Ad
Ad
Ad
Pages: [1]   Bottom of Page
Print
Author Topic: ACR and PS Process a Photo Differently - Why?  (Read 4830 times)
dmerger
Sr. Member
****
Offline Offline

Posts: 686


« on: December 11, 2010, 02:46:39 PM »
ReplyReply

I scanned a negative and in my scanning software selected aRGB color matching and saved it as a tiff, but the aRGB profile didnít get embedded.  I then made a copy of the file.  I opened the original in ACR, with aRGB output color space selected, but everything else at zero, i.e. no changes to the photo. From ACR I opened the photo in PS.  The photo opens in ACR and PS looking normal and the same in each.

I then opened the copy in PS without assigning any color profile.  The photo looked too red and saturated.  I then assigned aRGB.  The photo looked normal and similar to the original version processed with ACR, but the photos donít look the same. There is a very obvious difference in the photos. Why?

When I opened the original file in ACR, the photo didn't have a profile embedded. Why would the photo look fine in ACR and not red and over saturated like it looked in PS before I assigned aRGB?  Changing the output color space in ACR doesn't affect how the photo looks in ACR so it doesn't appear that ACR is using the output color space in determining how the photo should look in ACR.

 
Iím using CS-4 64 bit with Vista 64 bit and the Minolta scanning software for my Minolta 5400 scanner.
Logged

Dean Erger
digitaldog
Sr. Member
****
Online Online

Posts: 9004



WWW
« Reply #1 on: December 11, 2010, 02:55:44 PM »
ReplyReply

Pretty sure LR and ACR assume untagged docs are in sRGB. Bottom line is, donít deal with untagged files! Assign the profile and move on.

IF you bring a document into ACR or LR and then have those open into Photoshop, unless in LR you tell it ďopen originalĒ, its going to be processed through the ACR engine and while with no alterations of the settings, it should look the same, its still undergoing processing (in this example for no reason at all).
Logged

Andrew Rodney
Author ďColor Management for PhotographersĒ
http://digitaldog.net/
dmerger
Sr. Member
****
Offline Offline

Posts: 686


« Reply #2 on: December 11, 2010, 03:29:19 PM »
ReplyReply

Thanks, Andrew.  What you say makes a lot of sense and will be a big help. I had a feeling that somehow sRGB was being applied to my files somewhere but couldnít put my finger on it

I think I confirmed that ACR assumes that an untagged document is sRGB.  I compared the photo that I had originally opened in ACR with the copy opened in PS, but this time assigned sRGB.  Now the two look identical, although the ďmeanĒ and ďmedianĒ values in the histogram vary by 0.35 and 2, respectively. 

In addition, in PS I made a duplicate of the photo that I originally opened in PS and had assigned aRGB, then saved the duplicate, closed it in PS and opened it in ACR, then from ACR opened the photo in PS.  Now the photo originally opened in PS and the copy of it processed through ACR look identical, although the ďmeanĒ values in the histograms vary by 0.17. 

I donít know why the mean and median values vary, but the variance is so small as not to be noticeable.

Again Andrew, thanks.  Iíve been conducting a lot of tests to determine how to get the best out of my scanner when scanning negatives and I was getting some very unexpected, confusing results.  In one fell swoop, youíve cleared up a lot and made my life a lot easier. 
Logged

Dean Erger
dmerger
Sr. Member
****
Offline Offline

Posts: 686


« Reply #3 on: December 11, 2010, 03:52:44 PM »
ReplyReply

Another question.  My scanner software has a lot of options for color spaces.  The options are sRGB, Apple RGB, SMPTE-C, PAL/SECAM, ColorMatch RGB, aRGB, Wide-Gamut RGB, NTCS, CIE RGB and monitor RGB.   (Unfortunately ProPhoto is not an option and if I donít select color matching I suspect it defaults to sRGB.)  Which of the available options has the widest gamut?
 
My goal would be to scan with the widest color space I can, then in PS assign that profile and convert to ProPhoto.  Although if the widest gamut available is aRGB then I assume there would be no benefit to converting to ProPhoto, so Iíd stay with aRGB.

Also, how does the gamut of color negatives compare to aRGB or whatever would be the widest color space available with my scanner software?
Logged

Dean Erger
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6903


WWW
« Reply #4 on: December 11, 2010, 04:43:46 PM »
ReplyReply

Another question.  My scanner software has a lot of options for color spaces.  The options are sRGB, Apple RGB, SMPTE-C, PAL/SECAM, ColorMatch RGB, aRGB, Wide-Gamut RGB, NTCS, CIE RGB and monitor RGB.   (Unfortunately ProPhoto is not an option and if I donít select color matching I suspect it defaults to sRGB.)  Which of the available options has the widest gamut?
 
My goal would be to scan with the widest color space I can, then in PS assign that profile and convert to ProPhoto.  Although if the widest gamut available is aRGB then I assume there would be no benefit to converting to ProPhoto, so Iíd stay with aRGB.

Also, how does the gamut of color negatives compare to aRGB or whatever would be the widest color space available with my scanner software?


What scanner software and scanner are you using? Once I know that, I MAY have some workflow suggestions for colour-managing colour negative scans.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
dmerger
Sr. Member
****
Offline Offline

Posts: 686


« Reply #5 on: December 11, 2010, 05:45:36 PM »
ReplyReply

I using a Minolta 5400 (original version).  I'm doing tests with the Minolta software, Vuescan and ColorNeg.  I've also tried several methods for scanning negatives as positives and then removing the orange mask and inverting in PS.  So far, the Minolta software is giving me the best results.  By best results, I mean it gives me a photo with the most accurate color and the least color cast "from the get go." 

At first I was having problems with the Minolta software, but once I discovered that it is essential to select color matching in the Minolta software and then open in PS and assign the appropriate color profile, things improved a lot. (I was under the mistaken impression that I could scan a negative without selecting color matching and then assign the provided Minolta profile, but that profile is for positive film only.) The Minolta software is now giving me very good results. 

I've read a lot of good things about Vuescan but it isn't working very well for me.  In my tests I've consistently gotten rather strong color casts.  I've tried three types of film and have set Vuescan for the specific film types, but I still get strong color casts that would take a lot of time and effort to just come up to what I'm getting from the Minolta software without any tweaks.

Until recently I was scanning positive film almost exclusively.  The Minolta software does exactly what I want for positive film.  It outputs the unaltered raw linear data directly from the scanner CCD which I can then process in ACR and PS.  Itís not so easy with negative film, however.  I just couldnít get a good conversion of the raw linear data from a negative with the orange mask to a good positive.  Also, although ColorNeg is designed to convert the raw linear negative data into a positive, I havenít had any success with it.  So, I think Iíll have to go with scanner software to make the negative to positive conversion. 

I havenít tried Silverfast, yet.  Iím getting a little tired of all the testing and Iíve read that Silverfast has a steep learning curve.  So, Iím reluctant to try yet another piece of software and Iím now getting such nice results from my Minolta software that Iíve lost some of my incentive to try Silverfast.

Photos with a very wide dynamic range, such as snowy mountain scenes, still present problems.  Itís difficult to get a good scan exposure of both the bright parts and the mid level to shadows.  One ďtrickĒ Iíve been exploring with positive results is to do several scans with different exposures and then blending them in layers in PS.  The auto align function in PS now makes aligning the different scans easy.  Then I use the blend if sliders to seamlessly blend the various exposures.  It takes a little practice to get it right, but it looks very promising.

Iím new to scanning negatives, however, so Iíd welcome any suggestions for getting the best results.
Logged

Dean Erger
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6903


WWW
« Reply #6 on: December 11, 2010, 06:30:51 PM »
ReplyReply

I using a Minolta 5400 (original version).  I'm doing tests with the Minolta software, Vuescan and ColorNeg.  I've also tried several methods for scanning negatives as positives and then removing the orange mask and inverting in PS.  So far, the Minolta software is giving me the best results.  By best results, I mean it gives me a photo with the most accurate color and the least color cast "from the get go." 

At first I was having problems with the Minolta software, but once I discovered that it is essential to select color matching in the Minolta software and then open in PS and assign the appropriate color profile, things improved a lot. (I was under the mistaken impression that I could scan a negative without selecting color matching and then assign the provided Minolta profile, but that profile is for positive film only.) The Minolta software is now giving me very good results. 

I've read a lot of good things about Vuescan but it isn't working very well for me.  In my tests I've consistently gotten rather strong color casts.  I've tried three types of film and have set Vuescan for the specific film types, but I still get strong color casts that would take a lot of time and effort to just come up to what I'm getting from the Minolta software without any tweaks.

Until recently I was scanning positive film almost exclusively.  The Minolta software does exactly what I want for positive film.  It outputs the unaltered raw linear data directly from the scanner CCD which I can then process in ACR and PS.  Itís not so easy with negative film, however.  I just couldnít get a good conversion of the raw linear data from a negative with the orange mask to a good positive.  Also, although ColorNeg is designed to convert the raw linear negative data into a positive, I havenít had any success with it.  So, I think Iíll have to go with scanner software to make the negative to positive conversion. 

I havenít tried Silverfast, yet.  Iím getting a little tired of all the testing and Iíve read that Silverfast has a steep learning curve.  So, Iím reluctant to try yet another piece of software and Iím now getting such nice results from my Minolta software that Iíve lost some of my incentive to try Silverfast.

Photos with a very wide dynamic range, such as snowy mountain scenes, still present problems.  Itís difficult to get a good scan exposure of both the bright parts and the mid level to shadows.  One ďtrickĒ Iíve been exploring with positive results is to do several scans with different exposures and then blending them in layers in PS.  The auto align function in PS now makes aligning the different scans easy.  Then I use the blend if sliders to seamlessly blend the various exposures.  It takes a little practice to get it right, but it looks very promising.

Iím new to scanning negatives, however, so Iíd welcome any suggestions for getting the best results.


Hi Dean,

I've been through exactly everything you're going through with the same scanner and the same software options, but years ago, and have been evolving with it to the present. You can see my various articles about scanning with SilverFast on this website going back a number of years to very recently. I've done a mixture of negatives and positives, but more of the former actually.

There is no question in my mind that SilverFast has the best negative scanning solution of anything you'll find anywhere. With the extensive collection of negative presets they offer, as well as the ability to tweak each one of them to taste, you can customize a negative "profile" that will fit your needs exactly for every kind of negative image you have. For positives, SilverFast also allows you to profile the scanner properly, which is a big help, although you can use their canned profiles for this scanner quite successfully. (It is an excellent scanner, my only complaint being how slow it is.)

You can download a trial version of SilverFast, read my articles, read articles by Ian Lyons on using SIlverFast, and by all means download and read the essential sections of the SilverFast manual and you can get on top of it before the trial period ends, easily. It's interface is "special", but still not rocket science to wrap one's head around it. I recommend trying the Ai6 Studio version on a trial basis. One of the features in SilverFast you may appreciate for those high-dynamic range images is the Multi-Exposure feature. When selected, it makes one scan for the highlights and mid-tones and another for the shadows, thereby extending the DR of the scan and saving time in post-scan processing. I believe this feature is available for the Minolta 5400 Scan Elite version - that information is available on their website.

You asked about colour spaces. SilverFast (and I think the other programs should too) allows you to select whatever colour space is listed in the profiles stored on your hard drive for colour management. So if you have a ProPhoto working space on the hard-drive, you can select it. If you adjust your image and scan it with ProPhoto selected in the scanner software's colour management settings, and that working space is embedded in the scan, when you open it in Photoshop in Pro-Photo the colours should be the same or very close. If you scan in a narrower space and then open the image in Photoshop and use a wider space the results are rather unpredictable. If you scan in a wide space and then wish to compress it to a narrower space in Photoshop, Photoshop usually handles that compression quite well. In all cases, it is best to scan in 16-bit per channel (what SilverFast calls "48-bit color") because that gives you the most data for protecting the image from issues such as banding and posterization in post-scan processing.

Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
digitaldog
Sr. Member
****
Online Online

Posts: 9004



WWW
« Reply #7 on: December 12, 2010, 01:26:44 PM »
ReplyReply

Another question.  My scanner software has a lot of options for color spaces. 

Is there an option for the scannerís native color space? That be a good start.
Logged

Andrew Rodney
Author ďColor Management for PhotographersĒ
http://digitaldog.net/
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6903


WWW
« Reply #8 on: December 12, 2010, 02:17:24 PM »
ReplyReply

Andrew, is what you are mentioning here the procedure of embedding the scanner profile into the scan, then opening the scanned image in - say - Photoshop and converting to the selected Photoshop working space? This is a viable workflow too, given the usual proviso that one has a decent scanner profile and can scan the images under the same conditions for which the profile was created.
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
digitaldog
Sr. Member
****
Online Online

Posts: 9004



WWW
« Reply #9 on: December 12, 2010, 02:26:03 PM »
ReplyReply

Andrew, is what you are mentioning here the procedure of embedding the scanner profile into the scan, then opening the scanned image in - say - Photoshop and converting to the selected Photoshop working space? This is a viable workflow too, given the usual proviso that one has a decent scanner profile and can scan the images under the same conditions for which the profile was created.

You could convert to a working space or just leave it in the scanner color space. Its the original/native color space, its the ďbest fitĒ and its questionable if you need to convert to an RGB working space (adding yet another conversion). A neutral may not be R=G=B but otherwise, might as well leave it in the native space the scanner produced, assuming a good scanner profile is provided.
Logged

Andrew Rodney
Author ďColor Management for PhotographersĒ
http://digitaldog.net/
dmerger
Sr. Member
****
Offline Offline

Posts: 686


« Reply #10 on: December 12, 2010, 04:40:49 PM »
ReplyReply

I located a profile for my scanner for negatives and it appears to work fine, so Iíll take your advice and use it.   However, if I assign my scanner profile in PS and donít convert to another profile in PS, what will happen when I reopen the photo in ACR 5.7, make some edits and then output to PS?  I assume that ACR will be able to recognize and use my scanner profile, but for output from ACR I donít see an option to continue to use the input color profile.  I only see options for aRGB, ColorMatch, ProPhoto and sRGB.
Logged

Dean Erger
dmerger
Sr. Member
****
Offline Offline

Posts: 686


« Reply #11 on: December 12, 2010, 04:42:01 PM »
ReplyReply

A little off topic, but when I use my scanner profile for a photo and compare it to the same photo scanned with sRGB, aRGB and Wide Gamut, they all appear very similar, but not identical.  Based on one test, the best looking photo is the scanner profile.  No surprise here, but when I click on gamut warning I get just the opposite of what Iíd expect.  All show a little clipping in the reds, but the photo with the scanner profile shows the most clipping, followed by wide gamut, aRGB the sRGB.  I guess I donít understand what Iím seeing because I would have expected the opposite result. Where am I going wrong? 

On the other hand, in the blues, all four photos show almost the same amount of clipping, but the scanner profile version shows slightly less clipping.  Whatís going on?  (There is a very slight difference in the appearance of brightness in the photos.  Could this difference account for the gamut warning difference, but if so, why the opposite for red vs blue?)
Logged

Dean Erger
digitaldog
Sr. Member
****
Online Online

Posts: 9004



WWW
« Reply #12 on: December 12, 2010, 05:00:42 PM »
ReplyReply

I located a profile for my scanner for negatives and it appears to work fine, so Iíll take your advice and use it.   

Odd, Iíve never seen an ICC profile for neg. You sure its not for the transparency?
Logged

Andrew Rodney
Author ďColor Management for PhotographersĒ
http://digitaldog.net/
dmerger
Sr. Member
****
Offline Offline

Posts: 686


« Reply #13 on: December 12, 2010, 05:20:03 PM »
ReplyReply

Yes, Iím pretty sure.  When I installed the Minolta software, it automatically installed two profiles, one for normal positive scans and one for positive scans in 16 bit linear.  The ownerís manual describes these profiles.

When I looked at the Minolta program files on my computer, however, I found a third profile named ďNega InputĒ.  This profile is not mentioned in the ownerís manual nor was it installed automatically. I installed that profile and tried it on a negative scan.  It worked great in my one test. 
Logged

Dean Erger
dmerger
Sr. Member
****
Offline Offline

Posts: 686


« Reply #14 on: December 15, 2010, 12:04:57 PM »
ReplyReply

Iíve completed tests of scanning two negatives with the Minolta software, Vuescan, Silverfast and ColorNeg.  One negative is Fuji Reala and the other is Kodak Gold 100.

I used the Minolta profile for the scan with the Minolta software, sRGB for Vuescan (apparently the demo version only outputs sRGB), ProPhoto for Silverfast and no profile for ColorNeg (it uses a positive linear scan). Based on my tests using the Minolta software and various color spaces, using sRGB for Vuescan gave it a slight disadvantage but not nearly enough to change my rankings in the tests.

All scans were with auto exposure (which worked well and fairly consistently for all but one of the scans) and otherwise minimal adjustments.  No adjustments for the scan with the Minolta software.  I merely assigned the Minolta profile to the file when I opened it in PS.  For Vuescan, I selected the appropriate film type but made no other image adjustments.  For Silverfast, I selected the appropriate film type and made several scans.  I followed Mark Segalís tutorial, Ian Lyonsí tutorial and various combinations of auto color cast corrections and auto image adjustments. For ColorNeg, I selected the appropriate film type but made no other adjustments. I tried to be sure that none of the software was making any adjustments other than as described above.

My goal was to see how well the software made the basic conversion of the data from the scanner CCD into a positive image.  Of course, the software contain a lot of image adjustment options, but in my view the best software produces the most color accurate photo in the initial conversion.

For the Reala negative the clear winner is the Minolta software, with Vuescan a distant second and Silverfast a distant third.  (I didnít do a ColorNeg test for the Reala.)  The Minolta scan is superb. It needs a minor white balance correction in ACR and some contrast boost, but I canít imagine a much better scan with my scanner. The Vuescan colors are good but need a fair amount more saturation and there is haze to the scan, which I assume could be corrected with a combination of white balance, saturation and contrast boost. Iíd say that the Vuescan scan is a very good scan.  It would need a lot more post processing than the Minolta scan but I suspect it would turn out very well.  The best of the Silverfast scans is way over saturated, so much so that it is not possible to even judge the color accuracy. 

For the Kodak the results are much closer.  The auto exposure scan with the Minolta software was a little over exposed, so I made another scan with less exposure to more closely match the scans with the other software.  This second scan is very good and the best scan.  The Vuescan scan is a close second.  A good argument could be made that it is the best scan for post processing. It has the same characteristics as Vuescanís scan of the Reala.  The scans with Vuescan are very consistent.  I suspect that the Vuescan scan would shape up very well with some post processing. The best Silverfast scan is again over saturated but not nearly to the extent as the Reala scan. It also has a very noticeable magenta/red color cast. The ColorNeg scan has a magenta/red color cast, but the colors otherwise look to be good, just a little under saturated (which is not a bad result). My hunch is that the ColorNeg scan would be easier to correct in post processing than the Silverfast scan, but I havenít attempted any post processing.

Obviously, my test is limited to two photos and not extensive enough for anyone else to use as the final word on scanning software.  Iíve decided, however, to use with the Minolta software for all my negative scans. 

In case anyone else has the Minolta scanning software and wants to set it up as I did, in the software select auto exposure for negatives, 16 bit and no color matching.  Then find and install the Minolta profile as I described in an earlier post above.  Open your scan directly into PS and assign that profile, and probably convert to ProPhoto.  If you need to adjust the exposure, use only the master exposure slider.

One last note:  Iíve read a lot of good things about Silverfast. Iím surprised and a little puzzled by Silverfastís very poor showing in my tests.  I was very careful in following the methods described in Mark Segalís and Ian Lyonsí tutorials, and tried several variations, all without success.  Iím wondering if somehow, somewhere I had something set incorrectly. 
Logged

Dean Erger
FranciscoDisilvestro
Sr. Member
****
Offline Offline

Posts: 500


WWW
« Reply #15 on: December 15, 2010, 04:00:51 PM »
ReplyReply

ProPhoto for Silverfast ....

... The best Silverfast scan is again over saturated


The problem could be related to the choice of color space. When using the negafix utility in Silverfast to scan color negatives, the results seems to be designed for AdobeRGB.

There is an old thread in this forum that deals about this

http://www.luminous-landscape.com/forum/index.php?topic=9158.0

It seems that the way color management work when scanning color negative film with Silverfast + Negafix is by "Assigning" the profile and not "Converting" to the profile.

I suggest you repeat the scans with Silverfast choosing AdobeRGB as output color space.
Logged

dmerger
Sr. Member
****
Offline Offline

Posts: 686


« Reply #16 on: December 15, 2010, 10:02:54 PM »
ReplyReply

So this is a known problem that has been around for over five years and Silverfast still hasnít corrected it? 

From what I gather from the thread you linked, Francisco, I really donít need to do another scan but just assign aRGB to the scans I already made. I did so and saw a major improvement in the Silverfast scans.  However, I also tried sRGB on the Reala scan and got an even better result. Iím not sure what Silverfast is doing, but based on my little test and what was stated in the thread you linked, Silverfast seems to have a unique color management system. (Perhaps Iím being overly generous in calling it color management?)

For the Reala, although the Silverfast aRGB scan is much improved, itís still a good deal over saturated, although not nearly as much the ProPhoto scan. The saturation in the Silverfast sRGB scan is good. The saturation and colors in the Minolta scan, however, are much better than in the other scans. So, after this new test, Iíd say that the Minolta scan is still easily the best scan, with Silverfast sRGB a distant second, followed by Vuescan, Silverfast aRGB and Silverfast ProPhoto.

For the Kodak, the Silverfast aRGB is better than the Silverfast sRGB.  I also tried assigning different color profiles to the ColorNeg scan and it too showed a large improvement by assigning aRGB to its scan.  The results of this new test are as follows: Minolta clearly still the best scan, followed by Silverfast aRGB, ColorNeg and Vuescan.

In my view, scan results are far and away the most important characteristic for evaluating scanning software.  Nevertheless, my ranking for (a) the most logical, well designed and easy to understand and use software and (b) color management are as follows:  Minolta far and away the best, followed by Vuescan, with ColorNeg and Silverfast lagging far behind.  One caveat about the Minolta software is that the ownerís manual doesnít adequately explain its superior color management system, but itís very easy to learn and super easy to use correctly.
Logged

Dean Erger
FranciscoDisilvestro
Sr. Member
****
Offline Offline

Posts: 500


WWW
« Reply #17 on: December 16, 2010, 04:56:28 AM »
ReplyReply

I donīt think it is a problem. It is  poorly explained in the documentation. As it was mentioned before, there is no input profile for color negatives (I donīt know what the Minolta software is doing or how they handled it)

The common practice then is to scan the negative, assign a profile and then tweak it until you get the desired results. Negafix is supposed to do this "tweak" and get close to the final result. As far as color profiles, most of the users think (from posts in this and other forums) that Negafix tweaks are based on AdobeRGB.

In my experience, you have to try different options, even different negative profiles, before getting satisfactory results
Logged

dmerger
Sr. Member
****
Offline Offline

Posts: 686


« Reply #18 on: December 16, 2010, 01:14:15 PM »
ReplyReply

Iíve made a couple of mistakes in my posts above, maybe more but two that I know of.  I said that I suspected that if I didnít select color matching in the Minolta software that it was defaulting to sRGB.  Wrong.  I said that the gamut warnings were opposite of what I expected.  Wrong again.  In fact, these two misconceptions are related. 

The reason I suspected that the Minolta software would default to sRGB is because that is what ACR did when I opened an untagged file.  So yes, my suspicion that sRGB was somehow creeping into my tests was correct, but it wasnít because the Minolta software was using it. 

In addition, the gamut warning is actually confirming that the native color space of my scanner has the widest gamut, followed by Wide-Gamut RGB, aRGB and sRGB.  In PS, I had incorrectly configured my proof setup and thus my gamut warnings were all wrong.  After correctly configuring my proof setup, the gamut warnings are exactly as expected.  For example, it shows that certain colors in the Minolta native color space are out of gamut with respect to sRGB, aRGB and even Wide-Gamut RGB. 

Francisco, maybe it isnít fair to call what Silverfast is doing a problem, but it sure is different from what the Minolta software is doing and what I understand to be sound color management.  Based on the linked posts, it appears that when you select a color space in Silverfast, it merely assigns that color space to your photo, but doesnít attempt to make any conversion or compensation for the different color spaces.  It treats them all the same.  Treating all color spaces the same isnít sound color management in my opinion.

On the other hand, the Minolta software clearly treats each color space differently and converts or compensates in some manner for whatever color space you select.  With the Minolta software, when I select the native color space, Wide-Gamut RGB, aRGB or sRGB, all the scans look very good.  None show the oversaturation like Silverfast.  The different color spaces show up as slightly different looking photos, but they are all very similar and all look good.  Plus, my description of the gamut warning above seems to confirm that the Minolta software is treating the different color spaces in the manner I would expect from software with sound color management.

When I started scanning negatives I wasnít happy with the results from the Minolta software.  Based on my initial results and on what Iíd read, I thought that the Minolta software just wasnít very good for scanning negatives.  This experience led me to test Silverfast, Vuescan and ColorNeg.  I was hoping to find better software.  Instead, this exercise has taught me how to use the Minolta software correctly and shown me that it is clearly the best software to use for scanning negatives.  In other words, I started out with a low opinion of the Minolta software and came away with admiration for the people at Minolta who designed some terrific software.   The people who wrote the ownerís manual, however, could have done a better job describing the color management of the software.

Andrew, Francisco and Mark, thanks for your comments in this thread.  Without your comments I wouldnít have discovered so much about the various scanning software and especially the Minolta software.   
Logged

Dean Erger
Mark D Segal
Contributor
Sr. Member
*
Offline Offline

Posts: 6903


WWW
« Reply #19 on: December 17, 2010, 05:54:50 PM »
ReplyReply

There is nothing "unsound" about colour management in SilverFast and its conversion process from negative to positive is the most sophisticated apparatus on the market, but (i) you need to know its strengths and limitations and (ii) you need to know how to use it.

Let's start with the strengths and limitations. The main strength in this area is that it is the only scanner software on the market which provides negative to positive conversion algorithms which are CUSTOM-MADE for every film type, sub-type and ASA speed AND all these for every film-capable scanner for which a version of SilverFast is sold - a huge number of custom "profiles", considered important for optimal conversion quality because the tonal characteristics of each scanner and film type/sub-type/speed are different. It makes for a huge number of "profiles", but LaserSoft Imaging provides them. The main limitation of this approach is that all of these "profiles" were designed for the ARGB(98) colour space. This means that outcomes will likely be more accurate and predictable if you convert negatives to positives using the working space for which the "profiles" were designed - ARGB(98). However, one can very successfully convert many images using ProPhoto colour working space and the results will also be fine. However, for certain kinds of images, the remapping from an ARGB(98)-based "profile" to a ProPhoto RGB working space can produce some over-saturation. The way to avoid this from happening is to do the negative to positive conversion at the scan stage using a good Negafix "profile", and embedding (i.e. assigning) the ARGB(98) workspace (you do this in Options>CMS) to the image; then scan it and open the image in Photoshop; if you wish to work this image in ProPhoto thenceforward, convert the file to ProPhoto. The file numbers will change but the image appearance will remain identical.

Turning to knowing how to use it: the main thing here is to get the best Negafix "profile" fit to the film stock you are scanning. As I mentioned above, for every relevant scanner, the Negafix tool has a very large number of canned conversion "profiles" which you can test, obviously starting with the one made for the film you are using if you know for sure what film it is (often one doesn't, because of the age of the materials, re-branding of film, etc.); if the resulting display image is not correct as you perceive it, you can select the Negafix "profile" which has the best fit of all you tried, and then tweak it in great detail - especially if you use the "Expert" mode for the Negafix tool. Once you have the look you like for a normal. representative image, that tweaked profile can remain available for all other scans of the same film stock.

The fidelity of the resulting image appearance from SilverFast to Photoshop is very good.

In light of a number of comments in this thread, just to make sure I wasn't dreaming in SilverFastland, I fired-up my Nikon Super Coolscan 5000-ED and made several test scans of two image types (low saturation and higher saturation) and two working spaces for the latter, in order to see whether I could reproduce unsatisfactory results by doing things the right way; that didn't happen. All of it produced very satisfactory tonal and colour results, albeit, as one would expect, the appearance of a ProPhoto scan and an ARGB(98) scan of the same image with the same settings (apart from the colour space) is not the same. I did nothing to the files beyond the scan stage except to blank out the license number on an automobile for privacy reasons.

I have up-loaded to a server full TIFF scans as well as some JPG screen grabs comparing the results in SilverFast and Photoshop. These are large files, the TIFFS being 126~165 MB each and the JPEGs in the range of 2.6~4.7 MB each. One of the screen-grab comparisons is done at 100% magnification to show comparative detail retention for the saturated image in the two colour working spaces. It appears to me satisfactory in both, considering that this is a very large magnification of colour negative film which has not been treated for grain and not sharpened. For those interested, I am willing to provide download links allowing you to examine these files in detail on your own computers, but this will be only on request (contact me via PM from this Forum's messaging facility) with the provision of your full real name, your Forum screen name if different and valid private email address. The files are proprietary, copyrighted and provided for your personal inspection only. By downloading the files you are agreeing irrevocably to these conditions. Between the file names and the labeling on the images, along with the information given here, these images should be self-explanatory. Any questions or discussion about the images should be posted in this thread, as I am not likely to sustain a private email correspondence beyond providing the download link.

Regards,

Mark
Logged

Mark D Segal (formerly MarkDS)
Author: "Scanning Workflows with SilverFast 8....." http://www.luminous-landscape.com/reviews/film/scanning_workflows_with_silverfast_8.shtml
Pages: [1]   Top of Page
Print
Jump to:  

Ad
Ad
Ad