Understanding Parallax error and viewfinders

[msbarnes]

OK so I just bought a 21mm skopar + 21mm CV metal viewfinder from. I plan on fitting this combo onto my M2 or M3.

My question:
Is the M viewfinder designed to mimick the lens's perspective without taking into account any offset? Are external viewfinders designed to take into account this offset?

I didn't measure the distances but I don't feel that the external viewfinders are THAT much further from the lens than the in-camera viewfinders...I could be wrong though, maybe this difference matters more and really makes the difference in terms of RF accuracy.

I know accuracy is SLR territory but truth is RF framing is usually accurate-enough with the way I focus/compose with my normal lenses. I'm just unsure in what to expect so I'm trying to understand the limitations. The Bessa R4M seems to be the most accurate RF solution....

 

[RallyFan]

From CameraQuest:

"The 21 brightline finder was introduced in May 2001 with the 21/4. It has glass optics and a hard to scuff outer plastic casing. The Voigtlander 21 finder is especially interesting as an inexpensive alternative to the expensive Leica 21 brightline, Contax 21 finder, or Nikon 21 finders. Many shooters find they prefer the Voigtlander 21 finder to the Leica. Like most wide angle finders, it has no manual parallax compensation -- but it does have parallax compensation marks in the finder. The lower dotted brightlines indicate the approximate view at about 3 feet. The upper solid brightlines indicate the approximate view at infinity."

I only know this because I've recently researched the same thing myself. I'm still not 100% certain that it will solve parallax completely, as I'd imagine the hot shoe location's height is not the same for all cameras.

 

[Doug]

...My question:
Is the M viewfinder designed to mimick the lens's perspective without taking into account any offset? Are external viewfinders designed to take into account this offset?
...

The Leica M body's viewfinder is offset from the lens centerline both vertically and horizontally, so there are these two components to the parallax error. So as you focus the lens closer, the viewfinder framelines move down and to the right as an automatic approximate compensation.

This automatic compensation is more or less accurate at one focusing distance only. That's because in a 3D scene there's still that vertical+horizontal discrepancy between what your eye is seeing and what the lens is seeing; a slightly different point of view. So if you line up a parking meter with a lightpole and a distant mountain, they'll look lined up in the viewfinder but not in the picture imaged by the lens. The only real correction for this error is to move the camera up and to the left so the lens is right where the viewfinder was when you composed the pic. Pretty tricky to do with any accuracy!

A shoe-mounted external viewfinder is positioned directly above the lens on many cameras, so there's no horizontal parallax error, just vertical. The guideline etched into the vf glass will approximate that edge of the frame at close distances, and you guess for distances between that and infinity.

As you suggest, it's workable, and still it helps to know the details. Useful question!

Edit: Twin lens reflexes also suffer vertical parallax error. Composing on the ground glass gets you a view of the subject from a slightly higher viewpoint than the taking lens. If it's on a tripod, you could raise the camera by the distance between the lenses before tripping the shutter. Then the picture would show what you had seen on the screen. There was a device made to do this conveniently, called a ParaMender; IIRC you just move a lever to raise the camera the correct amount.

 

[David Hughes]

Hi,

I wonder if I dare add a "yes, but... " or two to this thread.

Mostly I use 24mm and so on for landscapes and sweeping views. So it don't worry me if the tree on the horizon or the grass below is a couple of inches out in the frame either vertically or horizontally. If it was important I'd use the SLR but the important differences are at close ranges and 24mm and worst distort so much I don't use them that way.

On CRF's I often use 85 or 90mm lenses for portraits and I use them in portrait mode where I know a bit of the background will be wrong but also OOF, so why worry? (And, although the head is the main subject I seldom frame from one side of the face to another but usually from a little wider than shoulder to shoulder.)

And when it comes to printing we seldom use the entire negative anyway. Our 35mm film gives 5" x 7" for most people and that drops a half inch off the sides, and 8" x 10" is worse and drops two inches off the sides of the picture. Two inches of framing lost when printing is far worse than two inches of OOF background or a 2" displacement of a tree on a mountain in the distance.

Anyway, rant over. I'll go and lie down quietly somewhere and take a tablet.

Regards, David

 

[Doug]

I think you've put it in perspective, David. (Sorry for the pun.)

The greater the subject distance, the less the parallax error.