Komtanoo Pinpimai wrote:
>Winsor's and your readjusted pics make me realize the scanning
>problem. In my scanning process, I would like pictures to be produced
>exactly from what are in the films, so I disable all auto options like
>exposure controls, auto focus, tone curve histogram, brightness
>/contrast/color balance, hue/sat. This setting give me a colorful tone
>in xprocessed films (scanning by professional shop gives a little
>better result), so I had been believing since then that the setting
>returns correct result...
>
>
This Purity of Essence approach MAY work with slide film, depending on
the specific scanner and software involved. It will almost never give
the results you want with color neg film. There are two technical areas
of reasons for this.
First, scanners are not 100% accurate out of the box and may, probably
will, drift in exposure and color accuracy over time. There are various
ways to compensate for or to calibrate accurate adjustment for these
inaccuracies. By disabling all the auto controls in your software, you
make it impossible for it to compensate. If you are going to leave auto
focus off, which I understand you are doing, you then need to focus
manually, a big pain. Similarly, if you leave exposure fixed, you must
make your own adjustments to get it right, and so on. Not knowing
anything specific about your scanner and software, these generalizations
are all I can offer. except that many people are getting great scans out
of the Dual Scan IV, so it can be done.
When you say xprocessed films, do you mean slide film? I ask because the
seconr broad area of potential problems is in the difference between
slide/reversal and neg films.
Slide films are generally pretty straightforward. They have limited
esposure latitude, so any properly exposed slide has the range of scene
brightness that original exposure has selected spread across pretty much
the same range of film denisties from clear to almost fully opaque. They
also all have base colors that are quite close to clear and neutral. So,
if yo are happy with reproducing the color biases of the particular
slide film, scanning is straightforward. And that's what most people
want. If they shoot Velvia, they want the scan to look like Velvia, not
either like Ektachrome nor some result color corrected to neutral
characteristics.
With color neg, everything is different and more complicated.
1. The film base is an orange tone, slightly different within brands
and more so between brands. This has to be removed in processing of the
raw image data from the scanner.
2. Neg film captures a much broader range of scene brightness, but
records it in a narrower range of film densities.
3. This means that the same subject brightness range as a slide would
capture may be recorded over several different possible density ranges
on neg film and still be scanned to give the same range of brightness in
the results by using different exposures in the scanning process. That's
why we say neg film has great exposure latitude, because a considerable
range of different original exposures can still give the same displayed
result.
4. Sooo, exposure and expansion/contraction of brightness range are
important parameters for scanning neg film.
I generally agree with you about the idea of turning off things like
auto color. However, with neg film, one must either let the software
make some automatic decisions, tune each differently exposed scan
manually or profile your films.
Film profiling is a process that involves taking a picture of a standard
IT8 test chart with each film you use, then scanning it in software that
will make a table, called an ICC profile, that maps the colors/densities
on the film with the known colors of the chart. Applying that profile to
the raw scan data, either in the scanning software or an image editor,
corrects the results to be consistent and accurate between different
films. Here are some examples of scans otherwise properly done, but with
and without ICC profile <http://galleries.moosemystic.net/VuesProf/>.
Now, the non-profiled images can still be corrected in an editor, but
correct balance becomes a manual function for most shots.
When scanning negs without a profile, you have a range of recorded
subject brightness which is broader than can be printed or displayed.
Yet it is recorded across a narrower range of film densities than is
used by slides, comnpressed, if you will. So, getting a good image out
simply requires some decisions about what range of densities to capture
and how to map it to the output to be displayed.
You can do it yourself or let the software do it, but If it isn't done,
many or most of your scans will be poor. For myself, I have profiled
most all the films I currently use. For the too many old rolls of filma
that aren't made any more, I have to do without a profile. I use VueScan
software for scanning. What I generally do for all beu a very few,
difficult images, is set VueScan to spread the range of recorded
densities across the full range of a 16 bit output file.
This gives a very "flat", low contrast looking image, but means I have
all the brightness detail recorded on the film available to me in my
image editor. There, I can make my own decisions about what part of
highlights and/or shadows to clip off and or compress and what contrast
to set for the mid tones, etc. etc. Here is an example of scanned output
working this way
<http://galleries.moosemystic.net/Garden04/pages/3649_15s.htm>. The
result after adjustment may be seen by clicking on the up arrow. The
other images in this gallery all show the original scan and the results
after processing in PS. I've learned more since I did this exercise,
bought a higher rez scanner and use IC profiles for current work, but I
think they serve to illustrate my point.
If you are not willing to do all this, you need to either shoot slides,
find a combo of auto or semi auto settings in your scanning software
that give pleasing results without your intervention, or use profiles films.
>I still couldn't find a setting that produces good color/exposure as
>http://www.moosemystic.net/Gallery/Others/Komantoo/, especially,
>http://www.moosemystic.net/Gallery/Others/Komantoo/pages/6a.htm. Could
>you explain your adjustment ?
>
>
First of all, what I did is far from as good as could be done with
original scans of higher spacial resolution and higher brightness/tonal
resolution than these compressed JPEGS. Aside from the junk on the film
from poor washing/drying (which can probably be removed with rewashing
with something like photoflow), there is probably more on the film than
can be recovered from the JPEGs.
That said, I didn't do much to these images. Mostly, I used the
Highlight dropper in PS Levels or Curves and clicked on the brightest
bit of snow white in the image.Sometimes it took tring a few spots
before nice color balance showed up. This simple approach is appropriate
with these shots simply because they all have lots of nice snow in them
as a reference for pure white. What this does is two things. First, all
these scans have all the colors bunched up in the bottom part of the
histogram. Second, they all have bad color balance. Using the dropper
both spreads out the histogram and balances the image color. These two
steps would have to be done separately in an image without a pure white
in it.
As I recall, that and a bit of curves adjustment is all I did to most of
the images. For image 6, I also boosted the saturation.
>Since some pics such as http://rrboard.orisma.com/tmp/xxx/12.jpg looks
>okay with the same scanning process, I wonder if I need to adjust all
>scanned images, image by image in photoshop?
>
I already covered all this above.
>The scanning software from my scanner(Minolta DiImage Dual IV) has not many
>options to play,
>one of my friends suggests "vue scan", has anybody ever heard it ?
>
>
As I said above, I use VueScan exclusively. It is a great piece of
software which can do pretty much anything with almost any scanner,
often many things the software that comes with the scanner can't do. It
is constantly being upgraded, but last I looked, did not have a curve
function, which is fine with me, as I do that in PS, when needed.
On the other hand, the user interface can be confusing. Partly that's a
function of Ed Hamrick's abilities as an interface designer, partly,
it's a result of the extreme flexibility of the program. I think it is a
truely great piece of software. Some others have found it less than
transparent to use. I do know there are simpler modes that are to make
it easier for beginners and the non-technical, but don't know about
them, as I've been using it for years and use full advanced menus for
everything.
>Nice suggestion, having researched John camtech service, I feel like I
>really need to send the 1n to be "overhauled" -- the
>light-metering/shutter/apature should be working correctly and the 1n
>should swallow the normal battery, but before that I think I need to
>test a couple of rolls as your previous suggestions.
>
>
Well, at this point, it's really not possible to tell what parts of your
troubles are camera and which are scanning. Doing the tests series
should clear that up nicely, particularly as you can simply view the
slides without scanning to determine what's up with the camera separate
from the scanner.
Moose
==============================================
List usage info: http://www.zuikoholic.com
List nannies: olympusadmin@xxxxxxxxxx
==============================================
|
