Scanning wrong, or developing wrong? A tale of two rolls

Takkun

Takkun

Ian M.
Local time
1:51 PM
Joined
Jun 7, 2006
Messages
876
Hi everybody,
After spending much of the last few months shooting 4x5" and digital, I'm back to smaller format. Maybe I completely forgot what I'm doing.
The situation: Normally I shoot TMY-2, develop in HC-110 dil. H, 11 mins, agitation every 30 seconds. Recently I got ahold of a brick of HP5 medium format, and my brother bequeathed me a dozen or so rolls of Delta 400 and HP5 in 35mm format after a drawer clean out, so that's what I've been shooting lately.
I decided to rate it at 320 the way I do Tmax, and develop in dilution B to cut down grain. Massive Dev and Covington both say 5 minutes.

The MF scans looked pretty superb. Nice tonal range and good separation, or at least as good as it gets without a contrast filter. These were scanned on an Epson V500 I have access to.


See what I mean?

Just got done doing two batches of 35mm, scanned on a Minolta.




In a word: Ew. I tried every trick in VueScan to get a good histogram, but it always ended up a very narrow band, and increasing gain just moved it up and down the histogram scale. I know conventional wisdom says that one wants a relatively flat scan to apply curves to, but there's virtually no tonal separation here.
What's baffling me is that you can see I'm getting density in the highlights in both of those 35mm scans, and they're blocking up. And despite that, they're not getting past the halfway point on the histogram. Most confusing is that they were developed identically as the MF negs, that turned out beautifully.

So I tried clipping the levels in an attempt to regain contrast.

(click through for larger)
Well, now I'm getting contrast, but no mid tone separation. Fun.

So it looks like a couple things are going on. Looking at the pictures makes me think I'm developing too little. Or exposing too little. Looking at the histogram and the 'after adjustment' picture makes me think I'm scanning something wrong. Especially since the shadows are giving me a salt-and-pepper look, and not clean blacks. It's worse in Aperture, where everything starts to posterize the more I use levels and curves.

Any help would be much appreciated. I've got a lot more rolls I've shot and am afraid to develop them before I sort this out.
 
My guess is, you are including blank space in the scans, so your black point is way beyond where it should be, and the histogram flips all to the right. To avoid this, select for scanning only the exposed areas.
To verify if the development was OK, just wisually inspect the 35mm negs against the MF ones - if they look roughly the same - and they should probably look a bit lighter, as Tmax has a more clear base - then the development was OK. Should they appear dense, and the histogram still too high on the right, then select colour film, or at worst, slide film ( and invert later) in the medium type in vuescan.
 
Your 35 mm should look like 4x5 just smaller. The 35 mm scan looks extremely flat to me, but I can not see the negative.

If the neg looks flat, more time in developer is always the answer. If the negative looks normal, then there is scanning problem.

The first thing I tell people is get out of auto mode and use professional. Assuming the neg looks good, but the prescan looks light, then increase the density. Much like a darkroom, if the print is light, use more time. If you lost all the shadow detail going darker, the neg needs more contrast.

I would not use any default or auto settings. Go through the process, develop, scan until you get it right.

Since you have more film to do, make a test roll of a full tonal range subject using just 6 exposures or 12 inches of film. develop longer and see how it scans.

I would put the two negs on the Epson and scan at the same time. This may help to isolate the problem.

I never use the developing charts on the net. I find Kodak and Ilford times perfect for #2 paper and diffusion enlarger ,condenser or scanning, subtract 10 %.

The other issue is I never had any luck with view scan. I just use Minolta software with my Elite 5400 and Epson with 4780. Problem is MK software probably can not be found and may not work with new computers.

You might also try scanning the 35 mm on the epson.
 
To clarify: I wasn't using any auto settings here. I follow the locking exposure on the clear base technique, though I did try auto exposure settings (and set the buffer well within the actual frame to eliminate errors from the film rebate) with no improvement. Essentially it just moves that very narrow histogram further to the right; the highlights are lighter gray and the shadows less black with still little tonal separation.

I'll also note that the 35s here are also HP5, for sake of comparison, but I'm having this with Delta, too. It didn't even cross my mind that I should have saved this surplus film for experimenting, but I went out and did a fairly important shoot, nonetheless.
 
It's quite hard to say what the issue might be without seeing the negs and the full VueScan settings you're using.

I've been using VueScan since 2000. I recently put in a good effort such that I now feel I have a solid understanding of the processing controls. I use a raw workflow (outputting VS raw encapsulated in DNG) but output TIFFs as well, and the output TIFFs are very close to final rendering about 90% of the time.

G
 
Takkun said:
Just got done doing two batches of 35mm, scanned on a Minolta.
Click to expand...

This. I remember my pain when trying to scan B&W with Scan Dual III, finally sold it and later someone here on the RFF told me to scan it as a color positive and revert in PS.
 
jbielikowski said:
This. I remember my pain when trying to scan B&W with Scan Dual III, finally sold it and later someone here on the RFF told me to scan it as a color positive and revert in PS.
Click to expand...

I always have to do this on my Epson, but I haven't felt the need on my ScanDual IV. Of course everyone's tastes differ.
 
VTHokiEE said:
I always have to do this on my Epson, but I haven't felt the need on my ScanDual IV. Of course everyone's tastes differ.
Click to expand...

The scan as a transparency trick is great for negs that have very high contrast. It captures a wider density range than you get using the "Negative" settings on most film scanners. For normal negs, it gives an even lower contrast file that needs a little more postptocess work to get the tonality back.
 
Tried curves, and while it helped a bit, the midtones were exceptionally flat--no definition in skin tones or hair.
However, to add something else into the mix--I Just realized my Leica's battery was very, very low. Though I would assume that would mean things would be over-exposed, and not account for the negs not being particularly dense.
Ugh. Time to shoot some test rolls and avoid the computer for a few hours.
 
Crawford is an expert and you can tell by his on line photos.

I prefer to work a bit differently in that I try to get a histogram that fills out left to right and color balance correct or at least centered. It is then like a raw file from digital camera which I process in photoshop. By using the scanner software, the process is slowed too much.

Nothing wrong with using levels as a first step, then use curves for the contrast. If you want to use curves only, then use the black and white droppers to set end points. Actually that the how to color balance , use threshold to mark the high and low, then channel by channel use the droppers. Magically color balanced to perfection, but if you did not get it right in camera and have no reference, it works.

You might also try scanning in color, dup the layer and then use one of the contrasting increasing blend modes. Then change to black and white.

You can also try an intensifier. It will not bring up empty shadows, but will provide some contrast.

The rest of the rolls need more development if under exposed. They will have lots of grain and still no shadow detail.
 
I do exactly as Chris says, works every time (good to see you back Chris).
My 2c is that something is up with development, they look VERY flat. How do they look with your eyes directly? That's a good fool proof way of checking.
My scans always look quite flat like Chris says, but never that flat.

Michael
 
Chriscrawfordphoto said:
The scan as a transparency trick is great for negs that have very high contrast. It captures a wider density range than you get using the "Negative" settings on most film scanners. For normal negs, it gives an even lower contrast file that needs a little more postptocess work to get the tonality back.
Click to expand...


Thanks a lot for the info, I didn't know what was going on behind the scenes. Why does the transparency trick capture a wider density range? Why not always do that?
 
VTHokiEE said:
Thanks a lot for the info, I didn't know what was going on behind the scenes. Why does the transparency trick capture a wider density range? Why not always do that?
Click to expand...

Color transparencies (slides) have a greater density range than negatives, and good film scanners are made to capture that range. When you set the scan software to "Negative" (color or BW), the software compensates for the fact that no negative (at least not a properly exposed and developed one) has any solid, pure maximum density black areas (black on the neg, meaning whites on the print). It compensates by bringing in the Levels endpoints somewhat, so the scans do not look as flat as they do when using the transparency settings.

Even using the Negative settings, the scans will be too flat, but it is much more extreme when you set the scanner for Transparencies and scan a neg. If your neg was badly overdeveloped or shot in very contrasty light (or has a large brightness range, as with an interior shot that has an open window), scanning as a Negative may clip the light tones, making it impossible to keep detail in them. Scanning as a transparency captures the full tone range of such very high-range negs.

That doesn't mean the neg will be easy to print. When you increase contrast enough to make the midtones look good, you're still going to lose the very light and very dark tones, but if you're good with masking and layers in Photoshop, you can overcome that.
 
Thanks for the further advice. I'm guessing my dev time was horribly off, since I couldn't tell much in way of mid tones on the negs. Of course, that's a lot harder than it is on a 4x5. I was just baffled by how off it would be, especially between two consecutive batches. I'm thinking I might just pawn off what little Ilford stock I have rather than re-learn everything.
Curves helped rescue a lot of my older photos, but not so much here. They end up passable, but not pretty by any stretch.

And the transparency option is something I'll consider when I'm scanning some of my 4x5s--I'm photographing an indoor construction site and generally let open windows fall where they may on the scale, but they actually develop with some definition. It'd be nice to capture some of that on the scans. It's giving me better results with this tricky batch of 35mm film, but still not satisfactory.

I've since downloaded a demo of Silverfast. I don't really want to re-learn new software, but the results are better. I get a histogram that isn't shifted completely to the left half. Anybody want to take a stab at that part?
 
Just a thought, but the main difference between the two rolls is the scanner itself. Could the light source be going?

I didn't get along with silverfast.

Michael
 
Chriscrawfordphoto said:
Color transparencies (slides) have a greater density range than negatives, and good film scanners are made to capture that range. When you set the scan software to "Negative" (color or BW), the software compensates for the fact that no negative (at least not a properly exposed and developed one) has any solid, pure maximum density black areas (black on the neg, meaning whites on the print). It compensates by bringing in the Levels endpoints somewhat, so the scans do not look as flat as they do when using the transparency settings.

Even using the Negative settings, the scans will be too flat, but it is much more extreme when you set the scanner for Transparencies and scan a neg. If your neg was badly overdeveloped or shot in very contrasty light (or has a large brightness range, as with an interior shot that has an open window), scanning as a Negative may clip the light tones, making it impossible to keep detail in them. Scanning as a transparency captures the full tone range of such very high-range negs.

That doesn't mean the neg will be easy to print. When you increase contrast enough to make the midtones look good, you're still going to lose the very light and very dark tones, but if you're good with masking and layers in Photoshop, you can overcome that.
Click to expand...

The reason your scan looks different if you scan a negs as a positive is it applies different profiles and curves. You can apply these curves even if you scan as a negative.

Transparencies have a much shorter tonal range and the profile will optimize for a shorter tonal scale. Scanning negatives will give the full range of tones from the negative. Profiles and curves are different than transparency scans. Curves applied for a transparency will not necessarily be correct for your tonal scale in the neg. it works but may not be optimum.

With both my previous Fuji scanner and my Imacon 848 I can can a neg either positive or negative without applying curves or profiles. I don't believe this is possible but might be wrong that you can not do this with my V750.

Do your negs appear flat or thin?
 
x-ray said:
I've never seen scans like that and I e scanned thousands of negs.
Click to expand...

You have a $10,000+ Imacon scanner. It has much better software than what we have with our lowly Nikon and lowlier Epson scanners. Your experience is really not applicable to the OP's question.
 
You must log in or register to reply here.
Back
Top Bottom