Hi,

I have started a new comparison between my Pentax 67 with Velvia 100 and my Sony Alpha 900. It's on my webpage:

http://83.177.178.241/ekr/index.php/photoa...-sony-alpha-900

The article is very short this far. I plan to extend it with some more observations the next few days.

One of the comparisons is below (Pentax to the left and Sony to the right)

Click to view attachment

Added:

Calculated LW/PH with Imatest

2271 for Pentax (kind of very low)
2890 for Sony Alpha (kind of low)

The article has a link to the original images.

Best regards
Erik

Interesting stuff. But you might wish to view your page in FireFox. The paragraph formatting is stuffed.

Hi,

Yes I'm aware of that, but don't know what to do about it. The images don't fit on the page and that also affect the text. I could make the images smaller, or try something else.

Thanks for input, much appreciated!

Erik

Erik,

Thanks for the comparison. A first comment coming to my mind is that the Minolta scanner is probably far from doing justice to the slide which might explain the gap between these samples and the good enlargements you have done from these slides.

Besides, as long as we are pixel peeping, is the A900 image really properlly focused/shake free? For a sharpened image it doesn't quite exhibit the sharpness feel I would expect from the A900 (judging from I see with my D3x when I didn't mess up with something). Was there a bit of wind when you shot?

Cheers,
Bernard

Hi Bernard,

You may check this page on DPReview: http://www.dpreview.com/reviews/NikonD3X/page29.asp
I may be somewhat in doubt about their observations, but here they observe some unsharpness which they attribute to stronger AA-filter on the Alpha 900. Could be right or wrong.

Regarding the enlargements they were done using the same scanner, but I have spent several hours on sharpening and noise reducing the pictures. An Imacon scanner would probably pull more detail, but the Minolta clearly shows grain and on most of my pictures it also shows chromatic aberration. As a side note, I paid about the same for the scanner as for the camera, about 3000 USD (in 2001? and and 2008).

It's a pity we cannot share prints on the web. Let's put it this way, the image I'm talking about is a picture of my old friend Pierre who has been working like 10 years at one of the leading color labs in Sweden working with repro, 8x10", APO lenses and he is impressed. I know another guy, one of my colleagues who used to work at another lab and he is impressed, too. So my Pentax 67, Velvia, my scanner are obviously capable of producing professional quality 70x100 cm prints.

Regarding the sharpness from the Alpha 900 I was also expecting some more, but I don't feel I have messed up the image. I used:

- Prefocus on the central sensor
- Aperture f/8 on 100 ISO
- Mirror lockup
- Self timer
- Antishake off
- Gitzo GT3541LS tripod

I also compared it with a twin image but could not see any difference. Also checked for focusing errors but I cannot see any evidence of that, depth of field is plenty and neither foreground or background are sharper than the detail shown.

Sharpening is landscape sharpening in Lightroom, that is not excessive. The idea was to apply pretty much standard processing to both. Imatest did indicate undersharpening by about 20% and that is fairly typical of "landscape sharpening" in Lightroom, but that was another image.

Best regards
Erik

Erik

I can't compare/comment on your digital work but I did have a brand new Pentax 67 ll for some months. I got rid of it because, although it represented the best thing to match 35mm convenience, it would never do sharp! I only used it on a Gitzo G410 - never hand-held - and always mirror up. With the 200mm I couldn´t even guarantee focus! (I ordered a split-image screen but sold before it arrived.) Generally though, the problems were induced by the shutter which had so much mass and kinetic energy going for it that nothing at the sides during slow shutter speeds remained sharp. An expensive disaster which drove me straight back into the arms of Nikon which I had abandoned for the sad Pentax trip.

Rob C

Hi,

I have perfectly sharp images from my Pentax 67. On the other hand, I'm very much aware of the issue with shutter induced vibration. It was a very major issue on my Manfrotto 55 tripod but much less on my present CF tripods.

I honestly don't think it's a problem here. There is some very fine detail which would not be distinguishable would vibration be a problem. I'm gong to explain later.

Erik

I used to get quite wrapped up in these questions of film v digital but I now feel it's slightly passed its sell by date, though I am certainly not criticising the purpose of this comparison. It's simply that we can achieve perfectly adequate quality now with either medium. Personally I haven't shot any film now for a few years but that is mainly down to a question of costs and convenience more than anything else. I still look at some of my old drum scans from large and medium format images with a certain admiration when they were done well. The problem is that very often the weakest link in the chain was either the scanner or the operator and with digital I now feel I have greater control over the end result. Printing black & white in the darkroom can be a pleasure but it sure is less messy in Photoshop.

After scrutinizing both original TIFFs on my 29inch Eizo, I find the overall look of the film scan (even despite the moderate quality of the scanner used) to be much more pleasing. The Sony file looks flat and plasticized. Just one opinion of course:-)

Thanks Erik for taking the time to do this. It would be interesting to see a non processed original crop from the film scan against an uprezzed and processed crop (to match size) from the Sony... or maybe you already did this?

Hi!

Sorry for not responding...

I'm planning to publish some more comparisons, and I will publish the images for download. Some of the original images are here:

http://83.177.178.241/ekr/test3/index.html

I cannot tell how much processed they are, unfortunately.

Some comments:
1) In my view a scanned image always needs sharpening, so I don't feel that comparing unsharpened images is not fair.
2) The "plasticized" look is simple because the digital image has very little noise and no grain structure.
3) Next time I'll also publish original RAW and TIFFs from scanner

Best regards
Erik

You can add some noise to the digital image to make it look remarkably like the film scan, and fool some people into thinking there is more detail than exists.

Hi!

Some more test images added: http://83.177.178.241/ekr/index.php/photoa...date_2009_10_30

http://83.177.178.241/ekr/images/VelviaTest/Scaled800_BW.jpg

All images scaled to 8000 pixels width
Left column Velvia scanned at 4800 PPI right column Alpha 900
Top row not sharpened
Bottom row sharpened

Best regards
Erik

Unfortunately your comparison is as much about the scanner and your scanning skills as the Pentax camera. The scanner is such a huge part of this that it becomes the deciding factor. Look at this comparison of scanners. It is not perfect but it is the only scanner comparison out there. Scanner Comparison. If I remember correctly the Minolta and the Polaroid are the same scanner? The best scans can't pull everything out of a transparency but to truly do justice to the Pentax you would have to get a first class drum scan (not an Imacon scan). For example I could do a similar test with my Hassleblad and my Canon 5D2 scanning the film on an Epson 750 and an Imacon, but I know that neither of these scans are the best so the scan would make the Pentax look worse than it could. So my comparison would be limited by my available scanner and not a valid comparison to anyone else unless they scanned the film on the same scanner with similar scanning expertise.

Hi,

The article discusses the limitation of this testing. See it as something intended for folks using this class of equipment.

Regarding the scanner the Minolta Dimage Scan Multi Pro is not identical or similar to the Polaroid scanner you mention. To begin with it has almost twice the resolution 4800 PPI vs. 2500 PPI. A lot of optical deficiencies of the Pentax lenses show up in my scans, even at 3200 PPI, lateral chromatic aberration is most prominent. So I cannot agree on the scanner being an obvious limitation. I also made pretty much perfect 70x100 cm prints on Durst Lambda from my scans. The article you refer to uses 2400 PPI resolution, so that's about half of what I have. There has been three scanners in the same group as the Minolta, the Polaroid Sprint Scan 120, the Nikon Coolscan 9000 and the Minolta. All of these wre CCD based MF scanners. The article you refer to does not list any of them, probably because neither of these scanners can handle largeer than 120 format film.

The major issue with the Minolta scanner, in my humble view, is that it is nowhere near it's theoretical DMAX of 4.8, it's probably closer to 3.5. The images I show here are pretty much enlarged, if we assume that your monitor resolution is about 100 PPI, the image you see would be something like 120" that is about three meter wide, did you ever enlarge to that size?

Yes, drum scanners using photomultiplier tubes would probably have better D-MAX and possibly better resolution. Finding a capable lab may however not beeing easy. Also, I made about 20 scans for the article. Would I find a good lab doing drum-scans at 100 USD each just this test would put me back something like 2000 USD, about 60% of the cost of my Alpha 900 DSLR.

This article has some very good comparisons: http://www.luminous-landscape.com/essays/back-testing.shtml , it is from 2006, so it does not cover the latest generation of DSLRs or MFDBs, and it includes 6x4.5cm and 4x5" Velvia but not 6x7 cm. This is part of the reason I made this test (the dollar note is about the same scale as in that test). The reason I made this test is mostly that I have both equipments.

I have no argument pro or contra film vs. digital just presenting my experience. For that reason I'm not referring to other articles.

Best regards
Erik

Somehow I don't think this will be the last comparison.

Here's the rub - while a drum scan is nice and all, unless we're talking about either Aztek or ICG, none of the other players are still around. Add to that there are only 5 companies still actively producing scanners; Canon, Epson, Hasselblad, Kodak, and Nikon, and you have yourself relatively few options grouped into two categories: affordable and insanely expensive.

Now, here's an idea - instead of comparing scanned film to a digital capture, why not compare the same piece of film, scanned across the last scanners standing? THAT would be something worth seeing, provided, as Kurt so rightly pointed out, the operator was skilled in the use of the particular scanner.

Hi,

Yes that may be interesting. My primary endeavor was to find out the capabilities of the stuff I have.

Experience is an interesting issue, I could probably pull out more shadow detail a couple of years ago when I was scanning frequently, resolution is probably less affected by operator experience.

Best regards
Erik

the only folks (primarily artists) i know still using mf or large format film are drum scanning.

I've asked some pro's I've either worked for or came across their work by other means, those who still use film for some or most of their work are using either Kodak iqsmart or eversmart scanners, or Nikon 8000/9000 series scanners with a regluar dry mount, sometimes wet mount. But many are turning to getting lab roll scans done (properly done, these scanners, primarily Fuji Frontier's are capable of tremendous quality and DR). This use of lab-scans is generally for a quick turnaround, but the photog can't bill the client more for post time.

I would imagine that some(1 that I know of definitely), use the Imacon/Hasselblad scanners. They are great, but crazy expensive. But cheap compared to a drum scanner and all the software to get it running.

-Dan

Hi,

Here is the best comparison of the Minolta vs. the Imacon I have found:

http://www.allari-photo.com/scanimages.html
and
http://www.allari-photo.com/body_scancomp.html

Here is a collection of links on the MF film vs. Digital issue:

http://83.177.178.241/ekr/index.php/photoa...vs-mfdb-vs-film

Erik

An established artist friend of mine who did a 70+ piece exhibit at Ace Gallery in Los Angeles, with most prints being 7' diagonal or larger, shot MF film with a Coolscan 9000.

Hi,

In my view CCD scanners are OK for large print, I have made very good prints from scanned Velvia and Provia. The largest prints I made were 70x100 cm. The major disadvantage I have seen with CCD scanners is that they cannot handle the DMAX of Velvia (it's about 4.0). The Nikon may be better than the Minolta scanner I had. Whatever DMAX the CCD scanners handle, slide film has a low dynamic range.

What this boils down to is simple. What can be done depends on subject and film. In the best case a CCD scanner can deliver, in other cases it may be hard. I have never scanned negative film, to my knowledge it is not easy,

Best regards
Erik

Hi,

When I wrote "last comparison by me" I essentially meant that I publish all my finding in this discussion instead of starting new ones.

Comparing scanners is not easy as there is is no easy way of sharing physical quantities on the net. We cannot share print, silver halide images or Idaho Potatoes over the net. One way is to try to reproduce some standard image, like the one here:

http://www.luminous-landscape.com/essays/back-testing.shtml

Could be done by anyone having a mini color checker and a dollar bill.

Best regards
Erik

yes, in my experience the coolscan 9000 produce mf film scans from my mamiya 7 which far surpass dslr quality in tonality and color in large 30x40 or bigger prints...most fine artist's will agree. i only use my d3x when absolutely necessary.

Harlemshooter, in your opinion:
Under the same circumstances a print of "only" 24X30" would far surpass a D3x? - Or at this size, the difference would be irrelevant?
Do you wetmount?
Back to the 30X40 print: Is print resolution better or just tonality and color depth?
Thanks so much for your answers
Eduardo

Hi,

Are you shooting slide film, negative color or Black and White? If slide film which one? I would expect Velvia having issues with blocking shadows, due to it's high D-MAX.

Best regards
Erik

Hi,

This comparison http://luminous-landscape.com/forum/index....mp;#entry322993
was made by some knowledgeable gentlemen (Bill Atkinson, Charlie Cramer and Michael Reichmann) and regarding film probably still valid.

Regarding the tonality and color depth I'm somewhat confused. With digital almost any color can be achieved.

Best regards
Erik

Not really. The camera gamut is limited by the Bayer colour filters and the demosaicing algorithms, and your ability to process digital images is limited by the gamut of your monitor. I do find it odd that the gamut of digital cameras is almost never discussed.

Hi,

My impression is that camera gamut volumes are huge, but so are scanner RGBs. I have not found any RGB volume for Velvia. Anyway the issue is complex. Cameras and scanners don't really have color space. Regarding film it works like a camera, but using subtractive filtering. Once we project or scan the film it has probably a gamut volume defined by some RGB coordinates corresponding to dyes. Scanning the film adds another level of complexity. We have a gamut mapping from film to scanner. Scanners have huge gamuts, so we can assume that film gamut is preserved. If we discuss scanned images we can definitively adjust them outside the gamut of the monitor, but cannot display on screen. The colors missing on screen may show up in print or even not.

Check this article: http://www.luminous-landscape.com/tutorial...photo-rgb.shtml


Best regards
Erik

Is that like when 4 out of 5 dentists recommend sugarless gum?

Better yet: toothless gum. There lies money!

Rob C