I've read, or tried to read, all of the mindbending 184 post "Minolta
5400 or Coolscan 5000" free-for-all.
You have my sympathies... ;o)
I have 60 negs to scan for an art
project. I don't want to re-do these, so I'm trying to nail this out
of the gate.
In my case it's about 3000 slides and negs, but basically that's
exactly what I said about 18 months and 2 scanners ago... :-( At least
you're not starting from scratch as I did and you don't have to worry
about Nikon's Kodachrome disease which Nikon denied for 6 months
(wasting my time) and then admitted after I sent in a slide... But
that's another story...
I own a 5400. I tend to subscribe to Don's theory of scanning, i.e.
scan without corrections because scanner sw modifies the scan
post-scan and so just use PS to modify. I own the CS version, so
working in 16 bit (or whatever, 15+ bit) isn't a problem. I am very
good at Photoshop. I have color neg and b/w neg., a couple of Fuji
velvia chromes.
My experiences are limited to CoolScan (LS-30 and LS-50) as well as
Kodachromes, so take what follows bearing that in mind...
What I am wondering is:
1) using Don's theory, is it ok to use the Minolta 1.1.5 software, or
would Vuescan or Silverfast make any difference USING DON'S theory? My
guess is the Minolta sw is fine using Don's theory. I'd rather use PS,
a tool I know well, than to use some other sw that I must become an
expert at.
With above caveats in mind, I think you'll be just fine using
Minolta's software, as long as you turn off anything that may "mess"
with raw scanner data.
In my case, the only things I let the scanner do are:
- auto-exposure (when there's enough dynamic range, and your scanner
certainly has that)
- auto-focus (I may fine tune focus to compensate for film curvature,
you'll just have to experiment with this)
- ICE, when applicable (doesn't work with "dark" Kodachromes, but
that's not your problem...) However, ICE doesn't work with B/W so keep
that in mind! Do note also that ICE "softens" (blurs) the image
slightly. The consensus (and I for one concur) is that this is
negligible, especially considering the advantages of ICE. Indeed, it
can be argued that this softening even helps in that it hides the
grain. But Minolta's light source is not as "harsh" as Nikon's LEDs so
grain is not such a big problem for Minolta (at least that's what I
read here...)
- gamma (2.2 in my case) although to scan truly raw, you would use
linear gamma (1.0)
That's my "digital negative" and I then "develop" the "prints" in PS
(in my case viewable images, I don't actually print to paper).
I ran Silverfast once for about 5 minutes. With that in mind, I found
it too "cartoon-like" for my taste and it doesn't do anything to
improve raw scans. Indeed, it's "pretty" approach seem to hide
information. But others here swear by it...
I would definitely warn against VueScan, but I'm known for disliking
it so take this with a grain (boulder?) of salt! Some of my issues
with VS are:
- VS is a rolling beta (major bugs pop in and out at a drop of a hat).
Witness people scrambling to find "old" versions where this-or-that
still worked...
- No user interface. Confusing settings with multi-level
conditionality depending on other, "hidden" settings elsewhere, etc,
etc. In short, one never knows what's being applied!
- VS may still "do things" to the data even when you tell it to scan
raw (this is a subset of the above problem).
Witness a recent exchange where the author admitted that even "raw"
setting is only raw *if* a bunch of other settings and
conditionalities are "just so"...
In other words, there is a definite possibility of a firm maybe you
could, by accident, get a raw VS scan... Perhaps... ;o)
I could go on, but that's already more than enough to get VS groupies
excited... ;o) although that's not the point at all.
Oh, yes! VS apparently has a major dislike for 5400! I seem to recall
many posts about "streaks" which, I understand, remain unsolved.
2) If other sw can do a better "Don-scan" by opening some secret
hardware-level nirvana, what are the sw settings to do so?
Not that I (for one) am aware of.
Another thing to, maybe, bear in mind is that even if you eliminate
gamma, auto-exposure, etc. and even get the data directly from the
sensors, you will never get the same scan twice! At this level of
precision the slightest change in temperature, for example, will
result in different values. This was a bit of a revelation to me, and
certainly puts things into perspective when chasing that (apparently
imaginary) "perfect" scan without any corruption.
So, by all means, strive for perfection, but every now and then I
found it worthwhile to step back and do a reality check to see whether
things I do make sense within the given context, in other words, I'm
getting as much out as I put in.
3) AND, whatever the answer to the 2nd question, what are the
setttings in the Minolta software for a raw-type "Don-scan"?
Again, bearing in mind that I don't own a Minolta, my 2c...
My guesses are:
- 16bit linear (vs 16 bit),
If you apply display gamma later, that's fine, although in my case
NikonScan does a better job than PS. Don't know about Minolta
software...
- don't know about about autoexposure vs manual,
I touched on this above. Autoexposure will automatically boost the
exposure (this is just like taking a "real" photograph) to get the
"optimal" results. Again, you may wish to adjust this if, for example,
your image has unimportant highlights which you don't mind blowing if
in return you get more detail in shadows.
The only thing (and I don't know Minolta software) may be that
autoexposure often has "default clipping". It's an estimate that a
certain amount of data at either end (dark or light) is really noise.
So the scan is intentionally clipped usually in the 0.3% to 0.5%
range. If you really want "everything" you may want to set these two
values (if applicable in Minolta) to 0% clipping.
I find that when scanning mounted slides, the film is quite curved and
focus is a bit of a nightmare. You set it for the middle and the edges
are all out of focus. Do note that I examine the image at 300-400%
(and more!) in PS and that under normal viewing conditions this is not
noticeable.
One option is to take the slide out and use non-Newton glass mounts
which will then press the film flat.
But wait, you said, negs... I don't know what loading Minolta has, but
Nikon has "automatic" and film-holder. I personally hate automatic
(there are also alignment problems, apparently) and always use the
film holder. This has the added advantage of keeping the film
flat(er). Although, if that's not enough, you could cut up the film
and use the above mentioned non-Newton glass slide mounts.
Finally, there is a vicious rumor going around that Nikon's film strip
holder window is too small and that it "clips" the image slightly, but
that's really only a rumor. I don't know if this rumor extends to
Minolta's film strip holder.
- don't know about multi-sampling
It's a Very Good Thing! As long as it's single-pass, which apparently
your Minolta does have.
I don't have it, so I personally scan (almost) every image twice (once
for shadows, and once for highlights) and then combine. But, depending
on your requirements simply using multiscanning may be enough.
Oh wait... you said negs. I haven't done negs yet but Kennedy will be
around shortly... ;o) Because negs are reversed, multiscanning is a
*must* because dark areas are where the noise hides, and once inverted
it will become very apparent.
- don't know about color matching output space (adobe RGB?) and ICC
profiles
I turn Nikon Color Management off because it's just broken. I then
assign sRGB in PS. Don't know about Minolta...
- Color neg: best to scan as neg or positive? If positive, invert in
PS?
- B/W neg: best to scan as neg or positive? If positive, invert in PS?
- Fuj chromes: no idea
I'm still wresting with Kodachromes so I'll let those who know better
handle it - and makes notes for when I need it! ;o)
- ICE on, at least with color neg.
Yes, as explained above.
- Grain dissolver off with color neg, don't know with b/w neg
Again, I yield to others...
Thank you for your answers.
Thank you for your patience... ;o)
Seriously though, sorry, I didn't mean to be so verbose, but I hope it
helps.
Add this to what others say and then "average out" to eliminate noise.
Sort of, like multiscanning... ;o)
Don.