PDA

View Full Version : First drum scans



paulr
10-Nov-2008, 15:57
I'm getting ready to do the final prints on my first serious color project. The negatives are 2-1/4, and with wishful thinking I scanned them on my home flatbed, which does an incredible job on large format black and white.

But with medium format color, the quality is only good enough for work prints, promo stuff, and the web. So I spent a bunch of my grant money on drum scans, which means doing all the photoshop heavy lifting all over again. Big drag.

Before I get going, I want to throw out a question about these scans. The color looks extremely saturated. This in spite of the film being color negative portrait film (mostly fuji 160) shot with an ancient single coated hasselblad lens. My own scans (done with vuescan) have the more muted look that I'd expect. Is it possible that the scanner operator boosted the saturation? Is this a normal thing to do?

I'm going to want to desaturate the scans a bit, and since I've never dealt with that before, I'm not sure what the best tool is for the job. The hue/saturation panel is the obvious tool ... is it the recommended one?

Thanks!

Bruce Watson
10-Nov-2008, 16:48
The color looks extremely saturated. This in spite of the film being color negative portrait film (mostly fuji 160) shot with an ancient single coated hasselblad lens. My own scans (done with vuescan) have the more muted look that I'd expect. Is it possible that the scanner operator boosted the saturation? Is this a normal thing to do?

Is it possible? Depends on the software. Some software perhaps has controls for this. Some (most?) does not. Is it a normal thing to do? I would say no. At least this drum scanner operator has never done it, or been asked to do it. What's normal is to try to get everything one can off the film.

What you are seeing is most likely the additional clarity you get from a good drum scan using a good solid fluid mount on a drum. A consumer flatbed scan often looks somewhat veiled in comparison. This is normal.

Another possibility comes to mind. That is, that the film in question had a restricted density range (yes, I'm guessing here ;-). Say for example, it was made at dusk on a very overcast day. The scene maybe metered just a couple of stops of SBR. What a scanner will do with this is to take the restricted density range and spread it across the entirety of the scanner's digital range. This can have the visual effect of "decompressing" a small SBR scene into a full range image. You can get this effect because most drum scanners have the software set the range on the log amp circuits, so the scanner really does scan just the density range the operator specifies for each channel when he/she sets black and white points. I've never seen a consumer flatbed have this capability which may be why you don't see this with your home scans.

Tell us more: What film? What's the scene like (post it so we can see maybe)? Exposure? Processing? What kind of scan did you specify? How many bits/channel? And what's your working space?


I'm going to want to desaturate the scans a bit, and since I've never dealt with that before, I'm not sure what the best tool is for the job. The hue/saturation panel is the obvious tool ... is it the recommended one?

That's the tool I use for that duty, so yes. There is a master desaturation, and you can also desaturate the individual colors individually. Using the master is similar in some ways to using a perceptual rendering intent. Desaturating the colors individually is similar in some ways to using the relative colorimetric rendering intent.

paulr
10-Nov-2008, 17:23
Thanks Bruce,

here are a couple of examples (flatbed scans and corresponding drum scans).
I've done work on the flatbed scans, but have not messed reduced the color saturation in any way. The drum scans are untouched except for conversion to sRGB to match the profile of the others.

Michael Chmilar
10-Nov-2008, 17:28
What was the original ICC color profile in the scans?

Did you do a "Convert to profile..." or "Assign profile..." to put the images in sRGB?

paulr
10-Nov-2008, 17:54
Original for both was adobe rgb, which is my working profile. I used convert to profile, which changes pixel data but attempts to maintain the appearance. There was no significant visible change when I did this.

Tyler Boley
10-Nov-2008, 18:23
Paul, there is no real way to utilize profiling when scanning color negs. Basically a space is simply assigned, and if the scanner software will allow previewing how that will look in the end, the rest is up to software color controls and the operator. I have seen a fair number of over saturated color neg scans for this reason...
So if color is not as desired when opening in PS and viewed in a color managed situation, you're left with editing in PS, as you mention. There is no reason to hesitate doing so...
Tyler

Bruce Watson
10-Nov-2008, 19:22
Looking at your examples, I'm thinking it's probably a little of everything. Some comes from the drum scanner lifting a few veils, some is from some image expansion due to the way the software can set the hardware parameters for scanning, some from the lower density ranges on most negatives compared to the density ranges of trannies.

And what Tyler is saying too -- can't ICC profile a scanner for negatives. A drum scanner usually works in its own internal working space (most drum scanners were designed and built before the Adobe RGB 1998 working space was invented). It hands the file to you raw without a working space assigned (generally, but some software may allow the assignment of a working space). You open the file in your photo editor of choice and either assign a working space or convert it to a working space. Like any raw capture it's going to need some editing to be "right" however you define that.

Bottom line is that these scans look to me to be pretty reasonable for raw scans. You think they are too saturated; we all must defer to your expertise as the artist. Edit to suit. I've found hue and saturation changes are pretty forgiving in Photoshop. That is, you can be pretty severe without creating much in the way of artifacts. So be as bold as your vision and good luck with your current project.

paulr
10-Nov-2008, 19:33
Cool, thanks everyone for the feedback.
I just wanted to run this by people with more experience dealing with color negs rather than risking repeating a mistake 20 times on 20 negs.
I'll just procede as normal and desaturate a bit.

Steven Barall
11-Nov-2008, 08:51
That was a very interesting discussion. I work in color neg as well and I now know stuff that I did not know before. Thanks to all.

Michael Chmilar
11-Nov-2008, 10:45
It sounds like assigning a profile to a neg scan is rather arbitrary.

If that is the case, you might try using "Assign Profile..." on the original data to assign something other than AdobeRGB, and see if you prefer the result to assigning AdobeRGB. For example, performing "Assign Profile..." to sRGB, instead of "Convert to Profile..." might get you closer to what you want. (Then again, it could also take you farther away....)

Lenny Eiger
11-Nov-2008, 11:19
It sounds like assigning a profile to a neg scan is rather arbitrary.

If that is the case, you might try using "Assign Profile..." on the original data to assign something other than AdobeRGB, and see if you prefer the result to assigning AdobeRGB. For example, performing "Assign Profile..." to sRGB, instead of "Convert to Profile..." might get you closer to what you want. (Then again, it could also take you farther away....)

It is easy enough to desaturate a little.. with the hue/saturation tool. I would stay very far away from sRGB. There is no reason to use such a compressed space...

Lenny

Michael Chmilar
11-Nov-2008, 11:59
Well, I only used sRGB in my example because paulr mentioned that he was using it as his working space.

Greg Miller
11-Nov-2008, 13:29
Based on your posted images, it looks to me like what Bruce mentioned in his paragraph #3 has happened. The result is manifested as an increase in contrast. Light tones have become lighter in the drum scan and dark tones have become darker. Increasing contrast usually has the appearance of increased saturation in color images.

I think applying a curve to reduce contrast would be worth a try as an alternative to lowering saturation. I would try this with blending mode set to "luminosty" in addition to "normal" just to see if one gets you closer to your ideal than the other.

Tyler Boley
11-Nov-2008, 14:19
It sounds like assigning a profile to a neg scan is rather arbitrary....

Well it is and it isn't, and this could be part of a larger thread about scanning color negs. It depends a lot on the software. Sliverfast, for example, is color managed through and through if one sets it up to be so. So, what color space you have set for it to tag your file will be used in the preview, as well as the monitor profile. Therefore, if a large space has been used that gives an oversaturated result, the preview shows that, and appropriate edits for correction are done at that stage in the software, the resulting file is tagged (assigned) properly and comes into PS with no surprises.
Color negs are frustrating to scan, color management barely applies, subjective decisions must be made, some scanner software tends to be less than optimized for it. When scanning for others one has no idea what the original scene may have looked like.

Anyway, if the software and PS edits get you where you want, I guess the appropriate assignment would be one large enough to contain useful color and works well down the line...
I use Beta RGB, to contain drum scanned color from trannies without clipping, but it requires work to keep negs from over saturating.
Tyler
http://www.custom-digital.com/

paulr
11-Nov-2008, 19:31
Well, I only used sRGB in my example because paulr mentioned that he was using it as his working space.

I use it as a destination space for images posted to the web.

Working space I use Adobe RGB.

paulr
11-Nov-2008, 19:33
I think applying a curve to reduce contrast would be worth a try as an alternative to lowering saturation. I would try this with blending mode set to "luminosty" in addition to "normal" just to see if one gets you closer to your ideal than the other.

I'll try this first. Will work on the first image tonight.

poco
12-Nov-2008, 05:07
Paul,

Aside from saturation, how do the drum scans compare to your flatbed scans? I ask because I recently had my first drum scans done and was completely underwhelmed by the difference in resolution, clarity, shadow detail etc... compared to my 4990 scans. The drum scans were a total waste of several hundred dollars.

Don Hutton
12-Nov-2008, 06:40
Paul,

Aside from saturation, how do the drum scans compare to your flatbed scans? I ask because I recently had my first drum scans done and was completely underwhelmed by the difference in resolution, clarity, shadow detail etc... compared to my 4990 scans. The drum scans were a total waste of several hundred dollars.Who did the drum scans?

poco
12-Nov-2008, 07:23
Buffalo Big Print did the scans -- apparently there's a reason they're not known for their work.

The first (reddish color) file is a tight crop of the drum scan, and the other one was done on the 4990. Note the writing on the crane boom (the word "equipment") and the blotchy transition of colors and shadow blockage on the drum scan.

claudiocambon
12-Nov-2008, 10:41
Paul,

Have you asked the service bureau to give you a less contrasty version of the scan? They should have the raw version, and be able to do this. I often work with mine (Digital Fusion) to come up with the ideal meeting point between where they start and where I am coming from, and I think both of us like where things end up (ie, it's often new ground for both of us). Sometimes it means more money in terms of PS time, but it's always been worth it.

Also, have you printed out the two types of scans out to compare? I work backwards from there to the scan, and find this helps.

Good luck!

Don Hutton
12-Nov-2008, 11:13
Buffalo Big Print did the scans -- apparently there's a reason they're not known for their work.

The first (reddish color) file is a tight crop of the drum scan, and the other one was done on the 4990. Note the writing on the crane boom (the word "equipment") and the blotchy transition of colors and shadow blockage on the drum scan.Why on earth would you throw "several hundred dollars" at an operation whose expertise in this area is completely unknown?... Scanners are machines and are dependant on a human operating them... It's astounding to me that you wouldn't have had a few samples done first at various places and then go from there.

Lenny Eiger
12-Nov-2008, 13:06
Why on earth would you throw "several hundred dollars" at an operation whose expertise in this area is completely unknown?... Scanners are machines and are dependant on a human operating them... It's astounding to me that you wouldn't have had a few samples done first at various places and then go from there.

I'm with Don on this one - and its a point I've made a few times. If you want a certain quality its always better to go with a smaller shop where you can discuss what you want out of the scan with the operator. I operate this way, and there are a lot of others, right here on this forum, that do as well.

I would stay away from labs, nancyscans and scans from India. Sure, its more expensive, but if quality is what you are after, I think its the only way to go - and you actually get what you pay for. The smaller shops are almost always fully guaranteed, they will try different settings for you and scan things over happily until you are fully satisfied - because their success is usually predicated on people passing their name along and/or coming back or more.

Lenny

poco
12-Nov-2008, 14:09
I didn't start with a single test scan because the guy was on the verge of sending the scanner in for repairs (it was blowing bulbs) and it was either get the total number of scans I wanted right then, or wait 5-6 weeks. But you're right -- it turned out to be a bad roll of the dice on my part.

Bottom line, I'm close to throwing in the towel on outsourcing any work to labs. One quite well known place in Toronto has apparently lost my files and refuses to answer my emails, this guy in Buffalo clearly doesn't know what he's doing... It's ridiculous.

Lenny Eiger
12-Nov-2008, 14:35
Bottom line, I'm close to throwing in the towel on outsourcing any work to labs. One quite well known place in Toronto has apparently lost my files and refuses to answer my emails, this guy in Buffalo clearly doesn't know what he's doing... It's ridiculous.

Sorry you've had such a rough time... FWIW, you're not alone, I had a very bad experience (not being paid, also known as being stiffed) with a well known place in toronto as well. Same thing, just stopped answering my emails and phone calls.

Lenny