Silverfast Advanced Targets ?

[David Mantripp]

Recently Lasersoft has been hawking around these "Silverfast Advanced Targets (Part 2)", which they claim:

"With our newly designed Part 2 Targets, users receive three times more measuring fields, while even more precise measurements provide a significantly larger color spectrum. The color profile is further supported by additional dark, grey and pastel toned color fields, which are supported by significantly truer measured values."

Do any colour gurus out there (a) have any idea what Lasersoft are talking about, and (b), if so, is there any tangible benefit likely over standard IT8 calibration ?

[digitaldog]

Recently Lasersoft has been hawking around these "Silverfast Advanced Targets (Part 2)", which they claim:

"With our newly designed Part 2 Targets, users receive three times more measuring fields, while even more precise measurements provide a significantly larger color spectrum. The color profile is further supported by additional dark, grey and pastel toned color fields, which are supported by significantly truer measured values."

Do any colour gurus out there (a) have any idea what Lasersoft are talking about, and (b), if so, is there any tangible benefit likely over standard IT8 calibration ?

Hawking? Not really, it's not their invention; it's based on a new spec (ISO 12641-2). But they do offer it. And yes, I've tested it.

Very initial impressions of two targets used to build a reflective scanner profile (Epson V750).
One made with the IT8 original from LaserSof: IT8.7/2-1993 2015:03
One made with the new target: ISO:12641-2 2017:11


The newer target has a larger color gamut and gamut volume which is to be expected. But this expectation is good! The gamut of a profile cannot be any larger than the target used to build it and thus, this new target is a major improvement there compared to the original as seen below. Now what I'm still testing (or waiting to hear back from LaserSoft) is why using the two profiles produces so little actual difference in two scans of my Gamut Test File printed on a 3880. I'd have expected the wider gamut input profile would produce a wider gamut scan and I'm not sure why the deltaE differences of the two are so small.
--------------------------------
dE Report


Number of Samples: 208500

Delta-E Formula dE2000

Overall - (208500 colors)
--------------------------------------------------
Average dE:   0.57
    Max dE:   6.01
    Min dE:   0.00
 StdDev dE:   0.51


Best 90% - (187649 colors)
--------------------------------------------------
Average dE:   0.44
    Max dE:   1.21
    Min dE:   0.00
 StdDev dE:   0.30


Worst 10% - (20851 colors)
--------------------------------------------------
Average dE:   1.73
    Max dE:   6.01
    Min dE:   1.21
 StdDev dE:   0.52


--------------------------------------------------
--------------------------------------------------
What you see is the deltaE differences in the two are tiny and largely invisible. I'd expect that colors that fall outside the gamut of the 'standard' scanner profile would show larger dE values than using the Advanced profile. Note I'm using Standard and Advanced to name the two targets and thus the profiles they create.

[David Mantripp]

Thanks for the detailed reply Andrew.  By "hawking" I guess I was referring to the calibration targets they're pushing, not the standard.

If I understand correctly, the older standard targets may have a narrower gamut than the hardware - since anyway that is the physical limit - but the new ones _might_ be able to better exploit the capabilities of the scanner.  Or maybe not...

[digitaldog]

If I understand correctly, the older standard targets may have a narrower gamut than the hardware - since anyway that is the physical limit - but the new ones _might_ be able to better exploit the capabilities of the scanner.  Or maybe not...

There's no question the new target has a wider color gamut. The profiles show us that fact. Now why I don't see a difference in the scans of a wide gamut print I can't explain but I'm far, far from an expert on using SliverFast. Mark may have some comments. I've tried, using differing CMS settings in the software to run some initial tests and the data below was scanned such that the image was provided in the scanner color space, not converted to a working space. My initial concept, which may be wrong (again, waiting on LaserSoft) is this would provide a scan to evaluate with the least effect on the scanned data. Yet with the two profiles, the deltaE differences are small. But again, there's no question in my mind that the two resulting input profiles provide a difference in color gamut resulting from the differences in the targets. How that affects the resulting scan when the rubber meets the road is still questionable for me.

[Mark D Segal]

Very useful insights Andrew. Please keep us advised of any feedback you get from LSI on the questions you put to them. I just received copies of these targets the day before yesterday and intend to get into some testing in the near future. I'll be reporting results as and when - could be a while, as some stuff is in the queue ahead.

[David Mantripp]

Well, Andrew, you are now being quoted on the product page over at Silverfast, so with such an impressive endorsement, I have now ordered my own advanced target to play with  .

Incidentally, while I have the attention of the experts, does anybody know what on earth the IT8 cal button in Silverfast HDR is supposed to do ?  It doesn't make sense to me.... what would I be calibrating ?

[digitaldog]

Well, Andrew, you are now being quoted on the product page over at Silverfast, so with such an impressive endorsement, I have now ordered my own advanced target to play with  .

Yeah, but they never got back to me as to WHY the two profiles behave the same. I even sent them the URL for this series of posts. So I'm not sure I'm happy about all this or not. Again, the profiles do have a wider color gamut. The scans do not. A tad troubling.

[Mark D Segal]

Hi David,

As I mentioned, I shall be testing these targets in the near future. I received them a couple of days ago.

Your question about why the IT8 Cal in the HDR application? First point, just to be picky and sticky, I'm not keen on using the word "calibration" to describe this procedure. It is primarily profiling, of which calibrating certain key parameters to put the scanner into a known state should be an internalized first step. At first blush it would appear to make no sense because HDR is not a scanning application and this process is meant for profiling scanners. So we'll talk about profiling.

The option to create profiles did not exist in the first round of the HDR 8 application, but has been added since we published the book. On pages 290-293 of my SilverFast book, I discuss the need for coherent colour management (profiles and working spaces) between HDR and Photoshop or Lightroom for scanned photos being working between these applications. The same principle would also apply to working scanned photos coming say from third parties with different scanners. In the SilverFast Forum, LSI once explained as follows:

<<The purpose of the IT8 calibration in HDR is exactly the same as in the scan software, create a calibrated ICC profile (for whatever your source is).

<<Imagine you have a friend constantly sending you pictures for editing. He/She scans with different scanners. To obtain a solid color reproduction, your friend sends you a raw scan (48 bit) of his/her IT8 target appart from the scanned pictures, in that way you can create a calibrated profile which you will be using for the editing the rest of the pictures.>>

(I.E. he means you would create this profile in the HDR application using the Auto IT8 function provided there, whose purpose you are asking about.)

[David Mantripp]

Thank you Mark - that makes sense.  After many, many years of (generally happily) using Silverfast I still now and then get a sinking feeling that maybe I‘m doing it all wrong...

(I did do a search on my HDR profiling question, but nothing came up. I should have searched the forum. I do wish they’d open up the Silverfast forum a little, maybe even (gasp) allow search engines to index it. Still, I guess eccentricity is core to Lasersoft‘s corporate identity)

[David Mantripp]

Yeah, but they never got back to me as to WHY the two profiles behave the same. I even sent them the URL for this series of posts. So I'm not sure I'm happy about all this or not. Again, the profiles do have a wider color gamut. The scans do not. A tad troubling.

I guess this is a really stupid question, but is it at all possible that the software is tagging the advanced scan with the basic profile? Or scrambling the metadata somehow? (Note I have not even started to research how profiles are „embedded“).  It wouldn’t be the first time they released a version with colour management bugs, especially as related to profiling.

[Mark D Segal]

You can join their Forum free of charge. It can be useful!

[Mark D Segal]

I guess this is a really stupid question, but is it at all possible that the software is tagging the advanced scan with the basic profile? Or scrambling the metadata somehow? (Note I have not even started to research how profiles are „embedded“).  It wouldn’t be the first time they released a version with colour management bugs, especially as related to profiling.

I'll be testing all that anon.

[digitaldog]

I guess this is a really stupid question, but is it at all possible that the software is tagging the advanced scan with the basic profile?

It's tagging with each profile which is what's so odd.

[David Mantripp]

You can join their Forum free of charge. It can be useful!

What I mean is that they could stop requiring that every single post and answer needs moderator approval, which often takes days, and quite often for strange policy reasons are not approved.  This completely strangles user to user debate.  The forum could actually be a strong indirect marketing tool for them, if they allow it to become a centre of user to user excellence on scanning. But they won't.  I honestly cannot think of any other consumer software support form that is managed like this in 2018...

The frustrating thing is that Lasersoft staff are very engaged in the forum, are helpful and patient, and present a very positive image. And their direct customer support (accessed from within the software) is also way above industry standards.

I'm not knocking Lasersoft/Silverfast, I'm a satisfied and enthusiastic customer, but I want them to stick around, and have a healthy business, and I'm convinced that their very, er, unique approach to marketing and customer engagement is not doing them any favours...

[ColourPhil]

You can join their Forum free of charge. It can be useful!

Hi Mark,
Their forum was of great help to me, as was their direct support, with what turned out to be a scanner fault (ICE hardware), but it seems to be no longer in use?
Also found your Silverfast book to be invaluable, even though I'm considered a bit of a scanning expert!
Cheers,
Phil

[Mark D Segal]

Hi Mark,
Their forum was of great help to me, as was their direct support, with what turned out to be a scanner fault (ICE hardware), but it seems to be no longer in use?
Also found your Silverfast book to be invaluable, even though I'm considered a bit of a scanning expert!
Cheers,
Phil

Hi Phil, many thanks - much appreciated.

ICE is software that depends on the scanner's ability to scan for an infrared channel, which is a software directed hardware function. So if you say you were having a hardware problem it was with the infrared, not likely the software. LaserSoft Imaging has not used ICE ever since I started using SilverFast back in the late 1990s. Their iSRD tool is far superior. That tool is still there, as you know, for working with scanner-enabled infrared to do the scratch and dirt detection. But LSI has developed and bundled a new tool - SRDx, which works on the same principles as iSRD but does not need an infrared channel for defect detection. This makes it more usable for Kodachrome and usable for cleaning-up B&W media, which iSRD with infrared could not do.

[ColourPhil]

Hi Phil, many thanks - much appreciated.

ICE is software that depends on the scanner's ability to scan for an infrared channel, which is a software directed hardware function. So if you say you were having a hardware problem it was with the infrared, not likely the software. LaserSoft Imaging has not used ICE ever since I started using SilverFast back in the late 1990s. Their iSRD tool is far superior. That tool is still there, as you know, for working with scanner-enabled infrared to do the scratch and dirt detection. But LSI has developed and bundled a new tool - SRDx, which works on the same principles as iSRD but does not need an infrared channel for defect detection. This makes it more usable for Kodachrome and usable for cleaning-up B&W media, which iSRD with infrared could not do.

Sorry Mark,
I meant IR, as it was def. hardware and was fixed by the manufacturer/importer Reflecta under warrantee.
Cheers,
Phil

[Mark D Segal]

Fair enough, thanks for clarifying.

[David Mantripp]

LaserSoft Imaging has not used ICE ever since I started using SilverFast back in the late 1990s.

I can't resist a touch of pedantry here... Silverfast v6 certainly used, or enabled the use of, ICE, for compatible scanners, such as the Minolta DualScan MultiPro.  To quote Alejandro Morales at Lasersoft:

"ICE is a patented technology developed by Kodak. In previous versions of SilverFast some scanners had ICE included because the scanner manufacturers had special licensing to leave ICE as it was.
This licences do not have effect anymore in SilverFast 8, hence we could implement our own infrared dust and scratch correction technology iSRD."

iSRD has better detection than ICE, and far more controls (arguably ICE didn't need them - I never saw the sometimes crazy alignment problem that iSRD has with ICE). But the patching that iSRD applies is truly abysmal.  Just a horrible, sharp edged, super coarse blur.  I've no idea why they don't use the same algorithm as the Clone/Heal tool, which works quite well.

[Mark D Segal]

I can't resist a touch of pedantry here... Silverfast v6 certainly used, or enabled the use of, ICE, for compatible scanners, such as the Minolta DualScan MultiPro.  To quote Alejandro Morales at Lasersoft:

"ICE is a patented technology developed by Kodak. In previous versions of SilverFast some scanners had ICE included because the scanner manufacturers had special licensing to leave ICE as it was.
This licences do not have effect anymore in SilverFast 8, hence we could implement our own infrared dust and scratch correction technology iSRD."

iSRD has better detection than ICE, and far more controls (arguably ICE didn't need them - I never saw the sometimes crazy alignment problem that iSRD has with ICE). But the patching that iSRD applies is truly abysmal.  Just a horrible, sharp edged, super coarse blur.  I've no idea why they don't use the same algorithm as the Clone/Heal tool, which works quite well.

I don't know whether it enabled the use of ICE, but it didn't bundle that tool.

iSRD is a detection tool. Once the defect is detected, it does use algorithms similar to clone/heal for fixing them. I have not experienced anything like the "horrible. sharp edged super coarse blur" you mention. Of course the success of using this tool, like for most others, depends on the tool settings. I have always found iSRD detection and subsequent clean-up to be amongst the more successful aspects of this software. It provides much more user control than ICE ever did, which is one of the things I like about it.