True resolution of dslrs?

[djacobox372]

Okay, so i was doing another film vs digital experiment and noticed something interesting: when i take a medium format film scan and scale it down to 12 megapixels, it retains noticably more detail then the identically framed shot taken with my 12mp d700. As I continued to scale down the film scan parity was achived at around 8mp.

So what is the true resolution of popular dslrs? Obviously you cant go off of mp numbers alone.

I'll try to post some example pics later.

 

[Josh66]

IMO, you're comparing apples to oranges.

Scale a medium format scan down to 35mm sizes? Why would you expect it to NOT blow the digital file away?

 

[Josh66]

A more fair comparison would be if you stitched together a 4 frame digital photo:
[1] [2]
[3] [4]

And compared that to the 120 scan.

I mean, medium format is f'ing huge compared to even a full frame sensor...

35mm = 24x36mm
6x7 = 56x67mm

A 4 frame (full frame) pano (2 wide by 2 high) would be 48x72mm, or less depending on how much overlap there was. Not perfect, but it's a lot closer.

 

[480sparky]

A couple weeks ago, I calculated the mp equivalency of the 10" film used in the old Cirkut cameras. Came out to roughly 7217.

 

[djacobox372]

IMO, you're comparing apples to oranges.

Scale a medium format scan down to 35mm sizes? Why would you expect it to NOT blow the digital file away?

Not an invalid comparison at all... I'm comparing two 12mp images (apples to apples), both images have 2832 vertical lines of pixels. However, the DSLR version does not contain as much detail as the image derived from a medium format scan and scaled to exact same pixel count. The later of which acts sort of like a "control" to tell me how much detail is possible in a 12mp digital image.

It's common knowledge that the tiny sensors on high-res P&S cameras don't produce as much detail as a full-frame DSLR, but some may be surprised to find out that even a full-frame, low-density sensor like the D700 only achieves about 2/3rds of the resolution that the pixel count could conceivably store. I find this interesting, because camera makers still have an opportunity to improve resolution without having to increase file size.

 

[Derrel]

Okay, so i was doing another film vs digital experiment and noticed something interesting: when i take a medium format film scan and scale it down to 12 megapixels, it retains noticably more detail then the identically framed shot taken with my 12mp d700. As I continued to scale down the film scan parity was achived at around 8mp.

So what is the true resolution of popular dslrs? Obviously you cant go off of mp numbers alone.

I'll try to post some example pics later.

Interesting post!!! The Luminous Landscape crew posts an article or two on this topic about once or twice a year. Their most recent one is here: Everything Matters

The author goes on and on and on about how superior medium format images are, and why. But then, he also puts forward and really,really advocates for that old "audio cable" myth...he even goes so far as to say that the audio quality IMPROVES after an audio cable has been used for a week or so....snort!

Medium format film versus 12 MP NIkon d-slr D700...you know what? I have some 120 rollfilm exposed with an Argus Argoflex, with an uncoated 75mm lens, around 1976...the detail resolution of the images made with that then 40 year-old, pre-World War II $5 fleamarket camera, with 1976-era ASA 125 Verichrome Pan B&W film are...almost amazing to me...bigger is better some people say...

A 4x5 Ektachrome exposed with flash and a Funjinon 150mm view lens can resolve the teeny-tiny lettering on the side of a 2-liter soda bottle 18 feet away...on 35mm color film of that same era, shot with a 35mm Nikon, the GRAIN limited the acuity, and the small size of the image on the film made the same lettering impossible to read. Bigger pieces of film contain bigger images than smaller pieces of film.

What would be great to have is a medium format rollfilm shot, made with say a 645 back, and then the same subject shot with a digital back of roughly the same sensor size, and then a 24x36mm d-slr capture of the same subject matter.

 

[Garbz]

Not an invalid comparison at all... I'm comparing two 12mp images (apples to apples)

You're not.

What you are comparing is 12 mpxl images which were derived in different ways, one of the ways giving one image an unfair advantage over the other by masking the faults in it's recording process. See our digitisation equipment does not detect colour. Both your scanner and your camera record only the lightness of each individual point, but they do so through a red green and blue filter. These filters then are used to derive a colour picture where each pixel is made up of a red green and blue value.

Film does not work like this. Any given point on the medium has a full colour range as that point can be dyed with any combination of dyes. Your digital camera on the other hand has 3 million blue pixels, 3 million red pixels, and 6 million green pixels. Interpolation is then used to derive 12 million full colour pixels, it's not perfect but it actually works quite well. This however relies on any given point source of light falling across a group of 4 pixels in order to derive colour, which is done by a layer of bifringent glass on the sensor called the Anti-Aliasing filter (or low pass filter since it has a coating that also blocks IR). Effectively for a sensor to work you need to blur the image then sharpen it.

This is true of scanners too yet their implementation differs. Where you're falling short in your fruity comparison is that by resizing the picture that was recorded you effectively sharpen these slight nuances, much the same as reducing a 12 mpxl DSLR to a 4 mpxl DSLR effectively doubles its sharpness.

But it's not only the film itself that you're now putting at an advantage, but also the recording system. Medium format has larger glass physically and thus very different lens characteristics such as less diffraction. By downsampling the image or scanning it only at 12 mpxl you're also giving the lens on your medium format camera an unfair advantage over the film lens.

Either way you're never going to have an apples to apples comparison of film vs digital. Although you could get close by comparing a film shot by cropping the appropriate area, scanning at high resolution, and rescaling it to be the same sized as an image from a camera equipped with a foveon sensor, using the same lens.