Luminous Landscape Forum
Raw & Post Processing, Printing => Digital Image Processing => Topic started by: PSA DC-9-30 on July 08, 2007, 09:35:13 am
-
I've been digging out 15+ years' worth of Kodachome (and some Ektachrome and Elite Chrome) slides, and I have been scanning a few on my Epson Perfection 4180 Flatbed scanner. It's not a high-end scanner, but it's far from a cheap piece of garbage either. Anyway, I know scanner resolutions is a whole other topic, but I've been scanning at 3200 dpi or higher. From what I've seen so far, the images look fine for on-screen viewing, and should be OK for making small prints (probably no larger than 5x7). I had one printed out at around 8x 12 and it looked grainy and not in the sharpest focus, despite a bit of unsharp mask applied in CS2.
I discussed the results with a couple people at my local camera store / lab and they said that Kodachrome is difficult to scan even in the best drum scanners and one will typically lose about 30% of the resolution from the original. They also told me that scanners have trouble capturing the different layers in a Kodachrome slide. The lab does offer 8- and 16- bit scanning on a Kodak Creo iQsmart3 Scanner. It's not cheap though--for a 3600 x 5600 scan (i.e., suitable for a 12 x 18" print at 300 dpi), the lab charges $26; for a 16 bit scan at that size, the price is $40. Obviously, I'm not going to have this done for every last one of my slides, just the very best ones that I think might have potential to sell, or to serve as prints to hang on my walls.
The questions are then:
1. For making large, very high quality prints, what is the best option for scanning Kodachrome--how does this Kodak scanner mentioned above rate? Should I be looking at a drum scanner instead?
2. Is 16- bit (vs 8- bit) going to be worth the money for my favorite scans?
3. What is the largest size I can realistically hope to make from a scanned Kodachrome slide? I was recently at the Fred Herzog exhibition at the Vancouver Art Gallery, and his original Kodachromes looked incredible in gliclee print form and some were quite large--maybe 24x36"!
4. Where do you get your Kodachrome scanned?
5. For slides that will not end up as large prints, I will continue to scan on my Epson. Assuming I'm using 3200+ dpi and saving as .tif with no compression, is there anything else I can do to improve quality??
Thanks for your help.
Kevin
-
Isn't is great to see old Kodachrome slides? The archival quality makes me wonder if all my backup drives and DVDs are really more efficient for long term storage.
1. For making large, very high quality prints, what is the best option for scanning Kodachrome--how does this Kodak scanner mentioned above rate? Should I be looking at a drum scanner instead?
Many people will say they get best results from their flatbed, but in side-by-side comparisons of a properly tuned flatbed scan to a properly tuned drum scan, even the most cynical skeptics will say the drum scan is better in every regard.
The current crop of drum scanners on the market have a top resolution of 8000 samples per inch (which becomes pixels per inch in PS) for all film formats. Over the years, I've used a UMax, Microtek, Epson, Imacon and finally settled on a Howtek (now under the name Aztek (http://www.aztek.com/premier.html)). The difference in the quality of the results will surprise you.
However, most people don't want to learn to operate a drum scanner. Mastering film mounting, mastering the software, keeping the unit clean and lubricated and buying supplies (overlays, fluids, tape, cleaners) tend to put people off.
2. Is 16- bit (vs 8- bit) going to be worth the money for my favorite scans?
Only if you plan to heavily edit the file in PS. If you correctly scan the image with accurate endpoints, grayscale gamma and color gamut then 8-bit scans will suffice.
If you send out for scans, be sure to tell them you don't want any post processing done. This will help assure that you get the best scan for your money. Most scanner operators will tweak a scan before they send it out.
3. What is the largest size I can realistically hope to make from a scanned Kodachrome slide? I was recently at the Fred Herzog exhibition at the Vancouver Art Gallery, and his original Kodachromes looked incredible in gliclee print form and some were quite large--maybe 24x36"!
An 8000 dpi scan of a 35mm slide gives you approximately 8000 ppi across the short edge of the film. When converted to 300 ppi, the resulting image is 26"x39". If you want to go larger, you'll need to interpolate and sharpen accordingly.
4. Where do you get your Kodachrome scanned?
With my Howtek.
I'll also point out that once you familiarize yourself with a drum scannner, the payoff is efficiency. You would be able to scan more pieces of film per day than you every could with a flatbed or Imacon. This is because mulitple pieces of film can be mounted onto a single drum. When I had the volume, I'd scan 30 slides per drum, three drums per 8-hour day. This included setting the scan parameters and copying the files to workstations, but not rotating, cropping or spotting.
5. For slides that will not end up as large prints, I will continue to scan on my Epson. Assuming I'm using 3200+ dpi and saving as .tif with no compression, is there anything else I can do to improve quality??
Master the software to maximize results. Knowledgeable operation of your scanner is the only way to get the best results. This means knowing what to set endpoints at, what to set the gray balance at, where to set your histogram gamma midpoints at, and having your scanner profiled for a good color workflow.
Good luck!
-
Flatbed scanners are not usually ideal for film scanning unless they have a dedicated film scanning attachment. Even then you may be better off using a dedicated film scanner, for example - a Nikon Coolscan 5000. The software you use also matters. Silverfast Studio Ai 6 is the best on the market for film scanning. With a good desktop film scanner and good software you should be able to produce very acceptable high resolution scans from Kodachrome slides. I've done it.
The only real issue with Kodachrome is the use of Digital Ice. Digital Ice is a piece of automated retouching software for eliminating crud from slides. I have Kodachromes that are 50 years old, full of crud, and eventhough it is not supposed to work, Digital Ice (packaged with my Minolta ScanElite 5400 - unfortunately discontinued, an excellent scanner) did "de-crud" the slides very nicely. Because the colour fades over time, colour balance can be restored easily at the scanning stage using Silverfast, or in Photoshop afterward using the grey balance or white balance controls in Curves.
You should not need a drum scanner to do this work unless you need more than 4000 dpi, you should get very good results from a good desktop film scanner and software, and the investment may be worthwhile if you plan to do a large number of these scans. If you do it yourself you also have complete control over the whole process.
-
I use a Nikon 8000 film scanner with great results using Silverfast which has a Kodachome setting. I also made a Kodachrome profile for the scanner with a target I purchased through B&H which I think is still available. The combination of Silverfast and the Kodachrome profile gives me that nice Kodachrome glow.
-
The questions are then:
1. For making large, very high quality prints, what is the best option for scanning Kodachrome--how does this Kodak scanner mentioned above rate? Should I be looking at a drum scanner instead?
2. Is 16- bit (vs 8- bit) going to be worth the money for my favorite scans?
3. What is the largest size I can realistically hope to make from a scanned Kodachrome slide? I was recently at the Fred Herzog exhibition at the Vancouver Art Gallery, and his original Kodachromes looked incredible in gliclee print form and some were quite large--maybe 24x36"!
4. Where do you get your Kodachrome scanned?
5. For slides that will not end up as large prints, I will continue to scan on my Epson. Assuming I'm using 3200+ dpi and saving as .tif with no compression, is there anything else I can do to improve quality??
Thanks for your help.
Kevin
[a href=\"index.php?act=findpost&pid=127116\"][{POST_SNAPBACK}][/a]
1) Realistically drum scans still provide the best image quality. But a desktop scanner such as a Nikon Coolscan comes pretty close. I guess your best approach is to purchase a scan from each using the same slide.
2) If they screw up the exposure, or you want to do contrast masking and/or HDR, then yes. Otherwise, no.
3) Easily A4. A3 prints look good, assuming KR64, and a desk top scanner should do that, but you might want to post-process to remove the graininess.
4) I scanned mine at home with a Minolta 5400. The main problem is profiling Kodachrome. I could not find a Q60 slide. But someone indicates that B&H sold them? The Minolta is no longer made so is not worth considering. The Nikon Coolscan range is excellent, and provides for automatic dust removal (does this work with Kodachrome?), grain removal, and correction for fade.
Another alternative is to use a DSLR to copy the slide. Place the slide on a light box, in an otherwise dark room, and photograph the slide, preferably with a macro lens. A 10MP camera easily provides enough detail for an A4 enlargment. The advantage of this is the simplicity and cheapness if you have the equipment to hand.
-
I'll throw this out here and see if it works. Been experimenting with perceptual curves combined with profile processing assigning 1.8 gamma profiles to high contrast saturated images using shadow/highlite tonal sliders in advanced layer blend mode to reduce percieved grainyness.
Produce a flat looking scan by setting absolute 5,5,5 black instead to around 30,30,30 or just lighten the black until you actually see a difference between the blackest black and an overall lightening of the entire image. Then once in PS apply a global contrast USM setting of around 20 Amount and a maxed out Radius and adjust depending on how flat the image looks combined with resolution.
Epson's editing tools aren't that great used in a gamma encoded environment. PS is much better, but you have to start out with a somewhat flat looking image.
Just a suggestion based on trial and error. Not sure if it'll work, but I do get beefier looking scans with less noise and grain from my 35mm negatives scanning in 16bit.
-
Much of this stuff is covered in my two articles below:
Near Digital Quality[a href=\"http://</a>http://www.luminous-landscape.com/tutorials/silverfast-scanning.shtml\" target=\"_blank\"]Silverfast (http://www.luminous-landscape.com/tutorials/ndq.shtml)
Mark
-
4) I scanned mine at home with a Minolta 5400. The main problem is profiling Kodachrome. I could not find a Q60 slide. But someone indicates that B&H sold them? The Minolta is no longer made so is not worth considering. The Nikon Coolscan range is excellent, and provides for automatic dust removal (does this work with Kodachrome?), grain removal, and correction for fade.
[a href=\"index.php?act=findpost&pid=127143\"][{POST_SNAPBACK}][/a]
I could be wrong, but I don't believe much purpose is served calibrating the scanner with a Q60 slide if the slides to be scanned have been subjected to fading.
I'm very impressed with the capabilities of my Dimage Scan 5400 II, which, as Mark mentioned, seems to be able to remove scratches through use of ICE with Kodachromes, something that other scanners, such as my Nikon 8000 cannot do without producing artifacts such as double edges.
I haven't been keeping up with developments in scanner technology but I wonder if there's now anything even better than the KM 5400 II at a similar affordable price.
-
Hi Ray,
I don't think the actual condition of the slides alters the basic arguments for calibrating and profiling the scanner, because once one corrects the images say in Silverfast before scanning, one wants the scan to begin a consistent colour-managed workflow. This site is of particular interest:
Hutchcolor (http://www.hutchcolor.com/HCT_overview.htm)
Like you, I'm basically very pleased with the Minolta 5400, so I haven't been looking around either. Sadly though, as you know, Minolta has abandoned this business so these models are no longer available. There's good post-capture software around for clean-up, so I think the key things to focus on for any one buying a new film scanner is the same old stuff: optical quality, maximum optical resolution and dynamic range. The rest can be handled with 3rd party software.
Mark
-
Hi Ray,
I don't think the actual condition of the slides alters the basic arguments for calibrating and profiling the scanner, because once one corrects the images say in Silverfast before scanning, one wants the scan to begin a consistent colour-managed workflow. This site is of particular interest:
Hutchcolor (http://www.hutchcolor.com/HCT_overview.htm)
[a href=\"index.php?act=findpost&pid=127188\"][{POST_SNAPBACK}][/a]
Hi Mark,
I can't quite grasp what's going on here. I thought that the purpose of the target was to calibrate the scanner to a certain color scheme as one might calibrate one's camera by making hue and saturation ajustments in ACR in relation to a GM color chart. Having made such adjustments for a particular model of camera, and even a specific copy of that model of camera, such calibration is not effective for another model of camera and perhaps not even effective for another copy of the same model which has its own peculiarities.
On my uncalibrated 5400, I can make any color adjustments I think are appropriate, either in Silverfast or Vuescan, and the result I see in the preview before scanning is very, very close to what I see when I later open the scan in Photoshop and compare with the preview still on the screen.
As far as I understand, my scanner is very consistent in its uncalibrated state, but would not be ideal for scanning recently taken slides.
The Hutchcolor site you refer to makes no mention of benefits for faded slides. The examples relate mostly to underexposed slides.
-
Ray, best is for you to read chapter 4 of Andrew Rodney's "Color Management for Photographers" where he explains the purposes and process for profiling a scanner. Once your own scanner is calibrated and profiled to an IT8 or some such target (useful only for positives, not negatives) it is colour-managed and should give predictable results whether you scan under-exposed or faded positives that you have adjusted with scanner software. For negatives the best we have is Silverfast's Negafix, which is a work-around to profiling.
-
Ray, best is for you to read chapter 4 of Andrew Rodney's "Color Management for Photographers" where he explains the purposes and process for profiling a scanner. Once your own scanner is calibrated and profiled to an IT8 or some such target (useful only for positives, not negatives) it is colour-managed and should give predictable results whether you scan under-exposed or faded positives that you have adjusted with scanner software. For negatives the best we have is Silverfast's Negafix, which is a work-around to profiling.
[a href=\"index.php?act=findpost&pid=127205\"][{POST_SNAPBACK}][/a]
Mark,
I don't seem to have a problem scanning old slides. I can't remember the exact color so I make whatever adjustments look credible. Here's a 43 year old Kodachrome of a couple of Tibetan refugee monks who'd fled from the Chinese into Nepal in the late 50's and were settled in a refugee camp at Trisuli near Kathmandu ..... scanned a year or so ago on an uncalibrated Dimage 5400 II.
[attachment=2779:attachment]
-
Hi Ray,
Good image with believable colours. Can't argue with what works. It's obvious your set-up is delivering what you need without further ado, and in those circumstances of course one leaves well enough alone. However, this did not HAVE to be the case. There is no guarantee that every scanner will do likewise - in fact there is a guarantee that not every scanner will - hence the general case for colour management solutions.
-
Mark
Will Silverfast SE work as well? Or do I need to upgrade to Ai6?
-
Kenneth,
Check here:
Silverfast comparison (http://www.silverfast.com/show/compare-ai-vs-se/en.html)
and see whether SE has sufficient features for your requirements. I think one of the most fundamental differences between the two is that SE does not provide for 16-bit processing (what they call 48 bit which is 3*16 for R,G,, so your files would be 8-bit depth and would therefore lack the advantages of 16 bit-depth.
-
Hi Ray,
Good image with believable colours. Can't argue with what works. It's obvious your set-up is delivering what you need without further ado, and in those circumstances of course one leaves well enough alone. However, this did not HAVE to be the case. There is no guarantee that every scanner will do likewise - in fact there is a guarantee that not every scanner will - hence the general case for colour management solutions.
[a href=\"index.php?act=findpost&pid=127240\"][{POST_SNAPBACK}][/a]
Thank you, Mark, but we haven't got to the nub of the issue, and that is; how the heck can a calibration routine take into account an inconsistent fading process that will vary according to humidity, temperature, and exposure to ultra violet?
It would be possible, I imagine, to create a computer program that could analyse any Kodachrome slide and determine that the fading was a result of a certain ageing process. However, can such a program make allowances for uneven ageing due to extreme and uneven conditions. If it can, is such a program a part of the IT8 calibration process. I doubt it.
-
Ray, the nub of the issue is that profiling does not depend on the quality of the image you are processing (whether it is faded, off-colour or whatever). It depends on the numbers in the patches of the standard target (IT8 or Hutchcolor) which are there to compare the colour values of each patch in the target with the values your scanner produces for each of those patches, and based on that the profiling software characterises the color reproduction behaviour of your scanner. That information is used for deriving consistent interpretation of colour values between the image file, the scanner, the monitor and the printer, which is one of the basic purposes of a colour-managed workflow. This is also discussed in "Real World Color Management" Second Edition Chapter 3.
-
Ray, the nub of the issue is that profiling does not depend on the quality of the image you are processing (whether it is faded, off-colour or whatever). It depends on the numbers in the patches of the standard target (IT8 or Hutchcolor) which are there to compare the colour values of each patch in the target with the values your scanner produces for each of those patches, and based on that the profiling software characterises the color reproduction behaviour of your scanner. That information is used for deriving consistent interpretation of colour values between the image file, the scanner, the monitor and the printer, which is one of the basic purposes of a colour-managed workflow. This is also discussed in "Real World Color Management" Second Edition Chapter 3.
[a href=\"index.php?act=findpost&pid=127280\"][{POST_SNAPBACK}][/a]
Mark,
It would seem to me that that's the nub of the issue if one is trying to get a perfect color match between output and a recently shot scene, (eg. a Dulux color chart, a fashion show, an advertisement for make-up etc.), but hardly seems relevant when the colors on the slide are miles out due to fading.
I've seen 'before/after' IT8 calibration comparisons on the net. The differences are noticeable but definitely subtle. Far more important than IT8 calibration in these circumstances, it appears to me, is a scan preview that consistently and accurately matches the shades and hues of the scanned output. Is IT8 calibration necessary for this match between preview and scan, whatever the colors may be? If it is, then I guess I'm just lucky that my scanner is close enough in its uncalibrated state.
I still use Vuescan as well as SilverFast. It was the latest upgrade of Vuescan that allowed me to use my Minolta 5400 II on my Win XP 64 bit system. The upgrade included a generic driver that fortunately works with XP64. SilverFast includes no such driver.
One has to tell Vuescan where the monitor profile is and specify the 'working space' profile one is using; ProPhoto RGB in my case. Before using the program for the first time, Vuescan goes through some calibration routine which appears to have nothing to do with IT8 calibration. Just what this calibration does, I don't know, but the results are very satisfactory.
-
Thanks for the advice folks. Much of this discussion is more advanced than my current level. I didn't even know about calibrating scanners; guess I'll have to start looking into it. (If anyone has any links they know of, that would be helpful)
Kodachrome is/was indeed a wonderful film. Any opinions on whether the gamut of RGB or CMYK is better suited to it?
-
Mark,
It would seem to me that that's the nub of the issue if one is trying to get a perfect color match between output and a recently shot scene, (eg. a Dulux color chart, a fashion show, an advertisement for make-up etc.), but hardly seems relevant when the colors on the slide are miles out due to fading.
I've seen 'before/after' IT8 calibration comparisons on the net. The differences are noticeable but definitely subtle. Far more important than IT8 calibration in these circumstances, it appears to me, is a scan preview that consistently and accurately matches the shades and hues of the scanned output. Is IT8 calibration necessary for this match between preview and scan, whatever the colors may be? If it is, then I guess I'm just lucky that my scanner is close enough in its uncalibrated state.
I still use Vuescan as well as SilverFast. It was the latest upgrade of Vuescan that allowed me to use my Minolta 5400 II on my Win XP 64 bit system. The upgrade included a generic driver that fortunately works with XP64. SilverFast includes no such driver.
One has to tell Vuescan where the monitor profile is and specify the 'working space' profile one is using; ProPhoto RGB in my case. Before using the program for the first time, Vuescan goes through some calibration routine which appears to have nothing to do with IT8 calibration. Just what this calibration does, I don't know, but the results are very satisfactory.
[a href=\"index.php?act=findpost&pid=127362\"][{POST_SNAPBACK}][/a]
Ray - if the colours of the slide are miles out and you are using Vuescan or Silverfast to bring them back, you are depending on the monitor preview to tell you when they look right as you make adjustments to the data governing the preview file. And then you will be depending on the scanner to reproduce those numbers in the scanned vesion of the file. Generally speaking, if either the monitor or the scanner are not characterized and profiled it is more than likely that the output of the one would not resemble the output of the other and therefore you wouldn't get the apperance of the file numbers you intended for the corrective adjustments to that faded, off-colour slide.
It would seem to me that the good behaviour you are getting from your combination of Vuescan and the scanner is either good fortune, or perhaps Vuescan is doing something with the scanner under the hood which produces this happy outcome. I can't profess to know why you are so fortunate, because I haven't used your set-up or done that particular work. Perhaps others could help here.
-
The problem with scanning Kodachrome is really the fact that it has a much higher dmax than either Fujichrome or Ektachrome and there is not a single CCD scanner on the planet that can come close to seeing all the way into those deep deep shadows. I've scanned a ton of Kodachrome on my Howtek, and, trust me, there is no problem scanning this film aside from the dmax issues. Kodachrome has a dmax of around 3.7 - 3.8. Drum scanners can see to about 3.9 and CCD's to about 3.0 on a good day.
I just finished making a bunch of 32 X 48 inch prints for a friend of mine from K64 and K25 slides that he scanned on his Howtek 4500 at 4000 ppi. The prints looked phenomenal even if they were a tad grainy at that size, but what great looking grain it was. These prints looked better than the Lightjets of the same size in Bishop that were scanned on a Primescan. There is a real difference between brands of scanners.
I have profiles made from Kodak, Fuji and Hutchcolor Velvia targets, and the only thing I have to do is on some Kodachromes, is to manually set the black point, which tends to be too high in the blue channel. Push the blue channel down in the blacks and the color just comes together.
-
Ray - if the colours of the slide are miles out and you are using Vuescan or Silverfast to bring them back, you are depending on the monitor preview to tell you when they look right as you make adjustments to the data governing the preview file. And then you will be depending on the scanner to reproduce those numbers in the scanned vesion of the file.
Absolutely right, Mark.
Generally speaking, if either the monitor or the scanner are not characterized and profiled it is more than likely that the output of the one would not resemble the output of the other and therefore you wouldn't get the apperance of the file numbers you intended for the corrective adjustments to that faded, off-colour slide.
Both Vuescan and Dimage Scan Utility need the monitor profile to be loaded into the software. Silverfast seems to be able to work this out for itself. The scanned images from Vuescan match the preview as closely as one could hope for. However, the scanned images from Dimage Scan Utility and Silverfast are often very slightly different. Don't know why, except possibly certain adjustments to the image are not fully reflected in the preview. I don't think an IT8 calibration has anything to do with it.
Below is an example of a 44 year old Kodachrome scanned using Dimage Scan Utility. The colors look identical to me, between preview and scan, except for the sky which is noticeably bluer in the preview. I never see such differences when using Vuescan.
This is a problem image however. I've been unable to get the colors to my satisfaction in either the scanning software or in Photoshop.
Did you know that Bangkok had trams in 1963, Mark.
[attachment=2800:attachment]
-
Hi Ray,
Thank goodness they abandoned the trams (wish we could progress that way in Toronto) - can you imagine what traffic in Bangkok would be like if they still had them on top of the rest of the situation there? The BTS helped I guess, but I've recommended to the Mayor of Toronto that he visit Cairo, Lagos and Bangkok if he wants a vision of where Toronto traffic is rapidly heading. No reply of course. We just vote for these people..............
Turning to your images, the sky is the place where you most obviously see the hue differences, but when I look closely and intently at those two images on my monitor - unless my imagination is running wild - I can detect there is a more generalized hue difference which is camouflaged by the relative lack of brightness in the non-sky areas. If I'm correct in that observation, it means that there is a colour management issue going on somewhere.
I was never impressed with colour handling in Minolta's scanner software. Vuescan is better and (for negatives) Silverfast is the cream of the crop despite the obtuse U.I. and crummy documentation.
I'm talking generally - not specifically for any particular image, but I don't understand why that image should be especially problematic. In particular, it should be totally fixable in Photoshop. If it's a general cast a tweak of a channel curve or two should do it; if it's a specific sky cast, the same treatment with a layer mask isolating the sky should work. Perhaps you've tried that already? Shouldn't that have delivered the goods?
-
I could be wrong, but I don't believe much purpose is served calibrating the scanner with a Q60 slide if the slides to be scanned have been subjected to fading.
I'm very impressed with the capabilities of my Dimage Scan 5400 II, which, as Mark mentioned, seems to be able to remove scratches through use of ICE with Kodachromes, something that other scanners, such as my Nikon 8000 cannot do without producing artifacts such as double edges.
I haven't been keeping up with developments in scanner technology but I wonder if there's now anything even better than the KM 5400 II at a similar affordable price.
[a href=\"index.php?act=findpost&pid=127185\"][{POST_SNAPBACK}][/a]
My comments related to unfaded slides which is what I have. I have no experience of using the colour restoration tools of the Coolscan scanners, but I suspect it still makes sense to calibrate to get a good starting point. Someone who owns a Coolscan can no doubt comment.
Kodachrome also has a tendency to colour casts and colour changes due to reciprocity failure. Again calibration can provide a starting point.
Many scanners have problems using ICE with Kodachrome. Some people say Coolscans are okay with ICE and Kodachromes. Some say they are not. It might be that the emulsion has undergone changes throughout the decades, and some formulations cause issues.
I am not sure there is anything better. The market has moved over to digital, and scanners probably sell in much smaller numbers. Hence there is less motivation to commit funds to a new model.
However ... this might mean there are some used drum scanners going cheap?
-
I'm talking generally - not specifically for any particular image, but I don't understand why that image should be especially problematic. In particular, it should be totally fixable in Photoshop. If it's a general cast a tweak of a channel curve or two should do it; if it's a specific sky cast, the same treatment with a layer mask isolating the sky should work. Perhaps you've tried that already? Shouldn't that have delivered the goods?
[a href=\"index.php?act=findpost&pid=127553\"][{POST_SNAPBACK}][/a]
Mark,
I now believe I must have neglected to perform a fresh prescan after making some adjustment, to cause that sky difference with the Minolta software. Below is another scan using Silverfast where I've done my best to correct for color hues and an excessively dark background.
To my eyes, the two images are very, very close, including the sky. The only difference I can pick up is the very slightly deeper shade of burgundy of the Volkswagon in the preview on the right.
I don't know why this image is a problem. The colors are just 'yukkie'. Perhaps it's because the main part of the image is basically underexposed. Exposure was for the sky. Perhaps it's because there's a conflict of different WBs in the image as a result of different rates of fading and deterioration of the colors. I wouldn't expect an IT8 target would help here.
[attachment=2801:attachment]
Yep! That sky train in Bangkok is just great. They should extend it as much as possible. I'm not sure if the current extension to the new airport is going to serve much purpose. I can't imagine thousands of travellers struggling to heave their heavy suitcases or backpacks onto a train just to save a few Baht, especially if they have to get a taxi from the train station to their hotel, having struggled with their luggage down dozens of steps to the roadside.
-
I have no experience of using the colour restoration tools of the Coolscan scanners, but I suspect it still makes sense to calibrate to get a good starting point. Someone who owns a Coolscan can no doubt comment.
Kodachrome also has a tendency to colour casts and colour changes due to reciprocity failure. Again calibration can provide a starting point.
There's always a starting point. Calibration might provide a slightly different starting point but I can see no reason why it should be a better starting point with faded slides where the color scheme and degree of deterioration is not known (scientifically and precisely, that is).
In fact, as a result of unknown changes to the hues and saturation of the faded slide, a calibration of the scanner might result in a starting point which is worse.
Many scanners have problems using ICE with Kodachrome. Some people say Coolscans are okay with ICE and Kodachromes. Some say they are not. It might be that the emulsion has undergone changes throughout the decades, and some formulations cause issues.
I used to have problems using ICE with my Nikon 8000ED, with Kodachrome. No such problems with the Minolta 5400 II with Minolta's own software or with SilverFast.
In the image in my previous post, the left part opened in Photoshop has had ICE applied in Silverfast. The preview doesn't show the effects of ICE.
-
Mark,
Below is another scan using Silverfast where I've done my best to correct for color hues and an excessively dark background.
To my eyes, the two images are very, very close, including the sky.
The colors are just 'yukkie'. Perhaps it's because the main part of the image is basically underexposed. Exposure was for the sky. Perhaps it's because there's a conflict of different WBs in the image as a result of different rates of fading and deterioration of the colors. I wouldn't expect an IT8 target would help here.
[attachment=2801:attachment]
Yep! That sky train in Bangkok is just great. They should extend it as much as possible. I'm not sure if the current extension to the new airport is going to serve much purpose. I can't imagine thousands of travellers struggling to heave their heavy suitcases or backpacks onto a train just to save a few Baht, especially if they have to get a taxi from the train station to their hotel, having struggled with their luggage down dozens of steps to the roadside.
[a href=\"index.php?act=findpost&pid=127561\"][{POST_SNAPBACK}][/a]
Hi Ray,
Dramatic improvement in color alignment.
Under-exposure of the non-sky area shouldn't be a particular challenge for getting the hues right - but brightness clearly suffers. The idea that different dyes fade at different rates sounds plausible, but that too shouldn't be a problem unless it happens disproportionately in different areas of the image. I can see that causing a huge colour balancing nightmare. And in that situation, while a properly profiled scanner for colour slides remains <a good thing>, I agree it would not solve that kind of problem - you'd have properly managed conflicting white balances!
Haven't been through the new airport at BKK. I hear there have been teething issues.
-
Mark,
Thanks. Just for the record, here's an improvement but the skin tones are still not right. I'm afraid I can't make do without Photoshop. Perhaps I should be able to achieve such results (and better) in Silverfast. I need more practice .
Haven't been through the new airport at BKK. I hear there have been teething issues.
Just a few subsidence problems due to the fact the place was once a swamp, cracks on the runway etc. Nothing serious .
[attachment=2803:attachment]
-
Ray,
It's pretty good! I think I would de-saturate the Yellows by something between 5% and 10% - that could eliminate the last vestige of muddiness and help the skin tones too - at least from what I see on my monitor.
As decent a program as Silverfast is, I don't rely on it to do everything. Most importantly, I can't soft-proof with the printer profile at the scanning stage, so that alone necessitates a trip into Photoshop. In fact, all I count on Silverfast for is the heavy lifting on luminance and colour balance for the scan, then I do everything else in Photoshop where the tools and plug-ins I have make it so much preferable.
-
Ray,
It's pretty good! I think I would de-saturate the Yellows by something between 5% and 10% - that could eliminate the last vestige of muddiness and help the skin tones too - at least from what I see on my monitor.
As decent a program as Silverfast is, I don't rely on it to do everything. Most importantly, I can't soft-proof with the printer profile at the scanning stage, so that alone necessitates a trip into Photoshop. In fact, all I count on Silverfast for is the heavy lifting on luminance and colour balance for the scan, then I do everything else in Photoshop where the tools and plug-ins I have make it so much preferable.
[a href=\"index.php?act=findpost&pid=127608\"][{POST_SNAPBACK}][/a]
Mark,
You're quite right. There is a yellow cast. What I find puzzling is that there's plenty of opportunity for gray balncing in this image with chrome bumber bars and tram lines, but it doesn't seem to work. Produces instead some pretty awful casts.
The following change I think is a big improvement, but still not quite right. What do you think?
[attachment=2807:attachment]
-
Kevin:
I was in a similar situation a few years ago - 30 yrs of Kodachrome to convert to digital. There is lots of good advice in the previous posts (e.g. Mark's article); as you mentioned, some may be beyond your knowledge at the moment, but I will offer a suggestion: I have a Minolta 5400 and a Nikon 9000. I have kept the 5400 because it does a better job with most properly exposed Kodachromes. So, my suggestion, in the absence of drum scans, is to find a 5400
.
-
I don't know what things are like where you are but here in the UK, you can get a drumscan for an amazing £10 per scan. Compared to the £50 GBP we used to pay just a couple of years ago this is simply unbelievable.
I have used these people http://www.drumscan.co.uk (http://www.drumscan.co.uk)
I have spoken to these guys and basically they broker scans for scanner operators who are finding things incredibly tough as the demand has just fallen away incredibly over the past couple of years.
I had 35mm (not Kodachrome though) scans made by them and they were incredible.
I am a professional retoucher with 15 years in the business, so I know a good scan when I see one.
The problem with doing your own scanning, especially from an extensive back catalogue is space and time.
If you can negotiate a hard bargain, I'd say get someone else to do your scanning, but work with them for the results you need.
-
I don't know what things are like where you are but here in the UK, you can get a drumscan for an amazing £10 per scan. Compared to the £50 GBP we used to pay just a couple of years ago this is simply unbelievable.
[a href=\"index.php?act=findpost&pid=127773\"][{POST_SNAPBACK}][/a]
Time is money. If you are very busy, a one-time high resolution scan of your best Kodachromes at 10 pounds per slide might be the best way to go.
However, I paid roughly 300 pounds for my Monolt 5400 II scanner. That sum of money pays for the drum scanning of only 30 slides at this amazingly low price of 10 pounds per slide.
The other issue is the color balance if the slides are old, as I've tried to demonstrate with those examples in my previous posts. Does that price of 10 pounds include individual color correction for each slide. I doubt it. I imagine a batch of slides is stuck to the drum with some sort of fluid and they are all given the same treatment.
-
Hi!
Just want to mention that Vuescan is profiled. Ed Hamrick includes profiles for most scanners. Variations between scanners is supposedly small.
The idea with profiling is to characterize the scanning device. What you do is to scan a measured target (where each patch of color is exactly known) and calculate a conversion from scanned color to correct color. This characterization is for CCD, filters, light source and electronics like amplifiers and ADC (Analog Digital Converter).
There is some discussion about whether there is a need to calibrate for each film or not. The only reason I can see to do that is that different films can probably have different spectral transmission probabilities and IR (infrared) absorption.
I would assume that Nikon scanners which have a LED based light source may be less subjects to these effects than scanners using luminiscent panels or tubes.
Best regards
Erik
Ray - if the colours of the slide are miles out and you are using Vuescan or Silverfast to bring them back, you are depending on the monitor preview to tell you when they look right as you make adjustments to the data governing the preview file. And then you will be depending on the scanner to reproduce those numbers in the scanned vesion of the file. Generally speaking, if either the monitor or the scanner are not characterized and profiled it is more than likely that the output of the one would not resemble the output of the other and therefore you wouldn't get the apperance of the file numbers you intended for the corrective adjustments to that faded, off-colour slide.
It would seem to me that the good behaviour you are getting from your combination of Vuescan and the scanner is either good fortune, or perhaps Vuescan is doing something with the scanner under the hood which produces this happy outcome. I can't profess to know why you are so fortunate, because I haven't used your set-up or done that particular work. Perhaps others could help here.
[a href=\"index.php?act=findpost&pid=127364\"][{POST_SNAPBACK}][/a]
-
There's always a starting point. Calibration might provide a slightly different starting point but I can see no reason why it should be a better starting point with faded slides where the color scheme and degree of deterioration is not known (scientifically and precisely, that is).
In fact, as a result of unknown changes to the hues and saturation of the faded slide, a calibration of the scanner might result in a starting point which is worse.
I used to have problems using ICE with my Nikon 8000ED, with Kodachrome. No such problems with the Minolta 5400 II with Minolta's own software or with SilverFast.
In the image in my previous post, the left part opened in Photoshop has had ICE applied in Silverfast. The preview doesn't show the effects of ICE.
[a href=\"index.php?act=findpost&pid=127565\"][{POST_SNAPBACK}][/a]
The clue was in my posting when I said 'suspect'. That means I do not know and I am guessing based on my experience of f or 5 scanners. Which is precisely what you are doing since you also too seem to have no experience of faded emulsions.
"I can see no reason why it should be a better starting point with faded slides "
To state the obvious, if the fading is slight, then the calibration will get you close, given that in my experience of many scanners, the default uncalibrated scan of Kodachrome is way off. And again to state the obvious if the fading is heavy, then calibration might not help. Again the only way to know is to get feedback from someone who has done it. I think these points are both quite obvious, and I assumed it was not worth making them explicity.
My experience of slides with reciprocity failure is that calibration is a good starting point.
But then again you seem to argue for the sake of arguing which I find rather tiresome.
-
The clue was in my posting when I said 'suspect'. That means I do not know and I am guessing based on my experience of f or 5 scanners. Which is precisely what you are doing since you also too seem to have no experience of faded emulsions.
"I can see no reason why it should be a better starting point with faded slides "
To state the obvious, if the fading is slight, then the calibration will get you close, given that in my experience of many scanners, the default uncalibrated scan of Kodachrome is way off. And again to state the obvious if the fading is heavy, then calibration might not help. Again the only way to know is to get feedback from someone who has done it. I think these points are both quite obvious, and I assumed it was not worth making them explicity.
My experience of slides with reciprocity failure is that calibration is a good starting point.
But then again you seem to argue for the sake of arguing which I find rather tiresome.
[a href=\"index.php?act=findpost&pid=128011\"][{POST_SNAPBACK}][/a]
Your last sentence is an uncalled-for personal remark. Let us keep this discussion confined to the technical merits of the issues.
As I've said before, as long as you are NOT working with colour negatives (where it can't be done), characterizing a scanner is a *good thing to do* regardless of the characteristics of the image. It will at least return a correct (if correctly identified) neutral tone.
Where Ray has an issue - legitimate I believe - is how helpful calibration would be for these faded images, and this may have little to do with how faded the image is. Ray has mentioned the possibility that the different dyes may not have faded proportionately. This would obviously muck-up grey-balance or white-balance adjustments, because they all depend on mapping relativity between colours; hence if that relatively has been messed by disproportionate fading between the primaries, common sense would suggest that all the calibration in the world will not produce a balanced image in one click. However, it will get the grey or white right, and that is a useful starting point - IF the grey or white really is grey or white. If white/grey balancing does not work correctly because of this factor, it is necessary to use the individual RGB curves for balancing the image. Silverfast, for example, allows this at the scanning stage; if the scanning software doesn't, it is a task for Photoshop.
-
Your last sentence is an uncalled-for personal remark. Let us keep this discussion confined to the technical merits of the issues.
I disagree. Both of his postings were entirely negative, along the lines of "Well you are completely wrong but I cannot give any evidence, anyway I know this and you dont".
Now if he had some actual experience, and could make an observation, then that would have added something.
As someone else has said, calibration simply means that the scan is an accurate representation of the image, even though that image might have a cast or some fading.
FWIW I have owned a Coolscan LS30, a Minolta 5400, another scanner the name of which escapes me and a flatbed so I do have some experience on which to base opinons. And IMO Kodachrome is a devil to scan, and only the more recent scanners can do it justice. And I have found that it is better to calibrate first even when the slide has a colour shift. The uncalibrated scans are worse.
Incidentally I found most magazine scanner reviews to be worse than useless as they never tried Kodachrome. The OP should be aware of that.
-
Once you have profiled for the characteristics of the scanner it can then be modified by applying an inverted curve and re-profiling. You may have to average out several images representative of the color cast, of course more correction will be needed post processing. The Hutch site describes this method. This has helped me as I dive into more than 4500 for a client.
Ray,
In Vuescan you may want to try setting the Media to Image, Scanner ICC Profile (if any) to your modified profile, and Output Color Space to Device RGB. The scanner profile only affects the preview once you click scan. If you don't have a profile you can still preview your working space by setting it as you do now. My preference is to assign and convert in Photoshop. Ed's efforts at improving Vuescan's infrared cleaning paid off as it works very well.
-
I disagree. Both of his postings were entirely negative, along the lines of "Well you are completely wrong but I cannot give any evidence, anyway I know this and you dont".
[a href=\"index.php?act=findpost&pid=128042\"][{POST_SNAPBACK}][/a]
Slough,
That statement is completely untrue. Show what comments I've made in this thread that could possibly be interpreted in that way.
All my comments in this thread and all threads on this forum, for that matter, are based on either personal experience or personal opinion based on rational deduction from the evidence available as well as the opinion of others. I back up my opinions with photographic examples when possible.
If you would like to back up your opinion with a before/after example of how a scanner calibrated with an IT8 target can help a faded slide, then please do so, but make sure it really is a faded slide.
I have literally thousands of faded slides & negatives. Do you think I wouldn't be calibrating my scanner if I thought it would make a difference? Do you think I'm mad? The fact is, I have few slides that are not faded to some degree. My most recently shot film before I switched to digital about 6 years ago was Fuji Reala and Kodak Royal Gold color negative film.
I have no trouble getting the hues very close when scanning negative film because the scanner software includes a variety of profiles for different negative film types, which make a huge difference to the color and tonality. However, Vuescan also contains a few slide profiles, namely, Generic, Kodak Kodachrome, and Kodak Ectachrome.
With the slides I'm scanning, it makes not a whit of difference which profile I choose. The preview looks the same. As I mentioned before, the examples I've seen on the net lead me to believe that profiling with an IT8 target makes only a subtle difference. If your slide is faded by more than a subtle degree, I can't see the point.
I'm relying upon people like you, so convinced it does make a difference, to show me some comparisons. If we were to switch places and I was the one arguing the point that a calibration makes a worthwhile difference to the preview of a faded slide, I'd show you. I wouldn't expect you to take my word for it.
-
Ray,
In Vuescan you may want to try setting the Media to Image, Scanner ICC Profile (if any) to your modified profile, and Output Color Space to Device RGB. The scanner profile only affects the preview once you click scan. If you don't have a profile you can still preview your working space by setting it as you do now. My preference is to assign and convert in Photoshop. Ed's efforts at improving Vuescan's infrared cleaning paid off as it works very well.
[a href=\"index.php?act=findpost&pid=128053\"][{POST_SNAPBACK}][/a]
David,
Thanks for the tip. I'll try that. I tend to go through a phase of scanning activity, become familiar the quirks of the software, then leave it for a while. When I resume scanning perhaps a year or so later, I find I have to relearn.
-
Ray: I found your comments somewhat dismissive. As I stated I found calibration very useful when scanning slide with a slight colour cast due to reciprocity failure or cold/warm light. With significant reciprocity failure it was often impossible to get a good scan, which is not surprising if you think about what is going on in the emulsions.
As I clearly stated I have not tried scanning faded slides. But with slight fading I would choose to use a profile, based on my experience with colour casts etc.
"Do you think I'm mad? "
No. Is that a requirement for posting here? Well, not on the evidence it isn't.
"With the slides I'm scanning, it makes not a whit of difference which profile I choose"
I am very surprised by that statement as I have always found profiling to be essential as I see large differences between calibrated and uncalibrated scans, the main difference being a marked colour cast. I have found this to be true with many scanners and emulsions including Fuji Reala 100, Fuji Provia 100F, Fuji Sensia 100, several Kodak slide emulsions (I forget the names) and Kodachrome. In the case of Kodachrome I find the uncalibrated scan to be awful.
Incidentally the slide scanners I owned are the Nikon LS30, Microtek 4000T, and the Minolta 5400 and I still own the last one.
My experience is that emulsions from the same manufacturer using the same process e.g. E41 tend to scan similarly, and hence can use the same profile, which is perhaps not surprising.
I also use a calibration device for my monitor to ensure that the whole work flow is calibrated.
I gave up using Vuescan many years ago as I found it relatively useless with the Minolta 5400. The supplied software was much better and less clunky. I had no problem using Nikon Scan, the Minolta software, the Microtek software, or the German third party application supplied with the 4000T, so I suspect that Vuescan simply did not function correctly with that scanner at that time. Or perhaps the supplied drivers for the OS I was using were dodgy. Who knows.
From what you suggest it sounds as if Ed, the creator, has improved it greatly. I remember sending him some emails, and to his credit, he replied to each. Not many applications get support direct from the creator.
Anyway, it seems to me that the thread has gone way of course. The OP made no mention of faded slides. He did ask some specific questions though, and I am not sure if we have answered them all?
-
Anyway, it seems to me that the thread has gone way of course. The OP made no mention of faded slides. He did ask some specific questions though, and I am not sure if we have answered them all?
[a href=\"index.php?act=findpost&pid=128227\"][{POST_SNAPBACK}][/a]
Slough,
The OP mentioned digging out 15+ year old Kodachromes and Ecatchromes. They'd be faded to some degree. As I recall, Kodak Q60 targets should not be used if they are more than 6 months old. I actually have a Q60E3 target. Whilst searching for some slide material which is not significantly faded so I could test David Good's suggestion of scanning into Device RGB instead of the usual color spaces like sRGB or ProPhoto, to get a clearer idea of just what the differences are, I came across it, neatly folded in an envelope. The date, 1999:12.
I didn't see much point in using a 6 1/2 year old target as a reference point for calibration, so just used it as a normal slide to scan at differnent settings. The results seem quite normal to me. No color cast that I can notice. Natural skin tones on the lady on the right. All the colors and hues are clearly delineated, including the brightest and the darkest, the most saturated and least saturated. And, of course, there's plenty of opportunity for grey balancing on the Q60 target, which was another indication of the lack of a color cast because grey balancing wasn't required.
Scanning into Device RGB results in an image without embedded profile so the appearance is dependent on what profile is assigned when opening in Photoshop. Assigning ARGB seems to produce the most natural result. Assigning sRGB is too desaturated and ProPhoto RGB too saturated.
I get no sense that my scanner needs calibrating. I haven't got time to post the images right now because I'm due somewhere else shortly.
Of course, now I've found my Q60 target, you can be sure I'd at least make an attempt to use it to create an IT8 profile, just to see the magnitude of the difference it makes with the 5400 II, if for no other reason.
However, I failed to successfully implement the calibration. The Vuescan instructions on this process seem unclear. I'm supposed to rename a data file to scanner.it8 (which I downloaded from Kodak) and copy it to the VueScan.ini folder.
Problem is, my computer can't find any such file. I did a search for files and folders. There's a Vuscan folder of course with a heap of files of various descriptions, but no VueScan.ini. This is something I'd have to contact Ed Hamrick about, but I'm reluctant to waste everyone's time for the sake of being able to profile my scanner with a 6 1/2 year old target, especially considering I've seen no evidence that such calibration would help me get better results with faded Kodachromes and other faded film types.
Nevertheless, the Vuescan software does contain 3 slide profiles, Kodachrome, Ektachrome and Generic. None of these appear to be activated. It makes no difference which I choose, if any or none. If I'm able to use these 3 profiles as a result of an IT8 calibration, then that clearly increases my options and it could be worthwhile buying a recent Q60 target.
A google search on the subject of calibrating with an IT8 target, specifically using Vuescan, gives me the impression there's a lot of confusion here. I'm not the only one who finds Vuescan's instructions here inadequate.
I should add, I have no problem scanning negative film, except sometimes when I can't find the profile for my film type. If the software includes a profile for the specific film type I'm scanning, results are close to perfect. If it doesn't, results can be miles out.
-
The term calibrate seems to be substituted for profile. Calibration just resets the device to a known state, you may want to re-calibrate after a temperature or humidity change.
The profile generated from the target then describes how the scanner behaves, Vuescan has a separate tab for Profile Scanner. I have not tried profiling with Vuescan as I'm using a 6x7 Hutch Velvia target that is not supported. A search should turn up a number of freeware programs that can profile the Q60 targets.
Ray,
In Vuescan's output tab make sure the color balance is set to None, this way the scanner data gets transfered unaltered. Also verify that Tiff Profile is not checked that way you can assign your profile then convert (maybe Ekta space) in Photoshop.
-
Ray,
In Vuescan's output tab make sure the color balance is set to None, this way the scanner data gets transfered unaltered. Also verify that Tiff Profile is not checked that way you can assign your profile then convert (maybe Ekta space) in Photoshop.
[a href=\"index.php?act=findpost&pid=128282\"][{POST_SNAPBACK}][/a]
David,
I don't seem to be getting any transfer of scanner data when I attempt a profile. On my Win XP 64 bit system, there is no Vuscan.ini file to be found. I think the problem might be, Vuescan is designed for many scanner models and operating systems and the precise IT8 profiling procedures for each scanner model might be slightly different, and is not detailed in the PDF manual.
I vaguely understand there is a difference between profiling and calibarating with regard to scanning, but I've not gone into it, so can speak with no authority.
In my current situation, all 3 of the slide profiles available in Vuescan are not working. The paramount question in my mind is, do I need to profile my scanner with an IT8 target in order to get these profles working, or is there something simply broken?
-
Whew, I didn´t know it could get so complicated, said the bishop about corset stringing!
I have a Canon scanner, a CanoScan FS4000US which I bought with one objective in mind: to scan old Kodachromes from my calendar days and to play with Velvia too.
I never use any of the devices like ICE because it is impossible for me to imagine dirt or scratches being softened out of sight without legitimate information going the same way. It is boring, tiresome and very bad for DVT, but PS allows you to clean up your images as everybody knows already - just try to file your transparencies carefully and thus minimise the time it takes doing the spotting job.
Now, I don´t have any idea if the scanner would do a better job of altering my scans than can PS but as I would have to go into PS anyhow, why bother? I have had very good results from scanning Kodachrome and find that in my medium - people shots - the skin tones you get in black and white prints is out of this world, as is the detail. In fact, as far as people shots go, if Kodachrome were available today in 120, I´d be very tempted to go back into that format and scrub 35mm from the menu.
The trouble with running both film 35mm (F3 - one of the very last to be had!) and digital (D200) is that there is this tendency to compare which often leads to paralysis and NOTHING gets photographed. Yes, a little off-topic, but part of the hybrid world we have fallen into.
So, my route, for better or worse, is to let the scanner do a straight scan and PS do the rest. I admit that as I work mainly in B/W now, that I might have been saved from colour problems from scanning thus.
As for drum scanning - I was always led to believe that high-quality drum scanners were in the many thousands of pounds area, not really amateur or self-employed pro prices at all. When I was getting scans done back in the mid-eighties I was lucky to find anything under about UKP 75 for a smallish image.
As for the suitability of Kodachrome for large reproductions, just go look at some of the old Pirelli and Mintex calendars or even Lichfield´s Uniparts. They, and others, were the reason I used Kodachrome 64 and Nikon instead of Hasselblad and Ektachrome. I had both, used both, but not for calendars where the 35mm won hands down. Needless to say, I´m not talking landscapes.
Ciao - Rob C
-
Ray,
If you use Image in the Media tab there is no need to choose those profiles. Set Scanner color space to ICC profile, not built-in. When you have a profile built you can select it in the next tab, it is used for the display and is assigned when you choose to embed the profile into the tiff (don't).
The procedure of profiling a scanner is essentially the same for all models, scan, align the squares as best you can in the profiling software and it does the rest (simplified pipeline). I can't speak for the .ini file, have you tried any other software to build the profile from? Nothing's broken, just have to get all the right settings. As this thread has gone off topic you could e-mail me your settings if that would help.
-
I never use any of the devices like ICE because it is impossible for me to imagine dirt or scratches being softened out of sight without legitimate information going the same way. It is boring, tiresome and very bad for DVT, but PS allows you to clean up your images as everybody knows already - just try to file your transparencies carefully and thus minimise the time it takes doing the spotting job.
[a href=\"index.php?act=findpost&pid=128289\"][{POST_SNAPBACK}][/a]
Rob,
You've underestimated the ability of infra red rays (and associated software) to distinguish between physical scratches and image detail. There's a difference and this difference cannot be recognised in Photoshop and other noise reduction programs such as Neat Image.
What surprises me is, after all this time of technological development in scanning, there's no equally satisfactory method of removing physical scratches from old silver based B&W negatives. I've got many hundreds (possible thousands) of those and it's really tedious removing the scratches and blemishes, although the spot healing brush helps a lot.
-
Ray,
If you use Image in the Media tab there is no need to choose those profiles. Set Scanner color space to ICC profile, not built-in. When you have a profile built you can select it in the next tab, it is used for the display and is assigned when you choose to embed the profile into the tiff (don't).
[a href=\"index.php?act=findpost&pid=128292\"][{POST_SNAPBACK}][/a]
David,
I'm away from my studio at present, on another computer, so my descriptions might not be completely accurate, but from memory, when choosing a profile in Vuescan one has to specify the location of the profile. The profile has to exist somewhere. If I choose to use my moinitor profile, I have to specify in another box where the monitor profile resides.
I cannot find any scanner ICC profile in my computer, after going through the motions of profiling with an IT8 target.
-
Rob,
You've underestimated the ability of infra red rays (and associated software) to distinguish between physical scratches and image detail. There's a difference and this difference cannot be recognised in Photoshop and other noise reduction programs such as Neat Image.
What surprises me is, after all this time of technological development in scanning, there's no equally satisfactory method of removing physical scratches from old silver based B&W negatives. I've got many hundreds (possible thousands) of those and it's really tedious removing the scratches and blemishes, although the spot healing brush helps a lot.
[a href=\"index.php?act=findpost&pid=128293\"][{POST_SNAPBACK}][/a]
Indeed. To expand on this, I use a Coolscan 4000 with Vuescan and you have the option of being able to view just the infrared layer, which can be quite an eyeopener. Image detail in the RGB layer of E6/C41 film is transparent to infrared and it is only the defects, (scratches, dust etc.) which are recorded there.
Vuescan (and presumably ICE), then uses this information to do a clone operation of pixels immediately adjacent to the defects - no other processing is done on the rest of the image.
The reason this does not work with silver based B/W negs, is that the silver is opaque to infrared, rendering ICE unusable.
Regards,
Richard
-
After some stuffing around and after installing the very latest Vuescan update of 16 July, I succeeded in creating an ICC profile for my Minolta 5400 II scanner using my Q60 target.
[attachment=2834:attachment]
I was then able to scan the Q60 target choosing either the scanner ICC profile I'd just created or the scanner's built-in profile, which is the one I've been using to date.
In the comparisons below, the scan from the calibrated profile is always on the left.
[attachment=2835:attachment]
Now, to my eyes there's no doubt that the image on the left is more accurate. The lady in the right image might have a slightly healthier complexion, a more saturated and colorful complexion which is perhaps more appealing to some, but the primary colors in the squares are slightly out. Specifically, the blues have a touch of green; the greens have a touch of blue; the reds have a touch of yellow and the yellows are not as saturated as the yellows in the image on the left.
Over all, these differences would be significant when scanning relatively new slides. If I was still shooting slides, I would definitely get a fresh Q60 target.
Unfortunately, I wasn't able to get very far comparing the effects of these relatively subtle changes on other slides because the program became corrupted, but before it did, I was able to make a comparison of the 44 year old Kodachrome of the scene in Bangkok.
Both of these scans have had exactly the same settings applied, the only difference being the scanner profile used. I must admit I was surprised at the magnitide of the differences. I suspect these differences have been amplified as a result of ticking the 'restore color' and 'color fading' options in Vuescan.
The image from the Q60 calibrated profile is on the left.
[attachment=2836:attachment] [attachment=2837:attachment] [attachment=2838:attachment]
Even though I might have to swallow some pride here, I think I would rather begin working on the image on the left . I guess I'll now have to search the net for the best price on a new Q60 target. A price estimate from Kodak in Australia is A$250. My local photographic store had no reference to such a product in their system. I guess these calibration targets are going out of fashion.
-
Hi Ray - no need to swallow any pride - you've just convinced yourself that profiling works, and thanks for showing us the outcomes - very interesting. I agree the Bangkok image has improved - the profiling tamed the yellows better - complexions are probably more accurate, and the VW is more like the wine colour I remember for those old VWs - but here we really are talking *memory* colours, because it's "ancient history" by now (not to even begin wondering about how much of BKK air pollutants the driver may or may not have wiped off his vehicle every day)! Seriously though, I think your a step ahead now.
-
Ray,
Glad it's finally working for you! The price quoted to you sounds excessive, there are several companies such as Chromix that can supply an IT8/Q60 for quite a bit less. At that price you might as well order a Hutch target for a little more. The Kodak site also lists a Kodachrome 35mm target (cat# 157 5141) that may be of interest to you.
Cheers,
Dave
-
Hi Ray - no need to swallow any pride - you've just convinced yourself that profiling works, and thanks for showing us the outcomes - very interesting. I agree the Bangkok image has improved - the profiling tamed the yellows better - complexions are probably more accurate, and the VW is more like the wine colour I remember for those old VWs - but here we really are talking *memory* colours, because it's "ancient history" by now (not to even begin wondering about how much of BKK air pollutants the driver may or may not have wiped off his vehicle every day)! Seriously though, I think your a step ahead now.
[a href=\"index.php?act=findpost&pid=128576\"][{POST_SNAPBACK}][/a]
Thanks Mark. What impressed me was the fact that although the yellow cast was largely removed, using the calibrated profile, and the monks' robes look closer to their usual reddish orange, the yellows elsewhere in the image, such as the wall on the far right, are actually stronger.
Reducing the yellow in the robes and complexions in the non-calibrated image is not too difficult but a consequence of this is a reduction in the yellows elsewhere in the image, which are already probably too weak. If using a scanner icc profile saves time and work in processing a faded image, then that's reason enough to use it.
However, we should not forget this was a calibration attempted with a target manufactured in the last century. There's no guarantee that a fresh Q-60E3 target would produce better or even equally good results. In fact there's a certain logic to the idea that for best results with faded slides the target should be equally faded.
-
Ray,
Glad it's finally working for you! The price quoted to you sounds excessive, there are several companies such as Chromix that can supply an IT8/Q60 for quite a bit less. At that price you might as well order a Hutch target for a little more. The Kodak site also lists a Kodachrome 35mm target (cat# 157 5141) that may be of interest to you.
Cheers,
Dave
[a href=\"index.php?act=findpost&pid=128579\"][{POST_SNAPBACK}][/a]
David,
Thnks for your concern. I now realise from the catalogue no. quoted that the Kodak lady on the phone gave me an estimate for a 4x5" IT8 target. Nobody seems to be using such targets over here. I think I'll probably upgrade to SilverFast v 6.5 which does multiple exposures and automatic calibration with IT8 targets. Sometimes I get easier and faster results with Vuescan but Silverfast has far more potential for fine adjustments.
-
In fact there's a certain logic to the idea that for best results with faded slides the target should be equally faded.
[a href=\"index.php?act=findpost&pid=128583\"][{POST_SNAPBACK}][/a]
No. The logic of a profile is that it characterizes the colour rendering behaviour of your scanner so that it returns the color data you chose for your image file. This is a necessarily separate function that is independent of the image data - whether faded or not faded.
-
No. The logic of a profile is that it characterizes the colour rendering behaviour of your scanner so that it returns the color data you chose for your image file. This is a necessarily separate function that is independent of the image data - whether faded or not faded.
[a href=\"index.php?act=findpost&pid=128600\"][{POST_SNAPBACK}][/a]
Hmm! I thought there was a distinction between calibrating the scanner and creating an ICC profile for the scanner. The differences between the preview and the scanned image, on my current set-up, have always seemed trivial to me, using my monitor profile; nothing as great as the differences in the hue of those saffron robes.
-
True - calibrating and profiling are separate functions, and you should get a close match between the preview and monitor image if your colour management set-up is OK.
-
True - calibrating and profiling are separate functions, and you should get a close match between the preview and monitor image if your colour management set-up is OK.
[a href=\"index.php?act=findpost&pid=128746\"][{POST_SNAPBACK}][/a]
Mark,
I'm probably revealing my complete ignorance of the subject here, but it seems to me there's one thing about the IT8 target which never changes and that's the descriptor data accompanying each batch of targets, which defines the colors in a mathematical and device independent way.
A scanner which has not been calibrated is likely to produce colors which are much closer to what they are supposed to be, with an unfaded slide, than a calibrated scanner will produce with a faded slide.
By calibrating with an IT8 target made from the same type of film (Kodachrome or Ektachrome) of approximately the same age as the Kodachromes or Ektachromes one is scanning, it should (might) be possible to get an accurate representation of the original colors of the faded slide.
If it's possible to physically calibrate the scanner by other means, then so much the better.
By the way, Vuescan's infra red scratch removal produces artifacts with Kodachromes using the Minolta 5400 II; specifically a tendency to slight double edges and blocky, deep shadows. The ICE feature in both Dimage Scan Utility and Silverfast seems to work well.
-
Mark,
I'm probably revealing my complete ignorance of the subject here, but it seems to me there's one thing about the IT8 target which never changes and that's the descriptor data accompanying each batch of targets, which defines the colors in a mathematical and device independent way.
A scanner which has not been calibrated is likely to produce colors which are much closer to what they are supposed to be, with an unfaded slide, than a calibrated scanner will produce with a faded slide.
By calibrating with an IT8 target made from the same type of film (Kodachrome or Ektachrome) of approximately the same age as the Kodachromes or Ektachromes one is scanning, it should (might) be possible to get an accurate representation of the original colors of the faded slide.
If it's possible to physically calibrate the scanner by other means, then so much the better.
By the way, Vuescan's infra red scratch removal produces artifacts with Kodachromes using the Minolta 5400 II; specifically a tendency to slight double edges and blocky, deep shadows. The ICE feature in both Dimage Scan Utility and Silverfast seems to work well.
[a href=\"index.php?act=findpost&pid=128919\"][{POST_SNAPBACK}][/a]
Ray,
Whether the slide is faded or not I really believe is irrelevant to the merits of calibrating and profiling your scanner. If you characterize your scanner properly, once you make correct edits to image file data such that the faded colours return to their appropriate values, what you scan will look like the corrections you see. If you haven't done so, I suggest a read of Andrew Rodney's book, or less specialized - Tim Grey's "Color Confidence" would be helpful.
-
By the way, Vuescan's infra red scratch removal produces artifacts with Kodachromes using the Minolta 5400 II; specifically a tendency to slight double edges and blocky, deep shadows. The ICE feature in both Dimage Scan Utility and Silverfast seems to work well.
This topic is a little bit older, but perhaps someone looking for info (like me) reads it.
For the sake of completeness:
The SilverFast developers implemented a few new kodachrome features this year, escpecially:
the dust & scratch removal tool iSRD is working on kodachromes now.
And for calibrating the scanner they are offering special Kodachrome IT8 targets just since a few days.
http://www.silverfast.com/show/kodachrome-targets/en.html (http://www.silverfast.com/show/kodachrome-targets/en.html)
So this might be the easiest and best way to scan kodachrome slides, using SilverFast calibration target and software.
regards
Frank
-
Here's a 20 year old Kodachrome scanned in today with my Epson V700 flatbed and the basic Epson software
It seems about right to me!