Jump to content

What's the point of 21.1 pixel camera?


grace_woo

Recommended Posts

"Assuming the scanner people are not just pulling my chain, is there information to that degree in these formats? I am hoping it is not all interpolation. ;-) "

 

John, your question is a good one. I'd love to know the answer. Scanning at 4,000dpi from Velvia and K25, I'd STILL prefer what I'm getting with my 10.2mp D200, enlarged to 20x30...

 

I've had expensive 70mb TIFF scans done of a few Velvias, and again, I wish I'd shot the frames with what I have now.

Link to comment
Share on other sites

I work in all three formats, film and digital. Each one has it's special advantages.

 

I have seen the argument that the eye cannot resolve more than 300 and therefore a print with more than 300 dpi is wasted. Nonsense. To me, higher resolution in a print adds a quality of aliveness.

 

In 2006, my wife and I photographed Bridge Day in Fayette WV. She used a 12.7 megapixel 5D, I used an EOS 3 with Fuji Velvia. Same lenses (300mm F4). Comparing the results, the film seemed to have a better presence. Comparing the digital images to the slides on a light box under magnification, there was ever so slightly higher resolution on the slides. (Don't compare scanned slides to digital images because consumer scanners have abominal resolution and do horrible things to slides). I'm really looking forward to comparing the 5DII to film. It seemed the film was about 16mp and the 5DII should do better.

 

Medium format is more image to spread the line pairs across. A really good lens will do about 70 lpmm no matter which format. 70 lpmm gives 2520 line pairs on 35mm film or a 24x36 sensor. 70 lpmm gives 3780 line pairs on 120 film (54mm) or 3360 on a 36x48 digital sensor. But medium format is expensive ($45k for a digital back)

 

To me, large format is the way to go if you want high quality images and don't mind the time it takes to set up. Large format gives 8400 line pairs on 4x5 and 16800 line pairs on 8x10. Scanning (digital) backs are near useless in my opinion because something always moves in the image during the scan. Anything that moves is fuzzy. You can see object details on a large format transparency where you barely make out object in a smaller format. Extremely sharp wall sized posters are possible. Another advantage of large format is the tilts, shifts, swings etc. With a little patience, it is possible to get both foreground and background as well as the subject in sharp focus.

Link to comment
Share on other sites

I have an Epson Stylus Pro 3800 A2 printer, which cost me less than a couple of my L lenses. A 5D MkI or D700 can only produce prints at a

decent resolution (240dpi) up to a size of A3 (11.7×16.5 in). So my question is: why shouldn't I want a 21.1 megapixel camera? (...assuming I could

afford one!) A2 prints aren't that huge, after all.

Link to comment
Share on other sites

I have a humble 10mp DX camera. I can easily see a quality difference between an 8x10 and an 11x14 print of the same shot. How is it possible to conclude that larger sensors and more pixels are not needed?

 

As for the lens thing, I use lots of full frame lenses without any problems at good apertures. The same sensor density translates to 25mp at full frame. So what is the problem? The high quality lenses that get a lot of praise on the forums should be adequate well beyond this point.

Link to comment
Share on other sites

RL, the problem is that the image quality at the center of the image will be higher than towards the edges. Canon's landscape sample image from the 5D Mk II illustrates this nicely (see the lower left corner). With this approach and lenses available, variation of sharpness across the frame is obvious. Other examples of this are the dpreview.com skyscraper (mirror facade) example. Same thing here.

 

To get even sharpness across the frame at this resolution level you need some special glass (ie. tilt/shift lenses would be a good candidate since they have a larger image circle) or just shoot a larger format in the first place, so lenses don't need to resolve so much in the corners. Of course, you might not care about this, depending on standards and the subject matter. But certainly if you make large landscape prints (measured in feet or meters) this is a consideration.

 

A second consideration is noise and dynamic range. Canon appears to have done well at least in the high ISO noise department. But all other things equal, more densely packed photosites translate to less dynamic range. If you print small enough, the paper resolution and the eye will limit the detail you can see. In this case, unless there was too much sharpening used to make the print, you can gain some of that dynamic range lost in making the photosite density so high. So not much of a problem here. But the quality of the electronics on the chip might not be as good if the photosite density is high, which may still allow the smaller pixel count to result in better dynamic range.

 

In any case, let's just make photographs.

Link to comment
Share on other sites

The point is to have more pixels and thereby higher resolution for large / hi res prints. But it's sure not aimed at me, I've gone from 2 10meg cameras (R1 and E410) and have regressed to a 5 meg Oly E1 and 7.5meg Panasonic L1..while my "highest" res camera is a Ricoh GRD with 8megs. I'm most interested in 9X16 or 8X10 prints with the occasional 13X19" than mural size work.
Link to comment
Share on other sites

  • 1 month later...

Lannie,

 

The effects of special relativitiy have of a much more pronounced and visible effect than that of quantum physics

here. Diffraction will long govern the practical limits of how useful a smaller digital photosite will be before the

uncertainty principle comes into play. The only way to overcome this limitation that is to build massively bigger

lenses (so smaller f-stop, less diffraction) like the lens they make for semiconductor manufacture. For the current

almost diffraction-limited photography lenses such as the Rodenstock HR lenses, 6 microns is a useful practical

limit (80 line pairs/mm equivalent). Not coincidentally, that is currently the smallest sized pixel on medium format

backs. You'll note that the new Phase One back with 65 megapixels therefore represents the practical limit of utility

of the classic 6x4.5 film back size.

 

Cheers,

 

Steve

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...