80mp digital better than 8x10?

[dsim]

Am I the only one consistently annoyed at Michael Reichman and colleagues inability to set up a proper test.

I don't think it's an inability Tim. Lula constructed a test to show that the IQ80 was better than 8x10 film scanned at 745dpi . To add insult to injury they equated f/32 on 8x10 to f/16 on medium format .

Great going Lula.

[Steve Smith]

I will repeat the answer I gave in another thread.

If you assume 80 square inches of film to be equivalent to 80 million pixels then logically that is one million pixels per square inch. A 35mm frame is 1.33 square inches therefore you must also believe that a 1.33 million pixel digital camera is the equivalent to a 35mm frame of film.


Steve.

[timparkin]

Getting the sharpest image out of a view camera can be tricky. I get images that are sharper than what's been shown, but I'm picky about that. If you want the sharpest images your camera can deliver, then it requires a lot of testing and adjustment.

I would say it's fairly straight forward to get sharper images than are show. Flat focus, use the same depth of field as the phase. Use a very good quality 8x10 camera and dark slide. I would be very surprised if a photographic novice with an hour of training couldn't get sharper images.

Tim

[timparkin]

As for 'Higher scan resolution would not have brought in more details' and 'we can already see the grain', I call bollocks..

Here is a Portra 160 scan at 4000dpi taken with a 150mm Sironar S..



The right hand side was reduced to the 745 dpi they used in the test and then upressed.

Pffft!

[GPS]

Am I the only one consistently annoyed at Michael Reichman and colleagues inability to set up a proper test.
...
Oh, and it was taken with the Nikkor T-ED 600/800/1200 which isn't the sharpest cookie in the 10x8 jar..

Unfortunately, your last sentence is in the same category as their testing methods you criticise. Trying to start yet one more urban legend of amateur photographers, perhaps?

[mcfactor]

I have been waiting for someone else to notice this ridiculous article. The scans looks incredibly soft, like the focus was slightly off and/or the scanner not properly calibrated. Just to make sure I wasnt crazy, I immediately opened an 8x10 scan that I did on an eversmart at 2500 dpi with kodak 160 nc and there was information up to about 2300 dpi and that is with an older ccd scanner.

[rdenney]

Not the only one annoyed by furtive comparisons. These are difficult to make in an apples to apples sense. In fact the technologies are so fundamentally different that apples to apples comparisons can't be made in any kind of a precise way.

Sure. But it is reasonable to have a reasonable comparison, in which Technology A is compared to Technology B in the manner in which both are likely to be used to get the best out of them.

Of course, there are many problems, the main one being that they are comparing direct digital captures with scans of negatives. They should, instead compare high-resolution scans of portions of equal-sized prints. If the objective is web display, nobody is going to use either of these technologies--they are therefore not a reasonable solution to a web-page need. But they are a reasonable solution to the need for a large print. So, compare large prints.

And scanning at <900 spi? They should scan at the maximum capability likely to be used by a real photographer. I'd be satisfied with the best scan from a properly tuned Epson 750, at least for one scenario. But then they should also include a well-made optical enlargement to the same print magnification in their evaluation. That way, they'd be closer to comparing big prints made optically, which should bring out the best of the 8x10 within the reasonable capability of many 8x10 photographers. PMT drum scans ought to be in there, too, but including that doesn't mean they should not also include the Epson scan. Not everyone can afford drum scans, and one of the main advantages of 8x10 is that an Epson is good enough even for large prints. The point of a test is to provide readers an idea of what they can reasonably expect.

Each test scenario should reflect a reasonable application of the test technology. Not doing that is fundamentally flawed and unfair. No amount of subsequent rigor can make up for not setting up the test scenarios reasonably.

I would propose five scenarios, evaluated using a large print: Three using the state-of-the-art processes (which for film would include two scenarios--optical printing and drum scanning), and one each using state-of-the-practice processes, so that people can read the article and know what to expect with processes they can reasonable afford. The state of the art and the state of the practice may be the same with the digital back, so they may only need one test to cover both scenarios.

I've seen tests where film was the favored medium and they were skewed in that direction, too. They would do things with film few could afford to do in practice (which really does include drum scanning for most amateurs and many fine-art photographers). The first half of any such article should present and defend as valid the test scenarios.

I've stopped reading LuLa's tests after they denigrated a 30mm Arsat fisheye based on a comparison with a 30mm Zeiss Distagon fisheye (in Hasselblad mount), when the Arsat was not focused accurately on the same target as the Zeiss lens. When challenged, Reichmann defended the result on the basis that it didn't really matter. Indeed. The Distagon routinely sold at that time for 10-20 times what the Arsat sold for, so a successful outcome for the Arsat might still underperform the Zeiss. But by not focusing properly, he didn't give the reader the tools to make an evaluation of what they might expect.

Rick "validation is as important as verification in testing" Denney

[Tyler Boley]

other than passing on usedfull information to someone experiecing a problem with which we may have some familiarity, or getting the same in return, or even perhaps taking one of the resident shills to task, I find the site is pretty much irrelevant to high end work...
Tyler

[Bill_1856]

http://www.luminous-landscape.com/re..._vs_8x10.shtml

An excellent review. My only quam is that they developed the B&W film in Microdol-X, not Pyro or even D-76.

[John NYC]

At f/32 it is getting diffraction limited. The scans look very poorly done, and I mean in comparison to my own scans coming from a lowly Epson V750.... And done at 745 dpi? Come on.