the finer lines, the ones that look like scratches, are probably either dust (bulb or ccd), or just the Imacon trying to dig out detail from dense areas. I have this problem a lot with my cross processed film, that highlights may end up with odd magenta streaks like that.

The green band is something i experienced with a machine at college also. It could be any variety of problems, but the main thing is that it is not you, and the technician should understand that he isn't doing his job, and the Imacon needs fixing. The techs at my college were a bunch of jackasses and cared as much about the Imacon as they did about me. Due to the width of the band, it could also be a problem with the gearbox.

Anyway, your best bet for resolving the fine streaking is through balancing the exposure of the scan, so scan a full res 3F file, and work on it from there (which you can do on any machine that you install Flexcolor on).

t

could it be a bad connection? a loose wire? happend to me once ..just carefully double check both connection and>or change the wire between the scanner and the mac. a scsi wire could be defective. if you plug it using a firewire i would suggest to try a scssi instead, i use to get those when in firewire with the stupid little black box supplied to connect it.

also, you could try resting the imacon to its default setting and run a calibration using the supplied target.

I would point out that these are all issues that the lab technician should deal with, and not you, and simply because you shouldn't have to deal with the hassle of a service provided which is not maintained correctly. If you owned it, the situation would be different. Fundamentally, you are not doing anything wrong.

t

Patrick: there is no black box with the Imacon 848... it just has a Firewire 400 port on the back of the machine. I'll see if I can get the technician to try a different cable.

T Feltus: you're right, but I just find that I have limited bargaining power here, as the lab is pretty much providing this service to us for free. That being said, I do agree that if you own a piece of equipment that is worth thousands of dollars, you may as well spend a little time & money to upkeep it.

Question though: you're saying you sometimes see these green or magenta lines in your own scans? Is that also with an Imacon? Because I've never in my life seen these artifacts with either a Nikon LS-4000, LS-5000, or LS-9000. So it can't be a limitation of the technology... right?

Thanks, Rishi

oups..sorry didtn get the 848 model..just read Imacon : )

as for the green line, i dont know for T, but i only saw this problem personnaly when the cable or the connection was bad. It not normal when a 10k scanner (maybe less with the 848 im sure but still) does *problem* like that.

yes, though mine are usually magenta. they seem to occur when the light is not powerful enough to get through the dark parts of the neg or tranny. and yes, mine is with my Precision II, a bit older, but not much difference in the hardware. they generally are on one channel, so i have spotted some out with the spotting brush in PS. or you balance exposure to eliminate them. those areas are probably meant to be black, as mine are meant to be white.

the service may be free, but remember, they are paid to do that, and you aren't paid to use it, so they ought to do their job. talk to their boss.

t

Oh, also..free or not a lab should take care of there equipment..theres no reason to have a bad scanner other than telling you loud and clear that you should move to digital and stop bothering them..is that what they want to tell you? maybe?!

Haha, Patrick, that's actually what I want to tell myself. Yet somehow I still found myself shooting 12 rolls of Velvia these past 2 weekends which cost me $200 in film + developing. If I do the math, my use of film over just the past year, I'm sure, probably already exceeds the cost of a 5D. Cognitive dissonance, I guess, allows me to still shoot film.

But either way that doesn't take care of the problem of what to do with my hundreds (thousands?) of exposures on film lying around!

T Feltus: you say that 'those areas are probably meant to be black'... but if you look at the 100% crop, the green lines still run across regions that are certainly not meant to be black; i.e., the shinier portions of the rock. Also, if you're scanning negs and you get magenta lines, that makes sense, since magenta is opposite green on the color wheel & you're speaking of neg film which gets inverted. But note also that I have some magenta bands, especially in the lower center portion of the 100% crop.

I may forward this thread on to the supervisor of the lab.

Thanks guys, Rishi

yea, I know what you mean. my highlights that have streaks still have detail in them too. and no, i have not yet found the solution. we probably both exposed slightly off for Imacon's taste.

the 5d may be a "good" camera, but my 6 year old Imacon gives me around 47 megapixel from a 6x6 negative. Even if that were the only reason, I prefer film. and I get to keep an original. and film has a better dynamic range than digital. an argument can be made, but i have my stance.

last year, before i had the Imacon, at the lab we worked with a series of negs of mine, on a 949. we could not figure out how to get rid of the streaks, though my instinct was to blowout to white, the tech was not keen. solution? oddly, I showed 6 images scanned on a Frontier minilab, as for some reason its larger sensor and LED light source was capable of avoiding this problem, thugh the inherent sharpening in the software does mean that there is more of a separation in high contrast areas, than the Imacon would have produced. Yes, the Fuji Frontier costs about 5 times the Imacon, and is operated by a GameBoy interface.

t

As far as I know, I think you're right that negatives have a larger latitude than digital, as a white paper by Canon states, and I quote:

"Canon’s full-frame sensors have reached another image quality milestone as well. Their gradations and dynamic range are now the equal of the best positive films, and their resolution and lack of grain are superior. No smaller sensor has achieved this level of performance."

Since negative film has much more latitude than positive film, we can conclude then that neg's latitude is considerably higher than that of current CMOS chips.

I'm surprised that you also had this problem on an Imacon and were not able to get rid of it. I can't believe that all Imacons have this problem, right? Like I said, the Nikons do not display this problem at all, and still get the shadow information out of the shadows... and these cost 10 times less than the Imacon. What's going on?!

-Rishi

These streaks are horrendous and look like scanning artifacts- nothing on the film. I occasionally see similar artifacts in shadows with my CCD based scanner.

Is there a calibration feature?

i also want to add that if the canner is close to a electrical box, or a huge wire in the wall that can also cause the banding effect, or the smaller line..

i know, iknow it sound esoteric..but i talk from my own experience?! if all fail..it doestn cost much to replace the wire and move the scanner some inch from the wall : )

Another neat trick; because of the scnning vibration that the Imacon sometime could cause or the vibration due to someone walking by or by a car passing by, i use to put my Imacon on rubber feet, the same kind thta DJ put there turntable on to minimize the effect of the crowd dancing...that also can cause the head to *shake* and give out of registration color.

So heres a resume of the thing i would check;

1_cable

2_put it further from the wall

3_get some ruber feet

Also, i dont think it will be really necesary, but ask them to run a calibration with the imacon target..it take 3min..why dont just do it : )

i like the suggestion of rubber feet. i am tempted. sadly my workspace is not large enough to move the scanner from the wall, but I also am pretty confident that there is nothing in the wall at this part of the room. British electricians have not yet discovered the joys of running wire inside flexible conduits.

Let me clarify, I had the big green band in your first image, on a Precision II at Edinburgh College of Art. It would happen when i tried to change the ouput size of the scan, and seemed to disappear if i quit and restarted the scanner, or rebooted, i don't recall, it was 4 years ago. The CCD noise is something that I had when working with the pro lab, on their 949 (i dont recall if i also had some done on their previous 848), and now also on my own Precision II (which ironically is the newest of the three by far). It may be an inherent defect in the Imacon scanners, when they reach a film density that was not deemed important.

As for the Nikon scanner, I have not used them, but I would assume that the film plane is not very flat. I always assumed the Imacon was the best, in the same way that Leica is best and Hasselblad is best. Quirky but with the best results.

um... as for what Canon says... I will believe it when someone proves it to me. A colour tranny has a better tonal range than 16bit, and a well exposed black and white neg has a dynamic range nearer 64bit, in my opinion. There was a show here in Edinburgh last year of Ansel Adams archive prints, mostly small, but masterful. and for some reason they put a mini show of a scottish landscape photographer in there, and they were big prints (C-Dip digital intermediate using Imacon and Lamda) from 6x7 (mamyia, i believe), and the tonal range was terribly embarrassing. they were, theoretically, cutting edge-ish, and yet they were no where near prints made around 70 years earlier. For photojournalism, not Cartier-Bresson or Michael Subotzky style, digital makes sense. snapshots, film is a waste. editorial, it is debateable. nudes, film... but landscape? of course it has to be film. an 8x10 negative will have a resolution that digital will not be alble to reach for at least another ten years, much less its dynamic range that would make an elephant blush.

t

not want to start a debate here, but im seing zillion of image a years due to my work, from film to digital for the past 16years...

i could say that film was good, so was slide in need. i have scan them for years, hundreds per week for a image bank when i was younger fresh out of school : ) (im 34 now, being in that biseness since i was 18) from a agfa duoscan, to a horizon to a imacon..

I could say that i have seen a lot of s*** coming from digital and film, but more from digital because people had to learn a new way of exposure and learn all about the new medium..and at first a lot of people where not really inclined to do so..even today.

i have in my hand file from canon 5d, canon mark IIIds and bigger P45, and i could say that i can see major improvement, and amazing details and color in any of them..given its a 13, 23 and 39 million pixel camera. Well exposed, well developed, well print i could say i think digital is closer day after day if not superior compare to film in many situation, when well execute by someone who now is stuff; i have seen bw print at some epson print academy that where greenish..yet they have the best printer and technician on site. i have seen bad digital print and bad tradional print as well in gallerie and exposition..but i have also seen a lot of good one.

I think its a new medium that people need to understand like when photography start and painter where blasting it because anyone with the press of a button can capture the landscape : )..

like anything old habit is hard to get by, and i undertsand people who prefer film, but you should get good result with film or digital if in both case all is well controlled.

some like film, some like digital, i had the chance to use both for years and i prefer digital for all my job, and prefer to be sit in front of a monitor than in my darkroom. I also prefer using Ligthroom than scanning zillion of neg and slide..but i miss the beauty of a 6x6 or a 4x5 well exposed slide on a lightbox ; )

just a open oppinion, a point of view only.

T Feltus: 'a dynamic range near 64 bit'? That doesn't make sense -- the bit depth determines tonal transitions (how many steps in between discrete tonal steps), not the overall range.

Realistically, in my own experience I feel like digital has a better latitude than Velvia simply because of Velvia's ridiculous exponential response to light. Think of an exponential curve... if the exposure is too low, you're on the bottom end of that curve where the slope is really low & you just don't get much silver halide reduction in response to light... and the whole image turns out dark. Overexpose and you're on the other end of the curve where the slope is much higher than 1... too much silver halide reduction & you start losing a lot of dye, thereby desaturating and blowing out highlights.

Digital capture, as far as I know, has a linear response to light. Better than exponential. Not as good as negative's logarithmic response. However, as for digital, any time I try to brighten shadows on Canon CMOS captures, I start seeing banding, presumably due to differential gain from the amplifiers per row (see the thread here: http://photo.net/canon-eos-digital-camera-forum/00RBTe). This reinforces the philosophy of 'exposing to the right' when it comes to digital capture. Big problem, that'll probably, however, become moot as CMOS sensor sensitivities increase... i.e. I'm really curious to see if this'll be a problem on the 5D Mark II.

As for scanners, we shouldn't get this banding in the shadows because there aren't separate amplifiers applying gain... it's the same row of the CCD (or same 3 rows of the CCD, what have you). As long as there's no temporal variance, we should not see banding in the shadows. This is truly unacceptable from the Imacon. The Nikon LS-9000 is very clean in the shadows. The LS-5000 is pretty clean, but DOES exhibit some strange patterns in deep shadows that can be subdued by multi-sampling when scanning slides... still, nowhere near as bothersome as the green lines by the Imacon.

Patrick, yeah, don't really want to continue this debate about film vs. digital here either b/c each have their own hallmarks. Yeah medium format & large format film are phenomenal; but as for 35mm, digital has surpassed the resolution of film. On the other hand, it's not JUST about resolution. The color response of the dyes in Velvia are what I'll call 'all-inclusive'... i.e. the dyes have broad spectral responses. But look at the transmission spectra of the R, G, and B filters on Bayer pattern CCDs... they're pathetic spikes across the spectrum. Interpolation, however, weeds out all the colors in between. Whether one can actually see the difference is up to debate. My bigger problem with RAW capture is different RAW converter's interpretation of the color information in the RAW file... Lightroom's conversion of all Canon RAW formats is absolutely unacceptable.

At least when I scan a Velvia slide with a color-profiled scanner, what I see on the slide is what I get on my calibrated monitor. I love that about film. Of course, I have yet to adopt a scanner/scanning method that I'm completely satisfied with, but that's another story/problem that I'm trying to solve by this very thread :)

-Rishi

Shyly, yea, I know, I exagerate a wee bit. and the tangent should not even be touched, as it is an argument with no right or wrong. For my purposes a Hasselblad 500c and Precision II yield much better results than the equivalent digital camera i could have bought for the same amount. I also like the machine. I don't like machines that insist on doing stuff I don't want to do.

And Velvia, good for landscapes, not good for portraits. With film you could know what you needed and how to get it. you learned. the pricing of digital is not really permissive. you cannot have one dSLR system for landscapes, and another for portraits, and another for wedding. I kind of like Erwin Olaf's approach of shooting on film and then going mad in photoshop. at least, in an odd manner, he is faithful to both worlds.

Ja, dynamic range is what it is. From what I can guage with my eye, though numbers are not my forte, a print from a good black and white neg onto even a multigrade rc paper has about twice as wide a range of greys as the same neg scanned 16bit on my Precision II. The bromide print can contain detail in black, and still have grain in white, without blocking up. And Ansel Adams was, essentially, doing optochemical HDR with his Zone system, and his prints had full tonal range from billowing cumulous clouds to the needles on the ground under a pine tree, which must, almost, be doubling again the range of captured tonalities. Ah, yes, this may be different to how many it can display, but it still may have captured the full 10 zones. The thing is that Adams was a technician and scientist in his art, and i am not. I see it very much in a subjective manner. I feel that I can get more pleasing results with older equipment. I like big viewfinders. i don't like batteries.

But as I said, I like cross processing my film. so I f*ck the whole argument of perfection, and I am not good at keeping track of my settings, so often i end up losing notes and getting dense negs. Patrick is lucky. he gets to make other people's work better, and has a system.

One think, to wrap this up, as we are far from the scanning straks which we cannot resolve and are unacceptable (shall we all write to Imacon together?), which I found interesting, was that 10 days ago was the first time I went into a larger scale shoot with a polaroid back. I know, i am slow. Anyway, I did my test shots on the dSLR, and exposed a pola off the hassie. The thing i found really interesting is how differently people respond to a small sticky print, as compared with the display on the back of a camera. They really cherish the print, and see it as much more trustworthy, for some reason. the screen, they dismiss. Bah. But I think I like having some degree of unknown. can it, and see what happens when you get back to the lab.

t

T, what can i say..im a lucky guy : )

im glad to have cross your path also..didtn have the chance to work on your image, but like what i have previously seen..far from what i do, but artisticly really interesting.

Ah, Patrick, you are lucky. and don't move over here yet, as it seems that nepotism and long term work relationships rule supreme in Europe.

I am still on a slow learning curve with the Imacon, but I try my best. and I am sure that loads of people envy me, though I see nothing enviable in my plight.

t

The lines look like what I get with my Phase One 4x5 scan back when its got crap on its scan bar.

Here is a Lenox Cascade W-319 Dinner plate; shot with a 127mm F4.7 Ektar on a Speed Graphic with a Phase One 35 M pixel 4x5 back; using Halogen tungsten bulbs and a IR Filter. The back had not been used for awhile; this one of the first scans has the dirty scan bar; not cleaned off just yet.:

Kelly,

Thanks. Interesting. These bands are of much wider pixel width than what I see. But very similar otherwise. Except that they actually vary in color across the scan.

So how did you get rid of these? Cleaning the 'scan bar'? How? Also, I don't even know what a scan bar is b/c I've never used a digital back. Google didn't even help. Please fill me in :)

Thanks, Rishi

Hi Rishi.The scan back is sort of like a flat bed scanner; the image is scanned maybe in a minutes; a scan bar moves that has a row of sensors; a lead screw moves he bar across the film plane.. One places the scan back into the 4x5 camera just like a 4x5 film holder. One focuses on the ground glass; then inserts the digital holder; then does a scan. There is a scan bar; ie as sensor array that that moves and does one scan line at a time;the device connects to a computer like a scanner. .One gets this banding on the camera type scan back if there is dirt; dust etc on the scan bar. Flatbeds also can show lines like this too. Once the sensor's bar is brushed or vacuumed off the bands are gone.The larger width bands are from a bigger chunk or dust/dirt. A small piece can make one only a pixel wide. I think my phase one back has a array with a sensor for Red, blue and green;and if the dirt/crud/dust is covering some and not others the band lookslike a different color.

Kelly,

Thanks for the explanation. Just to make sure I understand: the scan back is like trying to make a digital camera out of a film camera? I.e. there is no film involved at all; you just set up your shot... and... I just read an article on it & can't seem to understand how it emulates a shutter speed.

I can buy that the small thin 1-2 pixel wide lines are dust. But the big green band? Or how some images just came back half inverted (like a negative)? I can't attribute that to a bigger chunk of dust/dirt. That's gotta be something data-related, no? I.e. communication error of software error?

Rishi

Oh boy, and the problems with this Imacon continue.

Now there seems to be an issue of optical flare and/or ghosting. Which I noticed before just didn't pay much attention to. I just decided to investigate it further, and compare to the Nikon LS-9000, and here are the dismal Imacon results:

Link to Full-Size Image

Follow the link to the full size image to get a better view. There are strange artifacts that leak around the squares, as well as the arrows and numbers, in the Imacon scan. In the 'real world' example, note the right-most tree, that has light leaking onto it from the yellow light from the road.

Has anyone seen anything like this before on an Imacon?

On the other hand... jeez look at the resolution of the Imacon. This is the corner of the frame, so the LS-9000 could do better with fluid mounting, sure, but... a huge difference.

Thanks,
Rishi

Here's another strange problem -- look at the bottom of this image, and notice the strange colorful bands.

Link to Full-Size Image

And here's an example of what I'm more used to seeing -- some thin banding mostly in the shadows (images have had shadows enhanced using Shadows/Highlights tool in Photoshop):

Link to Full-Size Image

And another one...

Link to Full-Size Image

right, the last two scans are what I am used to when the neg or trannie is a bit too dense for the scanner, whether this is a defect or not, it seems to be a problem that the Imacons have, and it seems to not change between the older models or the more recent ones. hardware is similar. The thicker green lines above look like they might be dust, either on the sensor or the bulb. but don't dust it yourself, it is not worth the risk, and is the duty of the technician, or the Danes.

in the top image, both scanners look out of focus, and the flare does not look optical to me. it looks like it is too regular to be optical, so i would assume it is a digital artifact on the +50 shadow detail.

i dont know what Pat thinks, however over here in Edinburgh we process the 3f files in Flexcolor, as poor a software as it may be. and the "brightness" slider seems to be something one does not touch, as it artifacts.

t

Rishi; Re Thanks for the explanation. Just to make sure I understand: the scan back is like trying to make a digital camera out of a film camera? I.e. there is no film involved at all; you just set up your shot... and... I just read an article on it & can't seem to understand how it emulates a shutter speed.

The scan back of a decade ago transforms a film camera into a digital camera; these came out by several makers in the 1992 to 1996 era. Typically they were/are used by repro/print houses to "shoot" artwork; old posters; or by advertising agencies to shoot clothing and still life; it products in boxes etc. Phase One and Better Light are two well known brand names. One focus ones 4x5" camera with the ground glass;then inserts the scan back in place instead of a film holder. The scan back is tethered to a computer via a cable. There is a *shutter speed* associated with the scan back; just like there is a shutter speed associated with a narrow slit on a moving focal plane shutter. With the scan back the *slit* is but one line; or RGB bank; and its takes say several minutes to go accross the field of view at the focal plane. The same thing happens with a film scanner; often there is a mirror that pivots; scan bar that moves etc that scans one thin region at a time. The 4x5 scan back IS MAKING a digital camera out of a film camera; its just that its 1995 technology with a 35 megapixel scan back; or a 1997 one with a 50 megapixel scan back. Usages of thes old scan backs involves alot of playing around. When they came out they cost more than a new car; thus were not something most amateurs bought or used. My point about dirt on the scan bar area is that it does cause weird scan lines with different colors; whether its on an old scan back; or flatbed scanner; or even film scanner.

Here is a page on the manuals for the obsolete scan backs such as the Phase One power phase and Photo phase units

These scan backs are so old that the recomended CPU was a Pentium 166Mhz. Here I have done alot with a 200Mhz Pentium Pro with a motherboard thats capped at 512megs; a computer that in its days cost us 2600; with only 32 megs then.

Here is a screen capture of the digital backs scan software. One sets the iso and exposure and shutter speed. One does a few pre scans to not blow the highlights; also one tunes the focus spot on too. The CPU requirements are minimal; a 200Mhz cpu is loafing around during a scan. Its a scsi device; one typcially sets up the back adn lets the computer find the scsi device. With a PPro board "souped" up with dual 333 Mhz server CPU's and the same 512 meg of EDO 66Mhz ram; one can rotate a 35megapixel 105 meg image 90 degrees in 7seconds with two cpus; and in 9 seconds with one cpu.



Since the *shutter* with the scan back is the scan itself; one can just use barrel lenses like an enlarging lens to shoot closeups like artwork on a speed graphic.


Or one can place an old 1960's process lens like a 150mm F9 Apo-Ronar and use it too; like this scan with the old 1996 35 megapixel scan back on a early 1950's speed graphic.In this shot there are NO "banding and scan lines" becuase the scan bar on teh scanner was clean!

"in the top image, both scanners look out of focus, and the flare does not look optical to me. it looks like it is too regular to be optical, so i would assume it is a digital artifact on the +50 shadow detail. "

It was there before the +50 shadow detail to my naked eye. I just wanted to make sure everyone saw it easily :)

Plus, keep in mind that the Nikon scan was also subject to such +50 shadow treatment, yet shows no sign whatsoever of this artifact. Also, remember that the shadow adjust was done in Adobe Photoshop using its 'Shadows/Highlights' tool.

As for this just showing up in dense areas of film with the Imacon -- that just doesn't really seem to be 'acceptable' to me when the Nikon LS-9000 can give me quite a good deal of shadow detail (on par with the Imacon) without any hint of this!

Thanks, and have yet to read Kelly's post.

Cheers,
Rishi

i still think that the flare looks digital. i think Imacon/Hasselblad needs to service the machine. have you checked its service status? the Imacon debug panel thingy tells you whether it is up for a service. my guess is that it is.

I would suggest that you just use the Nikon, if it is giving you better results.

t

Yup, it's been flashing a message saying it needs to be serviced for half a year now at least, to my knowledge. Still trying to get someone to do something about it. No luck yet.

Sure, I'd use the Nikon -- if I could the film flatness problem solved. Which involves wet mounting... which I haven't had time or money to try just yet.

Also, the Imacon still does resolve a little more at 8000ppi than the Nikon at 4000ppi, as you can see above :) Rishi

well, let's assume that Imacon's service message has some truth to it.

Also, remember that wet mounting does damage your original, limiting your chances of a good second scan.

good luck, t

Really? Even the ScanScience kit? They say their fluid is harmless & just evaporates off...

Ah, I have no idea. Each trannie i have seen that has been wet mouunted has suffered. Granted, most of these were probably from a drum scanner, but the guys at the lab i use have the same instinct. try it on a test neg that is not too important.

t

I just had a look at the manual for my Precision II, and found a couple of things that I shall test later on.

Fringing: it states that it is due to the unsharp mask. I believe its minimum setting is -100, so worth a try.

Noisy dense areas of the film: indicates that this could be the unsharp, but also to try using "adaptive light" at 50 in the Setup, rather than trying to compensate in the histogram.

Also, it notes that the 35mm neg carrier has two positions with different qualities. i had often wondered. 24x36 produces a higher pixel density, however less light passes through the source as with the 36x24, which they clearly state is better for denser sources.

soon.

I compared Dmax performance of Precision II vs. 848, and was appalled by the Precision II performance. Of course, I did use it in the 'high precision' mode (6300ppi for 35mm film). It hardly extracted any information out of the shadows of dense slides. The 848 extracted what appeared to me to be orders of magnitude more information.

But, then again, the light microscope I viewed the same slide under seemed to have extracted orders of magnitude more information out of dense areas of the slide than the 848. So go figure...

Clearly even the Imacon 848 isn't good enough. I haven't tried the X5 though... not that I've heard that it does much better (Dmax of 4.9 vs 4.6 of 848? And these are arbitrary anyway, since the Dmax of Velvia is technically supposed to be 3.9 but the 848 clearly doesn't resolve all the detail that's there in the deep shadows of my slides, as is evidenced by the details I see in the same areas of the dense slide under a light microscope).

OK, so after 'CCD calibration' using a gray slide, I got better results... but I need to confirm this over many scans before I get my hopes up...

But here you go:

Link to Full-Size Image
I don't know why the blacks are darker on the left. I used the same profile created by LPROF and an IT8 target... perhaps the Shadow/Highlights tool in Photoshop is a little whack.

Regardless, there's hardly any banding after calibration.

The bigger question is: when the $%#! is FlexColor gonna integrate multi-sampling? These sorts of green dots are visible on ANY CCD scanner in shadows. Multi-sampling on the Nikon & Minoltas totally cleans up shadows... Imacon, WTF?!

Can any of you hack FlexColor to multi-sample? Ed Hamrick? Support the Imacons? This is a crime, to not have multi-sampling on the Imacons!

Now that this problem seems to be solved, for the time being, back to the optical flare problem:

So... this flare is NORMAL (perpendicular) to the axis of the CCD... in other words, in the direction the film is moving... how is this possible ? I've never seen flare in this direction... as it indicates some sort of temporal , not spatial , flare...

Ideas, anyone?

Thanks,
Rishi

Rishi,

Regards multi-sampling with Imacon, there is a parameter called "adaptive light" in the setup file in Flexcolor. My understanding of this function, and feel free to correct me if I've got this wrong, is that it alters the integration time. The bigger the number the longer the integration. Same as multi-sampling really. The other way to approach it is to scan the image more than once and combine the images in Photoshop. I've tested this for up to 8 scans and you do get a slight improvement in shadow noise with this technique.

Is that last image, the one with the flare, from an Imacon or another scanner?

Here's an example of 8x multi-sampling with Imacon 646. These are full res crops from a 5000dpi scan of 35mm Kodachrome 64, on the left is 1x, on the right is 8x.

Link to fullsize image:
http://www.lexharris.net/documents/combined_full.jpg

By the way, please view my images above at 100% to see the banding before CCD calibration.

Also, yes, the optical flare or whatever that is is on the Imacon 848. Not along the CCD axis. Very strange.

As for multi-sampling, Lex, that's entirely different from adaptive light (which just increases the overall exposure, blowing out highlights and such).

Also, your method of multi-sampling above, while reducing some of the noise in the shadows, is a very bad idea in the sense that it potentially softens the rest of your image... which it has -- look at the zipper on the jacket. This is because this sort of method involves multiple image alignment.

That's not how multi-sampling on the Nikons work -- these scanners multisample 'on the spot', rather than re-scan the image several times.

I'm sure Imacon could implement it in their software, if they weren't so hung up over increasing speed of the scan...

Rishi

I knew I spoke too soon. The thin green lines are back in the Imacon.

But it shows up after a few scans... so I'm thinking it has something to do with increased noise after the sensor heats up a bit? I could try calibrating again now, after it's hot...

Like I said, I hate the Imacon. My LS-4000 is so much more reliable right now, and that's just the epitome of 'sad'.
Rishi

Rishi, true, adaptive light can blow highlights if you go too far, but the right amount does seem to pick out extra detail and stretch the histogram without pushing the black point up. About 50 works for me with positives, 0 for negatives.

Agree with registration issues with multiple images, definitely some softening and it takes forever too.

Below is Kodachrome 64 with histograms for different adaptive light in Flexcolor.

Lex,

Guess adaptive light works for limited tonal range slides... but I usually set whatever works for a HCT or IT8 target slide, which typically has *almost* the max tonal range for a slide (I think... I hope!)... then scan the rest of my slides at the same setting to maintain an color-managed workflow.

If I were to use adaptive light for different slides... I'd have to redo the color profiles with those adaptive light settings. In a sense, I see adaptive light kind of like 'RGB Exposure' in Vuescan... scan times get longer (right?)... so isn't it just increasing integration time? Likewise, in my color-managed workflow in Vuescan, I always use the same RGB Exposure setting for all slides that I used when scanning the color target (I let Vuescan select the auto-exposure for the target, then used that to make the profile and subsequently to scan all slides).

Rishi

P.S. want to jointly hammer FlexColor support asking them to add multi-sampling to the software? I don't see how it's hardware mod... seems like it could be done at each scanning line in software.