6x6 vs 6 mp: what's wrong?

Discussion in 'Medium Format' started by sfcole, Mar 16, 2008.

  1. I recently did a comparison between my Pentax k100d, a 6 mp DSLR, and my newly-acquired Mamiya 6. I expected the 6x6 to smoke the digicam, but it didn't and I'm wondering why. The test wasn't entirely scientific, but here are the details: Pentax k100D wi50 1.7 lens, f5.6, tripod, ISO 200 Mamiya 6 with Portra 400, 75 mm @ f8, no tripod but high shutter speed of 500. Scanned on the Nikon 9000 @2000 dpi. The test might not be entirely fair considering the scan size, film, and lack of tripod for the Mamiya, but I didn't expect such a route for detail. (Don't look at shadow detail--it was a different time of day). My local photo shop thought it was the scan technique. Any ideas?
    00OoG7-42321884.jpg
     
  2. Here's the full scene in 6x6
    00OoGB-42321984.jpg
     
  3. Crop of 6 mp:
    00OoGF-42322084.jpg
     
  4. crop of 6x6 negative:
    00OoGI-42322184.jpg
     
  5. Scott,

    Give us the images.

    TG
     
  6. Scott,

    Out of focus & camera shake?
    Who devlopped the film?

    TG
     
  7. you are seeign the difference between a scan and an original. You also don't way what color spaces you are working in, bit depth, or how the scan and the digital original are being processed.

    it is also pretty clear that the light changed between the two shots but that is not germane to the question of why the DSLR photo looks better than the scanned 6x6 photo beyond noticing the color differences.
     
  8. Well, just due to the atmosphere and light, I like the first photo a lot more than the second. I can't see anything in these two small samples to relate to the native resolution, sharpness, or rendition of the cameras used.
     
  9. Ah, now the crops appear, and I think Ellis is right.
     
  10. If you don't use a tripod, you might as well be using a box camera. You certainly won't see any difference between a 6Mp camera, 35mm film or medium format film - the resolution is limited by the process not the medium.

    Secondly, using an LS-9000 at 2000 ppi is equivalent to throwing away 75% of the information otherwise available from the scanning process. Actually the Nikon scanner ALWAYS works at 4000 ppi and downsamples the results in order to output 2000 ppi. The only thing you save is storage space, ICE processing time and a little transfer time. Scanning is so time-consuming anyway, you should scan at full resolution to make the most of your efforts (Adobe RGB and full bit-depth too).
     
  11. 1. You are using 400 speed film? Try Velvia 50, Velvia 100 or Provia 100F.

    2. You are scanning at 2000PPI? Try scanning at 4000PPI, there is 4000PPI worth of information on Velvia 50, Velvia 100 or Provia 100F.

    3. Is the film in the scanner not flat or at an angle? Make sure that your film is flat in the scanner.

    4. Your camera is using a sharpening algorithm your scanner is not? Try after processing sharpening with your viewing software.

    5. You are looking at a raw product from your camera but a highly compressed .jpg file from your scanner? Try saving you 4000PPI file from the 9000 in .tif format.

    Is your camera focused properly and on a tripod? Is the lens good?

    I have never had any digital camera beat my Rolleiflex let alone my old Rolleicord (lesser model) made in 1956 after scanning in my 9000. My folding Kodak Vigilant Six-20 with an uncoated three-element lens made in the late 40s blows away any digital camera I've ever compared it to after a run through my 9000. Plus how do you get 8000+ pixels on a side from any digital camera?
     
  12. Most has been said, but best to redo your test with 50 or 100 ISO film and high
    quality scanning. Also, was your lens set at infinity? The autofocus digicam no doubt
    focussed at infinity.
     
  13. Try a transparency film and look at it through a loupe. Scanning adds a variable to the process that you need to eliminate for an accurate comparison. Once you have an image that can be examined with a loupe, you can scan to match what you see.

    The Mamiya image above looks like it was out of focus - camera calibration, scanner calibration or both could have a problem. You say newly acquired Mamiya - if you bought used, could be someone serviced the lens and did not properly collimate it. I'm just throwing out possibilities here, but the Mamiya image looks like it has too little information content in it, and that is uncharacteristic of medium format. There is significant likelihood of something in the imaging chain being miscalibrated or otherwise messed up. This is because I don't see the usual suspects that kill information content - camera shake and diffraction. The only things left are the focus calibration, the lens collimation and the scanner calibration.

    You have to eliminate each variable in turn to find out which is the culprit. Focusing the lens at infinity, using a tripod, using slide film and examining with a loupe makes sure that you eliminate everything other than the lens itself. If this is good, then proceed to scan the image and compare.
     
  14. Why scan, make a print from the film, make another print from the digital file then compare the prints.
    You are using two generations on the film camera and one generation from digital, which is not fair.
     
  15. Thanks for all the replies. I think the first thing I'm going to do is go out and do a test roll for
    focus accuracy. If the focus is ok, I'll try a more scientific test.
     
  16. Here's an interesting read, Luminuous Landscape's comparison of image taken by the original 1Ds and a Pentax 645:

    http://www.luminous-landscape.com/reviews/cameras/1ds/1ds-field.shtml
     
  17. There are too many variables which can contribute to this problem. As people have already suggested, it could be due to poor choice of film. A good Velvia scan should definitely be a step up from the Pentax.
     
  18. Start at the beginning and work towards the end of the chain. Step 1: put the 6x6 on a tripod and do that focus test.
     
  19. Something is wrong with your shooting, there shouldn't be this difference between digital and film, no matter of megapixels and film speed. Have you used polarizer for digital shot only? Scanning process, no matter how bad you are doing it cannot change you photo so dramatically.
    I would recommend comparizon of images made of more controlled object before we start another fight film-agaist-digital, kind of like Michael Reichmann did but 5-7 years ago, when 6MP was his top of the line toy.
     
  20. Scott, I had the same problem with my Mamiya 7. I just could not get the detail which I knew was possible from seeing ther people's Mamiya 7 shots. My LF view camera with a 120 roll film holder beat the Mamiya too. I never found ou what was wrong and gave up on the Mamiya with regret as I liked the handling. Ebay'd the lot and went to digital.

    All the best.
     
  21. Perhaps I am in the minority but I prefer the film picture. It's difficult to compare because of the difference in lighting and exposure but to my eyes the MF picture has more and richer detail. Sure the digital picture is cleaner but a cleaner pic isn't necessarily better.
     
  22. "Any ideas?"

    Lots of them, but I've learned to keep my blasphemous mouth shut here. ;) Use both cameras in the manner you are going to typically use them. Process the files/negs as you typically would. Have prints made like you typically would. Hang the prints on the wall, compare with your own eyes, and make your own decisions. It doesn't have to be complicated; just view the photos as they are going to be viewed. The results other photographers get don't matter, unless you are lending the cameras to them.
     
  23. Lots of good advice above, but I'll add just one thing. Portra 400 is not the problem - it's a
    very sharp film and holds loads of detail. The key is to extract what's on the film and not
    exacerbate the grain. The biggest difference in scanning in my experience is to fluid mount
    (the Scanscience system works extremely well). With fluid mounting and 4000 dpi scans, your
    images should be grain sharp, which is the ideal. If you're blowing up to beyond 20x24 so
    that you start to see grain in the final print, then use of smart noise reduction software like
    Nik Dfine will let you manage it.
     
  24. Scott,

    You have met the great equalizer. The scanner. We've all seen the arguments posted before. Here's my digi pic, here's my scan, I can't tell the difference, therefore digi is equal or better than film. I know this is not your situation, but it illustrates the futility of those arguments. Although part of your problem here is your loose technique (tripod vs hand held, etc), the real problem is the scanner. You are comparing a 1st generation image (the digicam) with a second generation image (the scanner). There is always some image loss. It isn't always discernable, but its there, waiting to be revealed with your larger prints. More money translates into better scans which results in better and larger prints with more detail and clarity. To put this in the digital lingo (and I know this is not 100% accurate, but it illustrates the situation), your 6X6 camera would probably equate to 40-60 megapixels (depending on film, technique, etc, etc). By scanning at 2000 dpi, you have dumbed that down to about 4mp (2000X2000). In other words, about 90% of its capability got left in the scanner. If you want to really see what the mamiya can do, go get some true BW film like Ilford Delta 100 and have an image printed optically. No comparison.

    Mel
     
  25. 56mm*56mm at 2000dpi equates to 19MP, @4000dpi it's 78MP
     
  26. My scans of 6x6 film on a Nikon LS-8000 are 8500x8500 pixels aat 4000 ppi (slightly cropped to remove the border). That is slightly over 72MP, or a 28x28 inch print at 300 ppi without resampling. Tell me again about optical printing - no enlarger weighing under a ton is going to make a sharp print that size :) Have you ever tried to process a print that size. It's not a rhetorical question - I've been there and done that, and scanning is sharper from corner to corner. The cropped view is not grain-sharp as one would expect at 4000 ppi. This could be due to the reduced resolution or lack of flatness in the negative near one edge. My experience is that roll film is too floppy to stay flat in the standard holder. You need the glass holder to keep things sharp. Some traces of grain are visible, yet the detail in the rock face is unsharp by another order of magnitude. It is what I would expect if a tripod were not used - and probably the best of the lot at that. I have attached a scan from Reala that is sharp in all regards (Hasselblad, 80mm lens). Some of you may have seen this before, but this question seems to never die either.
    00OptB-42368084.jpg
     
  27. Everyone else pretty much said it all. I would add the following observations comparing the 2 submitted images: the light was different; the composition slightly different; the metering etc different; seasonal differences (1 snow, 1 NOT). You make a good case for both images to use a tripod next time. I have been guilty of taking some really gorgeous shots and discovering that 4x6 is as big as I could print due to blur and hurried technique. How disappointed . . . and I still get caught occasionally. For me, the solid tripod is a terrific image enhancement! :)
     

Share This Page

1111