Elmarit-R 180 f2.8 (non-APO) in-camera lens correction on SL?

[toyfel]

Advertisement (gone after registration)

I am not advocating for or against in-camera correction of R lenses. If anything I am advocating for transparency. If someone could definitely say if there are already in-camera corrections implemented (or not) I wouldn't have to spend time on doing lens tests to find out and could instead enjoy doing real photography. You know, "Das Wesentliche", if you ever heard of that. 

Edited December 26, 2020 by toyfel

[jaapv]

Well, if you need tests to know what your images look like, it cannot be very Wesentlich.

[Bobitybob]

Why don't you email Leica Customer Services and ask them if any corrections for R (or M) lenses are applied in camera and for which lenses?

[toyfel]

6 hours ago, jaapv said:

Well, if you need tests to know what your images look like, it cannot be very Wesentlich.

If there ever was a point where you understood what this thread is about, I apologize, because I must have missed it.

[toyfel]

5 hours ago, Bobitybob said:

Why don't you email Leica Customer Services and ask them if any corrections for R (or M) lenses are applied in camera and for which lenses?

yes, this option is and always has been there and I did think about it before posting here. But it seemed easier to ask here first because I assumed that some die-hard R user would have figured this out long time ago. And I am actually glad I started this thread. There were many comments here that I considered helpful in the sense that I learned something new. Hopefully they were also helpful for other people.   

[tritentrue]

Although this website is vastly a treasure trove of helpful information for Leica shooters, it is littered throughout with posts by those who aggressively peddle their speculation, extrapolations, musings, and "beliefs" as absolute fact; all while disparaging others' experience-based posts that they either half-read or didn't understand.  This is why most of my posts are in the picture-sharing threads, and why I usually end up regretting it whenever I attempt to share my experience in a gear-oriented discussion thread.

I'm sorry you were unable to get a direct, definitive answer for your question about your 180mm Elmarit-R.  My guess remains scenario a).

[adan]

Advertisement (gone after registration)

Without supporting evidence, experienced-based posts aren't really any more reliable than any other extrapolation, speculation, musings and beliefs. Which may also be experience-based.

It is all just "someone saying something on the internet" - until it can be demonstrated, and not just claimed.

[jaapv]

30 minutes ago, tritentrue said:

 My guess remains scenario a)

And mine... And I am still trying to figure out what a lens profile would be supposed to correct for a lens of this focal length with virtually no distortion. Maybe a bit of CA?

[ramarren]

My take on this, after using a largish array of Leica R lenses on the SL, M typ 240/262, and CL, is that Leica's intent in providing lens profiles for both M and R lenses on the digital bodies is NOT correction or improvement of the lenses' rendering qualities but preservation of the lenses' original rendering qualities and intent. Particularly with the R lens system, I've compared the rendering of my range of lenses (15, 19, 28, 35, 50, 60, 90, 100, 135, 180, & 250 mm) between their results on color film in a Leicaflex SL or Leica R6.1 with the results out of the Leica SL: found that the lens profiles improve fidelity of the digital capture compared to captures made with the profiles turned off, they looked more alike. The amount of change from "profiles on to profiles off" decreases quite a lot with longer focal length lenses (over 60mm); the more pronounced differences tend to be seen with the shorter and faster lenses. 

Corrections (removal of edge darkening, color shifting, etc) are generally limited to the effects of using a lens that was designed for film on a digital sensor and are the most pronounced on short focal length lenses, particularly the M-mount lenses (due to the short mount register). I haven't seen much in way of rectilinear correction with any of the lenses; I have seen a bit of reduction of chromatic aberration and fringing with a couple. 

Differences between a 180mm v1 vs v2 are likely almost undetectable in any event, both comparing the "profiles on to profiles off" case and comparing differences between the two lenses. I wouldn't worry about it. 

The result of all my testing is that I turn on the lens profiles for use with all my Leica R lenses and feel confident I'm getting what Leica intended as the lenses' original rendering behavior to the highest fidelty possible given the radical difference between what a piece of film records vs what a digital capture sensor records. 

G

[ph.]

Looking as far into the future as the time passed since the best R lenses were made, the current chips and motors of the present L lenses may no longer be available (example some earlier R chips & the contax+ conurus experience)

.No relevance to this discussion of corrections of lens faults, but our heirs might wish to have easy access to good optics, so even if new future varieties will be designed, taking good care of  the Rlenses , especially since they are orphaned by Leitz,, is important..

p.

[masjah]

OK, this may shed some light on things. On page 20 of the Leica R-ADAPTER L instruction booklet we read:

When using the Leica R lens with a contact strip, the relevant lens profile that has already been saved is automatically activated. This ensures the optimum image quality, and ensures tha the lens data are written in the appropriate EXIF field for the photo.

 

[frame-it]

via apotelyt.com >>

ROM-lenses are lenses for Leica's R-system of SLR that are fitted with Read Only Memory chips. The latter store the characteristics of the lens for transmission to and use by the R8 or R9 cameras. In particular, these cameras can use the lens-specific information to correct for lens vignetting (in connection with the Digital Modul-R), to adjust the zoom reflector on flash guns according to the focal length, or to correctly display aperture information if accessories, like tele-extenders, are attached to the lens.

[scott kirkpatrick]

I bought some R lenses in anticipation of the L-Mount cameras and tested every R lens profile on my (early) SL[601] by assigning each of them to one of my lenses.  There definitely are vignetting corrections (color drift is not a problem for SLR lenses), applied in-camera.  Some of them were unusably buggy.  Leica fixed that in due time, in release 2.0 of the SL firmware, and those errors have not reappeared in subsequent models of SLs.

Distortion and LCA corrections are applied only when RGB data is available at every pixel position, when outputting a JPEG or by the post-processing software when the camera saves only a DNG file.  You can read those corrections out from the DNG and inspect them with a little effort.  They are zero for the longer telephotos but do have an effect on some zooms and shorter focal lengths.  In cases where Leica was not the lens designer (the 15mm's, for example), there are no distortion corrections.

[jpattison]

On 12/22/2020 at 5:10 PM, toyfel said:

I assume I could make a series of tests to figure this out. But I'd be happy about any advice. 

Looking forward to the results of your tests.

Please post the results here - with no lens selected, type 1 selected and then type 2 selected.

Simples.

John

[frame-it]

there is probably a way to disable these in the DNG

 

0xc740	OpcodeList1	undef!	SubIFD	1 = WarpRectilinear 2 = WarpFisheye 3 = FixVignetteRadial 4 = FixBadPixelsConstant 5 = FixBadPixelsList 6 = TrimBounds 7 = MapTable    8 = MapPolynomial 9 = GainMap 10 = DeltaPerRow 11 = DeltaPerColumn 12 = ScalePerRow 13 = ScalePerColumn
0xc741	OpcodeList2	undef!	SubIFD	1 = WarpRectilinear 2 = WarpFisheye 3 = FixVignetteRadial 4 = FixBadPixelsConstant 5 = FixBadPixelsList 6 = TrimBounds 7 = MapTable    8 = MapPolynomial 9 = GainMap 10 = DeltaPerRow 11 = DeltaPerColumn 12 = ScalePerRow 13 = ScalePerColumn
0xc74e	OpcodeList3	undef!	SubIFD	1 = WarpRectilinear 2 = WarpFisheye 3 = FixVignetteRadial 4 = FixBadPixelsConstant 5 = FixBadPixelsList 6 = TrimBounds 7 = MapTable    8 = MapPolynomial 9 = GainMap 10 = DeltaPerRow 11 = DeltaPerColumn 12 = ScalePerRow 13 = ScalePerColumn

[frame-it]

there is actually a simple command to disable all the opcodes, its on the exif forum if you search

 

meanwhile, here is the opcode function in the DNG spec that handles vignetting >>

FixVignetteRadial

Opcode ID 3
DNG Version 1.3.0.0

Parameters

k0 – DOUBLE

k1 – DOUBLE

k2 – DOUBLE

k3 – DOUBLE

k4 – DOUBLE ˆ

cx – DOUBLE ˆ

cy – DOUBLE

Description

This opcode applies a gain function to an image and can be used to correct vignetting.

Parameters (k0 k1 k2 k3 k4 ) define a radially-symmetric gain function, explained below. Note that if all ki terms are zero, then the gain function is the identity (i.e., no gain will be applied).

ˆˆ
Parameters (cx cy ) are the normalized x- and y-coordinates of the optical center, relative to

the top-left pixel of the image. Example 1: specifying (0.5; 0.5) means that the optical center lies exactly at the image center. Example 2: specifying (1, 0) means that the optical center lies at the top-right pixel of the image.

Processing of this opcode is performed as follows.

Let I'i(x y) be the uncorrected pixel value of the i th plane of the image at pixel position ( x y ) (i.e., before opcode processing).

Let Ii(x y) be the corrected pixel value of the i th plane of the image at pixel position ( x y ) (i.e., after opcode processing).

For each pixel ( x y ) in the i th plane of the image, compute: Ii(xy) = gI'i(xy)

where

g = 1+k0r2 +k1r4 +k2r6 +k3r8 +k4r10 c = x + cˆ ( x – x )

x0x10

c = y + cˆ ( y – y ) y0y10

mx = max(x0–cx x1–cx) my = max(y0–cy y1–cy)

m= m2+m2 xy

122 r = --- x–cx +y–cy

m
x0 y0 = pixel coordinates of the top-left pixel of the image

x1 y1 = pixel coordinates of the bottom-right pixel of the image

Note that m represents the Euclidean distance (in pixels) from the optical center to the farthest pixel in the image, and r represents the normalized Euclidean distance in0 1 from the optical center to a given pixel.