Just some stuff about photography

BY TAG

Isole Borromée

in Photography , Thursday, June 15, 2017

A couple of days ago I discovered on my desk a couple of sleeves of 120 film. These turned out to be from a small set I made nearly 2 years ago in the Borromean Islands in Lago Maggiore.  They are all 6x7 shots taken on Kodak Portra 400 (it’s what all the cool kids use, you know) using the Voigtlander Bessa III 667 (probably the best fixed lens medium format camera ever made - certainly the last, along with its 667w close relative).

drm_B667_Oct15_11_07.jpg
drm_B667_Oct15_13_06.jpg
drm_B667_Oct15_13_05.jpg
drm_B667_Oct15_11_05.jpg
drm_B667_Oct15_11_06.jpg

Portra 400 - photography’s answer to Dad Dancing.

Posted in category "Photography" on Thursday, June 15, 2017 at 04:25 PM

How sharp is your scan ?

in Scanning , Wednesday, May 03, 2017

The object of this review is perhaps not the most enthralling I’ve ever covered. It is, to be exact, an exposed piece of film, a piece of positive “slide” Agfa film in fact. It isn’t even in colour. It is in fact a Resolution Target (USAF 1951), produced by LaserSoft Imaging. You can read all about it here.

There was a special off on these a few weeks back - they quite often happen - and so it was quite a handy retail therapy opportunity. But apart from feeding the inner consumer, you may well ask “what is the point” ?

Resolution target

Resolution Target (USAF 1951)

Well, there are two points. The first is perhaps the most initially compelling, as it gives you a standard reference target with which, following the detailed instructions, you can determine the actual optical resolution of your scanner. This then gives a warm fuzzy feeling and bragging rights over all your scanner owning friends (or of course it might leave you feeling extremely depressed).

Unless you own hundreds of scanners, that’s pretty much the end of it. I own two, so I was kept happily entertained for over 5 minutes!  But, actually, it isn’t the end of it all, because as the instructions explain, there is something else you can use this target for, which turns out to be very enlightening.

The second use is to experiment with different scanner sharpening settings to determine what the optimum settings are. This was quite an eye-opener. I’d always tended to leave sharpening to Photoshop, and some mumbo-jumbo laden plug-in, as I’m too lazy or stupid to really get the details of all this radius, strength and masking stuff. Using the target in conjunction with Silverfast’s USM tool allowed me to quickly determine the best settings for restoring sharpness lost in the scanning process, without overdoing things. It also revealed that the adaptive “Auto Sharpness” mode in USM actually does a pretty good job.  By way of illustration, if I used USM, I tended (for no really good reason) to keep Power in a range of around 100-150. Auto Sharpness, for a full resolution positive 6x9 MF scan set Power at 300 - it goes up to 500. Zooming in at 1:1, it was clear that 100 was too weak, and 300 was not introducing any artefacts. I don’t know yet if this means I’ll switch to Silverfast USM for capture sharpening, but this experience certainly boosts my confidence in it, and it would be quite a time saver over doing it in Photoshop.

So, if you are interested, one, in seeing how much resolution your scanner actually delivers, and two, making the most of that resolution, this Silverfast Resolution Target (USAF 1951) is a good investment. I’m sure it would be equally useful for Vuescan users, there is nothing that ties it to Silverfast software.

Oh, and my results ? 4096 dpi real optical resolution for the Plustek Opticfilm 120, which while less than the advertised 5300 dpi, is pretty much as good as it gets for desktop film scanners. The revered Nikon Coolscan 8000 / 9000 reputedly do hit their advertised 4000 dpi. And apparently, the Opticfilm 120 could do even better with some way of fine tuning focus.  Anyway, 4096 dpi is quite enough for 35mm and 120 film, and this test confirms what I could see subjectively, comparing with scans from my old Minolta Dual Scan Multi Pro.

My Canon 9000F only managed 2048 dpi, less than half the advertised 4800, but that’s pretty much what I expected.

Posted in category "Scanning" on Wednesday, May 03, 2017 at 10:34 PM

A roll of CineStill

in Film , Wednesday, March 29, 2017

I'm quite impressed with my first experience with CineStill 50 film. As promised, it is very fine grained, and allows for very sharp results, provided of course that operator issues such as focussing and scanning are carried out correctly. The exposure latitude also seems very good, probably quite similar to Portra 400. The character of the photos is interesting. More saturated than Portra, certainly, but not excessively so like Ektar.

xpan_cinestill1_14.jpg


Of course, at 50 ISO, even over-exposing by two stops, considering that the fastest XPan lenses only open up to f/4, hand-held it is strictly a bright daylight film.

This first roll is really pretty much throw away, just trying it out, and I had it developed by a 1 hour lab which does ok, but has no packaging for uncut film, so it ends up scratched and dirty.

And my somewhat interrupted love/hate relationship with Silverfast, and indeed Silverfast's makers, has resumed. Silverfast has had some more half-baked or oddball features added, but major issues remain (for example, why does it not cache iSDR results ? Why recalculate and reapply every single time, even if I just change the display type from "Corrected" to "Original" ? Why can I not add extra frames for batch scanning on my scanner ? And why, for heaven's sake, is their idea of a forum so unbelievably user-hostile ? I don't suppose we will ever know.

Anyway, here's a few more CineStill 50 shots. Up until now I'm using Negafix standard settings and correcting grey balance in Silverfast - one of the things it does exceptionally well. This grey balancing might be actually masking some special attribute of CineStill 50, but I'll think about that later.

xpan_cinestill1_07.jpg


xpan_cinestill1_08.jpg


xpan_cinestill1_02.jpg


xpan_cinestill1_16.hdr.jpg

Posted in category "Film" on Wednesday, March 29, 2017 at 01:22 PM

Silverfast 8.8 revisited

in Silverfast , Monday, January 04, 2016

It’s timely that I set out my motivations for writing this blog. Quite simply, it’s personal. I have nothing to sell, I’m not harvesting clicks, I’m not looking to grow a big audience or profile. Yes, to some extent I’d like to draw attention to my photography, despite having long ago realised firstly that this is futile, and secondly, being dependent on recognition is not a path to happiness or fulfilment. But I have no illusions or expectations that my opinions on anything amount to much, or provide anything other than very mild entertainment. I like to engage with readers, and try to answer queries and questions as best I can (most of these come via email, as a rule people don’t like commenting or debating in the public arena).

So, why is this timely? Well, I’ve been taken to task in quite a major way by a representative from Lasersoft for my last, rather unappreciative, posting on Silverfast. Since that also was via email, I will not go into details here, apart from to comment on the fact that posting it as a comment would have been far more beneficial to Lasersoft, as it shows that the company both listens and cares, and is hurt by what it sees, probably correctly, as unfair criticism. But it also reinforces that their handling of customer relations via social media needs improvement, and does their image and probably their sales a lot less good than it could.

But I will readdress my criticisms of Silverfast 8.8, and try to correct some inaccuracies.  First of all, it is again important that I set out my ground rules here. I am a paying customer of Silverfast. I have no special relationship with the company, indeed no relationship at all beyond being a customer. I have been using Silverfast for around 15 years, on Minolta, Canon and Plustek scanners, as well as the HDR version which reprocesses generic Silverfast scan files. Film photography is still important to me, and I scan standard 35mm, “panoramic” 35mm, and a range of medium format from 6x6 to 6x12. Silverfast is therefore “mission-critical” to what I do. I pay for it - and it isn’t cheap - and I expect it to work as advertised. When it works nominally, I’m very happy. When it does stuff that impresses me, I praise it. When it lets me down - which is not often - I get annoyed. When it lets me down badly, as it did recently, I get angry. And then I turn to Vuescan, and 5 minutes later, decide to give up film photography… (I know that Vuescan works for some people, but it doesn’t for me. Believe me, I’ve tried).

So, my position is that Lasersoft has a duty to me, and to other paying customers, to provide reliable software which performs the functions it is advertised to provide. In return, if I choose to write about Silverfast, I should do so fairly, and with attention to factual accuracy. Possibly neither party here has quite achieved this is the last iteration.

There is very, very little written about Silverfast on the web these days, at least not in English. Actually there’s not a great deal in Germain either. And a large proportion of that little amount is either recycled PR or individuals declaiming that “Silverfast sucks”, that it has “the worst GUI”, that it is “the worst software ever”, bla bla etc etc. None of these people have read the manual of course. Or, rather, they would not have read it if one existed. Which it doesn’t, still, for version 8, and this is a point that Lasersoft badly needs to address seriously. So, since I do write about the software and the process from time to time, reasonably fairly, and quite positively, I suppose it isn’t that surprising I get some traction on the topic, and very occasionally, some reaction from Lasersoft.

So let’s get back to 8.8. My basic problems seem to be associated with side effects of the introduction of SRDx. My rather rude dismissal of SRDx was more to do with the fact that it seemed to be destroying my workflow, rather than any actual evaluation. In fact I still haven’t evaluated it. But eventually I did what probably I should have done to start with, and used the “Reset Software” feature in both AI Studio and HDR Studio. This seemed to clear up some issues - although unfortunately it also appears to delete all of my stored presets - and I was able to complete the 64bit HDRi scans of the batch of 6x7 negatives I was working with. I ran out of time for running them through HDRi, though, although I did process one sample with SRD turned off. I’m not yet sure about the other issues I noticed, in particular the “black screen” where Silverfast seems to get confused about previews. This needs some dedicated time in order to try to replicate and report to user support.

I do retract the derogatory comments I made about the ColorServer feature, which in any case I’m in no position to evaluate. Actually I’m rather reassured that there is still professional customer demand for such a high end pro feature. And of course I made the classic mistake of assuming that since I could not, immediately, see any reason for using SRDx for digital camera file processing, then neither could anyone else. All of that was down to frustration arising from my own specific issues rather than any real complaint.

I think it is worth remembering that Silverfast has some absolutely world-class features. Just a few examples include the one-click scanner calibration, the 4-target neutral grey-balance tool, and the selective and global colour correction, which are at least on a par with recent and much lauded counterparts in CaptureOne v9. These together allow to arrive at an ideally balanced scan in way, way less time than it would take in Photoshop, even if you could work out how to do it there. Some other favourites of mine are the simultaneous, live input/output histograms and the Kodak Portra 400 NegaFix profile, which I find to be remarkably accurate. And in more recent releases, the greatly improved Job Manager, and especially the copy/paste settings feature, has saved me a lot of time.

There are areas that could be improved though, and a certain tendency towards UI bloopers which could be avoided with a little care and attention.  Simply as an example, I’m going to come back to the SRDx / iSRD issue.

When newly discovering v8.8, the first thing I’m going to look for is the banner feature, SRDx. But it isn’t there. In fact, for my Plustek OpticFilm 120 scanner, it’s hiding behind the iSRD button. Clicking on that reveals the Dust & Scratches panel, with two tabs - SRDx, and iSRD.

SilverFast SRD1

Initially SRDx is on top, which again, is a little weird as I’d clicked on the iSRD button. This then starts to make me wonder - is SRDx actually being promoted as superior to iSRD in all situations, including those for which IR Dust & Scratch removal works ? Is SRDx an alternative, or a complement, i.e can they work together ... or indeed should they? None of this is very clear from the interface. The mechanism that Silverfast uses in general to turn off an effect or tool is a checkbox in the title bar of the tool. It seems in this case that switching tabs activates the visible SRD option and deactivates the other, but to my mind this is weak feedback. Of course, ironically this is nothing new - the precursor to SRDx, plain vanilla SRD, lived in the same spot. But as it didn’t interest me, and was never visible by default, I never noticed.

So my conclusion must be that yes, the tabs are intended also as on/off switches. This goes against generally understood usage of tabs - as for example evident in Silverfish’s preferences popup - so I don’t think it is a good idea.

But in fact my attention was drawn to this simply because after playing around with these features a bit, I realised that the real time preview wasn’t working any more, for any tool, for example gradation. Otherwise I’d probably never have noticed!

Note, when the source file does not have a recognizable IR channel, or, I assume, the scanner does not have an IR channel, you get an SRDx button instead of iSRD…

SilverFast SRD2

...which, I guess, makes more sense of it all. But still, my feeling is that there must be a better way to do this for cases where both modes are available.

Final conclusion, I retract my recommendation to not touch 8.8, but as with any new release, I would say be careful, and unlike me, don’t install it when you’ve got some urgent work to do. That much is just common sense, which this time around eluded me.

I still think Silverfast is a great tool, and one that the film scanning community in particular is fortunate to still have around. The last point is on pricing. I mentioned that it costs €999. Well, yes, but that is for the fully loaded Archive Suite I use, and for the Plustek OpticFilm 120. And actually OpticFilm 120 comes with Ai Studio, priced at €658, and includes an IT8 6x7 calibration target. Depending on your scanner, the basic Archive solution comes in at €299 at the moment (hint, if you’ve got HDR Studio, you only need SE Plus as the scanning part, not Ai Studio). So yes, €999 is a bit of an extreme worst case scenario!

And then there are the entry and mid-level options, at much lower cost (under €50), and still featuring the same core scanning engine. Clearly it isn’t cheap. The competition is - at first glance - cheaper. But taking into account the output quality and the time it will save you, it’s hardly an unreasonable investment in an overall film photography budget.

So to start the New Year on a positive note, I apologise to the Silverfast team for my bad tempered rant, and I wish everybody reading this a happy and successful year of scanning. Even Vuescan users :-)

Posted in category "Silverfast" on Monday, January 04, 2016 at 07:52 PM

Silverfast v8.8. Don’t touch it.

in Silverfast , Tuesday, December 29, 2015

Please take the following article with a LARGE grain of salt. It’s mostly rubbish. Having calmed down, and after receiving a very hurt email from Lasersoft, I wrote this.


Around 3 or 4 years ago, when Silverfast was at version 6.5, it was urgently in need of a complete rebuild, not only to replace the antiquated and extremely convoluted user interface, but also to make it compatible with modern operating systems.  Rather than grasp this nettle, Lasersoft, to all external appearances, fiddled about, adding fake Mac Intel “support”, and various new features which added precisely zero value, but which they charged for.

Finally, they released Siverfast v8, the rebuild we needed, and while it had a bit of a rocky start, there were swift iterations, and quite soon it became reliable and dependable. Various incremental features were added, culminating in v8.5, which I wrote about when it was released. Not only was v8.5 actually useful, but it was free.  Several incremental improvements to the Job Manager followed.

Now, we have v8.8. Also free. And a total, unmitigated disaster. We seem to have gone back to the bad old days (no, I have no idea where 8.6 & 8.7 went, but 8.8 should have gone there too). We have a new iteration of the algorithmic dust & scratch removal, for scanners not equipped with an infrared channel, called “SRDx”. It has an “x” in it so it must be fabulous. Well, maybe it is. It looks less unusable than the old SRD, but that wasn’t hard to achieve. As for the results, I don’t know, but since I only have IR-channel scanners, I don’t need it.  However, what it also brings to the table is that it has apparently totally screwed up the IR version, iSRD, in Silverfast HDR.  Clicking on “SRDi” (the only was to invoke xSRD) selects xSRD by default.  There is no really obvious was to switch to SRDi, expect by clicking on a tab, but that doesn’t tell me if both are invoked or just one. Also, the 1:1 or HQ preview modes just show a black screen, or sometimes a 1:1 unprocessed view.  All real-time corrections such as curves stop working, and attempting to process an image leads to a “quiet” failure at somewhere near the start of the SRD phase.

In summary, Silverfast Archive Suite, at a stroke, becomes unusable.

They have “fixed” the appealing DNG bug from the last version, in that HDR can now correctly interpret DNGs saved by AI Studio v8.8, but not and previous versions.

This is simply not acceptable for a software suite which costs €999.

So what else does v8.8 bring us ? Apparently the SDRx thing can be used on “Digital Camera Images”.  Sorry, Mr Zahorsky, you tried that with Siverfast DC.  It won’t work. Nobody in their right mind would use Silverfast to spot-correct digital camera images.  We also get a “History” feature.  Whatever. All I see here is a sequence of black frames.  I don’t need “History”.  It doesn’t work, anyway. And as far as I can tell it isn’t saved in the “HDR” format metadata.

And there’s an option for something called ColorServer, which is not aimed at people like me, but at production environments (do analog film production environments still exist ?) and which looks a whole lot like the old hot folder feature which I seem to recall existed in v6.5.  And it costs extra. What planet are they on, really?

I don’t much like criticising Lasersoft, because the contacts I’ve had with their staff make them seem to be decent, courteous people.  On the other hand, the way they’ve implemented user to user forums (every single post moderated, with long delays) and social media communication is incompetent to the level of comedy.  Two days consulting with a social media expert would bring them more benefit than a hundred “upgrades”.

In parting, here’s my current issues list.  I had intended to send it to Lasersoft, but really, they don’t seem to care. At a corporate level, anyway.

  • if the “Tool set” is supposed to remember / recall a window layout, it doesn’t work very well. For a start, the Tool-related panels (USM, iSRD, etc) get VERY confused.  They seem to show an activated state in the icon but dis-activated in the panel.
  • In 24 bit mode.  The Histogram tool is floated, and stretched. When I click on 48 bit, the Histogram tool dis-activates / disappears.
  • When 1:1, or HQ Mode is selected in iSRD tool, USM tool says “exact preview disabled” - it seems that previews are not shared between tools, even after they’re activated
  • I often use 720dpi output.  In the resolution drop down, an entry, “Custom (720dpi)” has appeared. Very nice - except it doesn’t work. Selecting this has no effect on resolution setting.
  • Histograms cannot remember that I prefer “Separate”.  SF 6.5 had no problem with this as far as I remember
  • Frame resize mode cannot preserve aspect ratio - e.g SHIFT+Drag
  • OpticFilm 120:  On first start up, first scan, the scanner makes a “clunk” noise, then a second, and then nothing.  To get it to work, I have to force-quit SF8, turn off the scanner, turn it on again, restart SF 8, and scan again. Usually it works on the second attempt, but not always.  After this, it will work fine for hours or days, until it is shut down.  I have tried different USB cables, different ports, USB2, USB3, same thing.  This only started with a quire recent update of SF 8 Ai.
  • v8.8.  The new “SRDx”, which apparently is non-IR based, is selected automatically.  Selecting SDRi makes the preview screen go black. It is totally unclear if SDRx or SDRi are coupled, i.e work together, or separate. From the GUI, it does not seem possible to select which one you want (or both). Both get activated (possibly) by clicking on the iSRD button. Eventually, clicking back and forth makes Silverfast crash. Quelle surprise.

The net result of v8.8 is make me switch to Vuescan.  Well done, Lasersoft.

Posted in category "Silverfast" on Tuesday, December 29, 2015 at 11:32 AM

Page 1 of 6 pages  1 2 3 >  Last ›