Nikon LS-9000 vs Imacon 848: Resolution & Shadow Detail

"I have found the the Nikon Coolscans as the best desktop scanners in terms of color/exposure accuracy, resolution and dmax."

 

Have you tested the Coolscans directly against an Imacon - or is this just your opinion from using the Coolscan?

Okay, let's get this thread back on track and away from personal attacks.

 

What we need are reasonable tests. Testing the Nikon 9000 has got to happen with ICE off and at normal (not fine). Ditto for Vuescan- if you use that do none and low cleaning. With Vuescan also use the manual focus tool to see where the focus differs in different parts of the image and find a value more or less in the middle.

 

I agree with the concern about using bicubic sharpen for downsampling, and would be interested in a comparison of the Imacon at 8000 vs 4000dpi to see if it is capturing any more useful information.

 

I'm not sure anything above is a reason why you can't use Vuescan. You do need to test to make sure IR is working right in your version. If it is, what's the problem? If it's not, search for past versions which worked or send raw scans to Ed Hamrick and he will fix it.

The wet scan that I posted has no sharpening applied.

Les,

<p>

Who's continuing to do the same thing hoping things will end differently? You, just trying to criticize me? Or me, by testing Vuescan, Silverfast, Nikonscan, a LS-4000 with and without Scanhancer, LS-5000 with and without scanhancer, LS-9000, and an Imacon 848? Or me, by taking people's advice when they offer it (read my last few posts on this thread), and trying new things/changing my method? How is that just 'doing the same thing hoping for different results'? If you're going to criticize me, do it well. Are you still bitter over me being right about the presence of pepper grain in my last thread, while you just thought it was my problem in the face of many other people clearly attesting to its presence in their own scans? If so, get over it.

<p>

And regarding the different colors with Vuescan's different 'white balance' settings, any knowledgeable scan operator knows not to use any of those settings, and either use Erik Krause's Advanced Workflow with Nikon scanners, or build a color profile instead. Or some combination of both. Contrary to what most people write, it *is* possible to build profiles even for negative film. So don't come talk to me about color accuracy of your scans until you've attempted to build a profile off of a color chart.

<p>

Finally, do you think your method of communication works? How far does it get you? It certainly doesn't get me to listen to you. Because all it's done for me is pretty much made me ignore most of your responses when you write them.

<p>

Whereas notice that I actually have *conversations* back and forth with everyone else on my threads. A bunch of other people on this very thread have criticized my scans and potentially my method, but have done so constructively, and note I have taken these criticisms with open ears, thanking them (read my post to Scott, above, for example). So, no I don't take criticism and the presentation of facts personally. I take your name-calling and illogical (as I've shown above in your assertion that I'm repeating the same thing) attacks personally.

<p>

Advise you to start waking up on the right side of the bed. And take your pent-up anger from who knows what and your 'looney toon' name-calling elsewhere. Not that I can stop you, but I can tell you that <b>you're not welcome in my threads</b> until you learn how to civilly communicate.

<p>

Rishi

Les: "While you are obsessed with the tiniest specs of creepy crawlies - doesn't matter in the end whether it is

printing or web posting"

 

Who says it doesn't matter? You? On sufficient enlargements, I see pepper grain in highlights from my Imacon

scans. Why do you think the Imacon 949 revision put a diffuser in their light path?

 

Why do you make unjustified claims? And then pretend like *you're* the rational, reasonable one? Someone correct

me if I'm wrong, but, isn't that what a 'loony toon' does?

 

Rishi

Roger, thanks. Yeah it's possible that IR or autofocus or something got broken on the latest Vuescan. I still haven't had a chance to test.

 

Also, none of those have IR cleaning set to 'fine'. I never use anything other than 'normal' for Digital ICE, or 'light' for Vuescan's IR cleaning.

 

I should have some time tomorrow on the LS-9000, so I'll try the same scans with IR cleaning completely off, as well as give you a range of manual focus readings from different parts of the film.

 

Cheers,

Rishi

Rishi,

 

"So you're saying that you do see scans out of Nikons, at 4000ppi, that look as sharp & detailed as the Imacon

scan above? I'm just checking to make sure I understand. "

 

Yes. Absolutely.

 

The one thing I see in your examples that does ring true to me is just a touch more shadow detail from the Imacon

- but as you note, it takes some fairly aggressive pulling up of the shadows to demonstrate, and it's hardly a

massive difference. As to detail, however...

 

"Reason I don't use NikonScan is because I can't figure out how to get a color-managed workflow from it ...

 

...Am I wrong in this assertion?"

 

Nope, you're spot on. And Les - if you're going to berate people for ignoring provable facts, this is one you

consistently choose to ignore. NikonScan is not and simply CANNOT BE color accurate. Color accuracy doesn't

come from the hardware, it comes from a combination of hardware, software and reference to a standard, and

NikonScan simply doesn't provide that reference. You may find that it serves your purposes and works well for

you, and is the easiest, fastest route to good scans for you, and I've no argument with that - I would never

argue against anyone's personal preferences. But your continuous claims of scans from NikonScan being "neutral"

or color accurate are simply wrong, and display a remarkable lack of understanding of the processes involved in

digital imaging. Sometimes NikonScan gets it pretty much right, sometimes it gets it a bit wrong and sometimes

egregiously wrong - just as you'd expect from any uncalibrated, "autocolor" type system.

 

I'm sure you've read the many discussions on color with Vuescan, and should know that the included profiles are

pretty much useless. That's not how we make good color negative scans in Vuescan. And since Rishi is

referencing positives in a calibrated workflow (something he very much has right), try repeating the experiment

with a well-profiled, properly set up Vuescan. Or better yet, do the same thing in Silverfast with IT8

calibration. You've berated Rishi for demonstrating a lack of experience in his examples, but all you do here is

show your own lack of experience in using Vuescan.

 

As for the personal attack, these "looney tunes" comments are absolutely beneath you.

 

And with that, I'll end my participation here. When these kinds of personal attacks start showing up, the

discussion becomes pretty much useless.

 

Good luck to you Rishi,

 

Scott

Thanks Scott, for your help & for your comments. That's too bad we're losing you over something so inane & trivial, but I understand your point & position.

 

I'll report back with new scans tomorrow, as I have access to the facility from 10-6 :)

 

Cheers,

Rishi

Rishi,

 

Just to be clear, I'm not ending my occasional participation in photo.net discussions - I'm just going to let this particular one go for now.

 

Take care,

 

Scott

<p>I'm not about to get drawn back into this discussion, but that doesn't mean that I'm going to allow my

thoughts to be so grossly misrepresented. This quote is both wildly out of context AND selectively edited in

order to completely twist my meaning.</p>

 

<p>Here's the full sentence I originally wrote: "NikonScan may well be the best application on earth when set to

defaults - <i>I don't know and don't really care to argue such a point, as it's still not nearly good

enough."</i></p>

 

<p>Good grief.</p>

 

<p>Scott</p>

Hahaha! I think the mods should consider slapping a warning on this Les character. Clearly I'm not the only one

that has a problem with him.

<p>

And by the way, Les stating his experience of 8,000 scans and never experiencing what so many others have

experienced (colors just turning out wrong) and, at this point, clearly just <i>ignoring</i> evidence from others

that contradicts his perfect little world of Coolscan/Nikonscan, reminds me of something... what could it be?

<p>

Oh, yes, I remember now:

<p>

<img src="http://staff.washington.edu/rjsanyal/science_vs_faith.jpg" width=800>

<p>

Need I spell out who's on the left side of the diagram and who's on the right?

<p>

Rishi

OK, the version of Vuescan I was using kept relocating the autofocus point (I think). That was the problem.

 

I'm actually ashamed at myself for not double-checking my results before posting here. I apologize profusely. I have scanned on the LS-9000 before, but always noticed that it was less sharp than the LS-5000 or Imacon, and this was months ago. When I returned to scanning and got these dismal results, I just chalked it up to what I remembered before: that the scans were typically less sharp than the LS-5000 or Imacon. But I didn't double-check against the sharpness of some of my previous LS-9000 scans.

 

Bad form on my part, really bad form.

 

I will post back with some new 1:1 crops & comparisons.

 

Can I petition to have this thread removed? I really don't want to feed misinformation. I'll correct it, but, still, I don't want people to stumble on this thread and just read the first post, for example :)

 

Shamefully,

Rishi

I gave up on Vuescan - too many of the patches created new problems.

"Advise you to start waking up on the right side of the bed. And take your pent-up anger from who knows what and your 'looney toon' name-calling elsewhere. Not that I can stop you, but I can tell you that you're not welcome in my threads until you learn how to civilly communicate. "

 

Sorry, Rishi, but this is photo.net's thread. If you want your own thread, you'll have to post it on your own site.

Robert, so I take it you use Silverfast? Or NikonScan?

 

I'm still having film flatness problems with the LS-9000... no surprise there. So my next test is to try wet mounting on the LS-9000 vs. the Imacon.

In my experience, wet scanning is absolutely necessary to achieve results on the LS9000 that can compare with Imacons. Film flatness

is the biggest shortcoming with the design of Nikon M/F scanners, and the rather ingenious design of the Imacons was expressly

developed to give the film the flatness of a drum scanner. It goes without saying that if your film plane is not flat, you are not going to

get professional scans, yet the designers at Nikon seemed to have treated that as an afterthought. Wet mounting negates that

shortcoming.

 

That said, wet scanning on the Nikon is a time consuming process which, if you happen to be scanning copious amounts of film, will in

itself drive you to spend the money for

an Imacon. I mostly use my Nikon now for website portfolio work. When scanning for a few, select, large format exhibition prints, I rent

time on an Imacon 949, which reveals superior shadow detail, but I've found that the Imacon lacks the near-highlight (5-10% value) detail of

the Nikon, and depending on the needs of the image, I occasionally I end up wet scanning on the Nikon instead. On 30 x 36 prints from

an Epson 9880, I still get grain sharpness that achieves or exceeds that of darkroom prints.

 

Has anyone noticed a similar advantage in highlight detail with Nikon 8000/9000 over Imacon? I noticed this when scanning Delta 400 and

Tmax. I can see detail on the neg but the Imacon won't see it and the Nikon will. My scan settings were flat in the highlight area and very

slightly exceeded the data on the histogram.

I make linear scans, as positives, using NikonScan. I make all of my corrections using Photoshop, particularly with the plug-ins by CF-Systems, ColorNeg does a nice job inverting negatives and color correcting. A rcent thread on this is here:

 

http://www.photo.net/digital-darkroom-forum/00MRuZ

 

As for film flatness, wet scanning fixes that. Or you could buy a Nikon glass holder. I use a sheet of ANR glass from Focal Point when I don't wet scan. It's cut to fit the Nikon film holder - simply place it on top of the film and close the holder. It works well.

I've heard the Nikon glass holder results in Newton rings. I tried the ANR glass from focal point &, no offense,

but I cannot understand how anyone would use it as it introduces serious artifacts, as I've shown here:

<p>

<img src="http://staff.washington.edu/rjsanyal/FH3_ANglass-Comparison1.jpg" width=800>

<a href="http://staff.washington.edu/rjsanyal/FH3_ANglass-Comparison1.jpg">Link to full-size image</a>

<p>

Above images are from a LS-4000, btw.

<p>

I'm resorting to trying fluid mounting from ScanScience, as your comparison has me pretty convinced. I may just

stick with the Imacon; I really don't know just yet.

<p>

Thanks,<br>Rishi

The difference in perceived sharpness in the first two images in the original post is purely a question of unsharp mask. I took each image in turn, pasted it into a new Photoshop CS2 image. Then I selected the left hand side (the Nikon scan) and applied a light unsharp mask. Low and behold the Nikon and Imacon images were indistinguishable. In fact they were so similar that they both showed the same sharpening artifacts. Proof positive that the Imacon is applying more unsharp mask than the Nikon.

 

I have heard that the Imacon are not the best pro scanners though I have no first hand experience so that might be twaddle.

 

Now maybe if you took the Nikon image and upsampled to 8000 DPI then compared the two images, there might be a difference, but I suspect not. Perhaps the real test is to photograph parts of the image with a compound microscope to determine whether or not there really is more detail to scrape off the emulsion i.e. is the scanner or the film the limiting factor.

Leif,

 

I had posted earlier that I wanted this thread removed -- b/c I've used a Nikon LS-9000 before, seen that it had less yet comparable resolving power in comparison to the Imacon with 35mm film, yet prematurely posted these results using a newer version of Vuescan that, for some strange reason, moved my autofocus point to the edge of the film.

 

So the softness is due to focus error.

 

I'll try and come back today to post the real results -- those that show that the Nikon LS-9000 is quite comparable to the Imacon.

 

Sorry for wasting your time man,

Rishi

Excellent. No problem.

Robert (Budding), did you compare wet scanning on the CS 9000 vs the Nikon glass holder?

<h1>Important Update!</h1>

<p>

Everyone who's partaken in this thread, I feel obligated to update you with some real results.

<p>

But first, let's play a little game... :)

<p>

For each side-by-side comparison below, guess which image is the Nikon LS-9000 scan and which one's the Imacon

848 scan... one is not necessarily consistently on the left or on the right; they can be mixed up from image to

image! Colors are equivalent between the two scanners because I've profiled both with a Hutch Color Target (gotta

love proper color management!). Note that the Imacon image was scanned @8000ppi, then downsized to 4000ppi using

a Bicubic (not 'bicubic sharper') algorithm. (Later on I'll post a comparison upsampling the LS-9000 to 8000ppi).

<i>Remember to follow the link the full-size image for a fair comparison at 100%</i>.

<p>

<b>Untouched scans, 1:1:</b><i> ok this one's easy</i><br>

<img src="http://staff.washington.edu/rjsanyal/Photography/LS-9000_vs_Imacon848/OriginalComparison.jpg" width=800>

<br>

<a href="http://staff.washington.edu/rjsanyal/Photography/LS-9000_vs_Imacon848/OriginalComparison.jpg">Link to

Full-Size Image</a>

<p>

<HR>

<p>

<b>After application of Unsharp Mask to the LS-9000 scan:</b><br>

<img src="http://staff.washington.edu/rjsanyal/Photography/LS-9000_vs_Imacon848/AfterUnsharpMask.jpg" width=800>

<br>

<a href="http://staff.washington.edu/rjsanyal/Photography/LS-9000_vs_Imacon848/AfterUnsharpMask.jpg">Link to

Full-Size Image</a>

<p>

<HR>

<p>

<b>After Unsharp Mask to LS-9000 scan, AND NeatImage application to both scans:</b><br>

<img src="http://staff.washington.edu/rjsanyal/Photography/LS-9000_vs_Imacon848/AfterUnsharpMask+NeatImage.jpg"

width=800>

<br>

<a

href="http://staff.washington.edu/rjsanyal/Photography/LS-9000_vs_Imacon848/AfterUnsharpMask+NeatImage.jpg">Link

to Full-Size Image</a>

<p>

<HR>

<p>

<b>Another portion of the image, after Unsharp Mask & Neat Image:</b><br>

<img src="http://staff.washington.edu/rjsanyal/Photography/LS-9000_vs_Imacon848/AfterUnsharpMask+NeatImage_2.jpg"

width=800>

<br>

<a

href="http://staff.washington.edu/rjsanyal/Photography/LS-9000_vs_Imacon848/AfterUnsharpMask+NeatImage_2.jpg">Link to

Full-Size Image</a>

<p>

<HR>

<p>

<b>A crop from the extreme left side of the frame, to check for edge sharpness:</b><br>

<img

src="http://staff.washington.edu/rjsanyal/Photography/LS-9000_vs_Imacon848/CornerFocus_AfterUnsharpMask+NeatImage.jpg"

width=800>

<br>

<a

href="http://staff.washington.edu/rjsanyal/Photography/LS-9000_vs_Imacon848/CornerFocus_AfterUnsharpMask+NeatImage.jpg">Link

to Full-Size Image</a>

<p>

If you get tired of guessing and just want the answers, then let me know :)

<p>

Suffice it to say the results are rather surprising...

<p>

Cheers,<br>Rishi

Nice results and yes, they're remarkably similar. any chance you can tell us which one is which? feel free to mail me if you don't want to post it here yet.

Looks like no one besides Alan wants to play my little game... so... here it goes:

 

For all images: LS-9000 on left, Imacon 848 on right.

 

Alan, you almost nailed it :)

 

Pretty surprising, huh?

 

I was particularly surprised by the fact that the Imacon DOESN'T nail extreme side focus! I use autofocus on the unit (for that matter, does anyone know how autofocus on scanners works? Is it contrast detection?), and so 90% of the image is tack sharp, but the extreme edges that are right near the edge of the magnet/metal turn out to be quite unsharp.

 

The LS-9000 held up pretty well after a few manual focus adjustments. But I still find it rather ludicrous to manually adjust focus until you get it right for every scan. Which pretty much means I'm gonna have to go the fluid mounting route.

 

Also, I'm repeating this test with a steady image NOT taken from a bridge. Will post those results soon. Additionally, the 8000ppi Imacon DOES resolve detail that the 4000ppi scan does not, which is apparent when one enlarges the LS-9000 scan to 8000ppi. However, how much of that is USEFUL information is up to debate.

 

Rishi