Multiple Manual Focus Points - Dimage 5400

[jpursley]

The following applies to the Minolta Dimage Scan Elite 5400

(original version):

 

Using manual focus for scanning mounted slides generally gives me

much better focus accuracy than autofocus, but I'm having some

problem getting completely reliable results - sometimes the focus is

spot-on, sometimes it's close but not quite ideal.

 

I've noticed that with both slides and negatives, the manual focus

black bar always has two 'optimal' focus points. If you move the

thumb in the scrollbar all the way to the left, for example, and

slowly press the right arrow, the black bar will become longer and

longer. At some point it will peak and start to become shorter. A

few presses of the right arrow after that, the bar starts to become

longer again, and generally peaks again at the same spot it

previously peaked. Continuing to press the right arrow makes the

black bar shorter and shorter after this second peak.

 

Perhaps the peaks are the close and far side of the film

surfaces...I'm not sure where the best focus would be if this is the

case.

 

Do you experience this double peak when using manual focus? Have

you found better results with the left peak, the right peak, or

maybe 1/2 way between the two (even though this means the black bar

will be shorter than otherwise)?

[mendel_leisk]

I'm not getting a double peak, sorry. Some thoughts on focus, anyway:

 

With (pretty cupped) slides, emulsion down in the holder, I need to go to the left quite a bit. I prefer the manual knob, and find I end up around 9 oclock, if you picture the knob and it's pointer as a clock face. Incidentally, the knob and it's pointer will revolve even if you're using the mouse technique.

 

I do pay close attention to pointer on the knob. You can do one focus, then after it re-pre-scans, tentatively try a second focus at a different point and see how far and which way you would need to turn.

 

I do not batch focus, i.e.: focus all four frames and then batch scan. I batch pre-scan, but then I scan them one at a time, right after each focus adjustment. I don't trust (from experience) the focus to return to the right spot, or the film not to flex in the interm.

 

I've found GENTLE warmup with a hair dryer "conditions" the slides to flex to their warmed curvature, and reduces focus drift, especially on the first frame.

 

After each scan, I immediately open the result in Photoshop (which I usually have open anyway: cloning whatever ICE has missed on prev. scans), and pan around a bit reviewing how well the grain is resolved. The firewire connection works good for this. Having PS and Minolta Scan Utility open at same time causes no speed reduction with the firewire connection, at least none that I can notice.

 

Focus is always a bit of a toss-up with mounted slides. I try to focus off-center a bit, at a point of interest (face, hand, etc.) and on an area of light, even tone.

[mendel_leisk]

Ok Jeremy, did see the double focus on one slide, focussing on a relatively dark, uniform area. As you speculate, perhaps it is focussing on the near and far side of film. I _think_, if you start with the slider all the way to the left, the first focus point would be the emulsion. That is assuming you are following the manual's guidance to face the emulsion down when loading into the holder.

[mendel_leisk]

All right. Tried to be as careful as possible, following same steps each scan, focussing on same spot. Started from left extreme (using the manual knob). Did first scan at first focus peak, second scan at second focus peak.

 

I'm seeing the first peak from the left extreme is VERY slightly sharper.

 

Jeremey, hope you're still check this thread. Please repeat my experiment and post your results.

 

For you info:

 

1. My slides cup on the emulsion side, fairly typical case.

 

2. The whole holder moves left to right similar to the mouse, or control knob.

[mendel_leisk]

The more I think about it, the more I THINK the opposite should be true. In the process of moving the adjustment from left to right, the holder is doing the same movement: left to right. I THINK the business bits of the scanner on the left, and the light source is on the right. You'd THINK the first thing it would find to focus on would be film backing, THEN emulsion. Arggh, too much thinking.

 

Anyway, do check it out, see if you concur left peak is sharper.

[mendel_leisk]

Hi Jeremy, I think I've nailed it:

 

The scanner's focus is FIXED, at a certain distance in front of the lens. With the holder at the left extreme, with the emulsion facing the lens, that focus point is beyond the film, to the right of it. As you move the knob/slider to the right, the holder is moving to the right, and that focus point will first encounter the right, backing side of the film (the first peak), then travel thru the film, and then the left (emulsion) side of the film.

 

Accordingly, the second peak is at the emulsion. My observations of scans at both peaks agree with this.

[jpursley]

Well I stumbled across this old thread today and followed up with a bit of testing: I scanned an image at the left peak, the right peak, halfway between the peaks, and with autofocus - all on the same spot on the mounted slide.

 

The left peak image was just slightly sharper than the midpoint image, which itself was just slightly sharper than the right peak image. These were very, very small differences, though, that probably would not make a difference in even large prints. The autofocus sample was significantly less shart than any of the manual focus images.

 

Interesting to hear that you get similar results, Mendel. Thanks for following up on this...hopefully other 5400 users can benefit from these posts.

[mendel_leisk]

Hi Jeremy. I think we should clarify our descriptions, and settings.

 

1. I'm assuming that you, same as me and per the manual, place the film in the holder emulsion down. This in turn means the emuslsion is facing the scanner's lens. As you face the front of the scanner, the lens is to the left, and the light source is to the right.

 

2. Moving the focus slider to the left, or turning the physical focus knob counter-clockwise, moves the holder and film to the left, towards the scanner's lens. And vice versa.

 

Now, when I move the holder to the left extreme (closest to the lens), and then start slowly moving it to the right, I get 2 peaks. For me, the second peak, when the holder is further to the right and further from the scanner's lens, is the one that gives me the sharper grain. As I said before, I believe the first peak (holder closer to lens) is at the film back, and the second (holder further from the lens) is at the emulsion.

 

You say you're getting similar results, but you say your left peak is sharper. Not sure if it's the terminology tripping us up, but I think you're saying your results are better with the opposite peak to mine. To clarify, my sharpest result is with the second peak, starting with holder at left extreme of it's travel (closest to lens), and moving it to the right.

 

Now, something tricky I've noticed, going thru the following procedure:

 

1. I move the holder to the second peak from the left limit of travel, and carefully optimize it, but don't click ok yet.

 

2. I turn my back on the scanner, for at least a minute. Check out "Unified View", or whatever.

 

3. Turn back to the scanner. What do you know: usually, the black bar has fallen back, to be maybe half the length of the white bar.

 

4. What is required to re-lengthen the black bar to the white bar extent is to move the holder slightly to the left, moving it closer to the scanner's lens. Since this process is likely still in flux, even after a minute or two, when moving the holder leftward, I will overshoot the correction, very slightly. Watching the bars for the next 15 sec or so, I'll usually see the black bar flicker a bit, getting a little closer to the white bar length.

 

What I think is happening here:

 

My film (fairly typically I think) cups on the emulsion side, domes on the backing side. Placed in the holder, with the emulsion facing lens, and backing facing light source, it starts to warm up. In that process, it curves even more, especially the backing side. Accordingly the emulsion is moving away from the scanner's lens, as this warming takes place.

 

I believe I can compensate for this factor by just letting everything "sit" after optimization of focus, and then moving the holder a little closer, with the _slight_ overshoot (to allow for the little residual expansion that might be still to come).

 

The following is kind of off-topic, but will likely be of interest to you, if you haven't been following it:

 

My current project is scanning mounted slides, from the 70's through to near present, and the older ones in particular are quite cupped around the emulsion. Because of the 5400's very limitted depth of focus, I've taken to dis-mounting them, and remounting in Gepe 7012 mounts, with shims inside the mount to try to:

 

1. Counteract the cupping.

 

2. Scan close to full-frame.

 

The 7012's are relatively expensive, so my strategy was to get only enough for a roll or so, enough for a scanning run, and then a VERY large quantity of inexpensive Gepe mounts (tho still decent quality, 2 piece mounts, easily dsassembled) for final (yeah right!) storage.

 

The following threads have my most current comments relating to the depth of focus issue, and my efforts with the shimmed Gepe 7012's:

 

http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=00DgKE

 

http://www.photo.net/bboard/q-and-a-fetch-msg?msg_id=00Dkd5

 

In the second link, the picture I posted, titled

 

"Gepe 7012 mount with shims"

 

shows my latest shim arrangement. It uses 3 layers of 3M Post-It notes. I place the film chip into the grey half, emulsion up, tucked into the metal mask clips at top and bottom. My technique for doing this is shown in detail in the first link.

 

One other idea, Erik, the Scanhancer maker, suggests reversing the film so that the cupped side faces the light source, to reduce the flexing due to warm up. I've tried that with inconclusive results, so a staying with emulsion facing scanner lens (and away from light source), for now.

 

Incidentally, between my findings on focussing technique, and the shimmed Gepe 7012 mounts, I am making headway, regarding improvements in both extent and accuracy of focus. The shimmed mounts alone place the center of the film and the edges at approximately the same "elevation". The domeing sort-of pops out at localized spots, near the four corner, but no where near as pronounced.

 

I am also entertaining getting another scanner, if that's what it takes, likely the Coolscan 5000. But that is a MAJOR step, and would likely bring new hassles. In that regard, I have mailed another photo.netter, who has a Coolscan 5000, a couple of slides. He has graciously offered to scan them, for focus comparison.

 

One further observation, very puzzling, very interesting: without grain dissolver, I DO NOT get double peaks. None whatsoever. Very strange...

 

Well, hope you check in on this thread. Do please clarify which peak it is that is giving you better results, closer to the scanner lens, or further from the lens. Again, for me, with emulsion facing the lens, it is the further-from-the-lens case that is sharper.