review by H.Hobo | ||||||||||||||||||||
Nikon D2X versus Mamiya 7 |
||||||||||||||||||||
Example 1 | Conclusion | |||||||||||||||||||
The idea to carry out this test came some time ago when more and more tests became available and 12 mpixel DSLR's where tested better or equal than 6x7 medium format cameras. Many will know the famous tests about this subject on the site of luminous landscape. On the other hand there were tests which said that well scanned 35mm films are almost similar with those same cameras. A rather unexplainable matter. Clearly was, however, that with or without set-up the tests had been carried out frequently badly. Medium format 6x7 cameras were not always equipped with the best lenses and often there were doubtful reasons for downsizing to make the images similar to each other. There is even a test where the photographer photographs through the window, and who knows how clean this window was !? The slides of the Mamiya7 i see are very sharp and detailed. It seemed the right time to do some study myself. It had to become a test as honest as possible, without postulated outcome and commercial interests. The aim of this test is therefore to examine to what extent 6x7 qualitatively can measure itself with a modern 12 Mpxel camera such as the Nikon D2x. I do not have the intention to make any statement whether digitally is better than analog, or the other way around. Both have a another use aim. For both cameras, we tried to make a good test situation, though always keeping in mind the practical approach. I used a firm tripod, good lenses, optimum diaphragm, mirror lock up with the Nikon etc. We have done our best. Concerning scanning, the opinions diverge what is the best method. Scanning is a very important factor for the final result. The price will be an important factor for a lot of people. It is clear that a cheap scan, like the Epson flatbed, does no right to the qualities of 6x7. On the other hand a drumscan is seen as too expensive for daily use. Since this is a qualitytest between analog and digital, there has been chosen for an ICG drumscan and for an Imacon flexscan.The difference can be seen immediately between both. On this page you can see the large difference between an Imacon scan and an ICG drumscan At several places you can hire the Imacon scanner, at a certain price per hour, making the price per slide very acceptable. For a single real large print, a drumscan nowadays does not really cost too much. At professional photolab Capi in Amsterdam a 100mb scan will cost you 45 dollar. For a 31x47 inch (80x120 cm) print, the price of the scan is 25% of your total costs, what seems acceptable to me. The first photograph from the test photographs below, with the boat, has been scanned with ICG drumscan. The photograph with the bottles has been scanned with an Imacon 848. , the first test photograph of the boat has been scanned with both scanners.
Working method.
|
||||||||||||||||||||
Click on the red squares to see details at 100%
|
||||||||||||||||||||
Mamiya 7
65mm F11 Drum scan ICG 2400 dpi 5350x6670 pixels de-noised and sharpened in Neat. Download this picture full size. This picture is made in the Netherlands. Here is the GPS location. |
||||||||||||||||||||
Nikon D2X Raw image converted to 8 bits TIFF with Phase one 3.7.1 4200 pixels resized about 180% to match the Mamiya scan Nikon 18-35 24mm on F11 |
||||||||||||||||||||
The picture below, the bottles, are in fact NOT the same, as you can see. The Mamiya 7 cannot come closer then 1 meter, not sufficient for this photo. Although in fact a failed photo, on second thought this photo is perfect to show the possible image resolution of the Mamiya, of cource it is an estimation but i think we come close to the truth. When you look to the details you can easily see that the Mamiya is considerably sharper than the Nikon D2x. By example the watch or the details in the lens of the Yashica camera.The question is; how much more Megapixel is ' a little bit sharper? When a picture is already sharp with small details, which is normal with a 12mpixel camera, you need a lot of pixels more to see a noticeable difference. On this site a comparison between a Canon 1Ds mark II (16mp) and Betterlight scanning back of 48 Mpixel is made. In this comparison between Canon and Beterlight you can certainly see some differences. But you talk about 3 times more pixels! The Canon is 16 Mpixel compared to Betterlight 48 Mpixel. The differences in detail of the Mamiya/Nikon D2x test photograph below, with the bottles, are at least of the same magnitude. But to be save i assume a difference in pixels of a factor 2. Also important is the fact that the Mamiya photo has been taken smaller. On the Nikon photo the champagne bottle is 15.3% of the total width. On the Mamiya photo the same bottle is 11.5% of the total width. The same bottle (and all objects) on the Mamiya photo can be made 15,3/11,4 = 1.34 times wider. The number of pixels increases then quadratic: 1.34x1.34= 1.8x. The Mamiya therefore is in any case 1.8x (12 Mpixel of the Nikon D2x) = 21.5 megapixel. But because the more detailed image of the Mamiya is sharper and probably 2 times more megapixels (as explained before), the Mamiya has a total number of megapixels of 2x 21.5=43 megapixel. In short the Mamiya has a total amount of something between 21.5 and 40 the megapixel. That corresponds also very well to the value for 35 mm film, between 6 and 10 megapixel. Since 6x7 approximately is four times larger that produces roughly 24-40 megapixels. |
||||||||||||||||||||
Mamiya 7 photograph 65 mm F11.
Imacon scan 8200 pixels (300 dpi) Minimum distance of Mamiya is 1 metre that is why there is so much space around the bottles. Colorrendering is not perfect because the Imacon was not color calibrated. Bottles Are 0.75 times the same bottles in the Nikon picture below. The bottle 'brut' is 11,4% of the total width. |
||||||||||||||||||||
Nikon D2x resized to match resolution of the Mamiya.
18-35mm with 35mm F8 1/6 sec. Very good color rendering. Bottle 'brut' is 15.4% of total width of picture. |
||||||||||||||||||||
Prints.
|
||||||||||||||||||||
Some other tests with Medium format compared to Digital 12 mpixels.Mamiya 7 compared to Canon 1Ds I think it is not a good idea to make a photo through the window. Even a little bit of dirt or scratches on the glass will lower the maximum resolution of the photo. This will have the biggest impact on the camera with the highest resolution, the Mamiya. Luminous landscape 645 compared to 1Ds This is a well known test and a lot of references on the internet. When i read this test i will immidiately buy a Canon 1ds. Strange i get completely different results. Various formats compared. 8x10, 4x5 inch, 6x7, 35mm leica, Canon 6mpixel DSLR. Very interesting but medium format and bigger film is scanned with an Epson scanner. The loss in resolution will be too big to get a good comparison. The lens of a Plaubel 6x7 is not very good according to this test with an overview in lp/mm with a lot of medium format lenses. The results of 6x7 and bigger are therefore not showing the real possibilities of these formats. Pentax 67 versus Canon 1Ds Seems a serious test, 4000ppi nikon scanner. Hasselblad versus Canon 1Ds. Scanned with a Nikon 8000 scanner on 4000 dpi. The Canon is almost as sharp as the Hasselblad. It is a pitty that the rope is photographed from above /side. In this way the sharpness of the photo depends on the depth of field. Not sure if the shown crop is the sharpest point of picture. Downloads.Original photo downloads. The Mamiya file is drumscanned and colorcorrected. Be carefull these are very large files (20mb) and take a long time to download. Please don't download if you don't need them, they cost a lot of bandwith for me. Mamiya 7 without water and sky minimal Jpeg compressed photoshop 12. Mamiya 7 without water and sky Lossless Jpeg 2000 !!VERY large |
||||||||||||||||||||