjust the colorcast and contrast, etc. It took me awhile to get my PS workflow down for the HDR scans, but I am very happy with the results (except for under-exposed color negatives).
For badly exposed (grainy) color negatives I do use SF's degraining software and scan in 16 bit color mode (tif files) without any other changes. This does soften the image a bit so you get to choose between obnoxious sahdow grain or a slightly softer image.
willie
jano
Evil Bokeh
jlw's technique can be summarised as such: bracketing your scanner exposure