Managing Viewing the Best from M240

[Lindolfi]

Advertisement (gone after registration)

Fig. F and Fig. G are the same motif taken respectively by M8 and D-Lux 6, shooting conditions are identical except the aperture size wide open.

 

 

In those figures you show images in which you varied

 

1] position of camera

 

2] orientation angles of camera

 

3] sensor size of camera

 

4] focal length of camera

 

5] aperture

 

When doing experimental comparisons, it is wise to only vary 1 variable to get conclusive results.

Sometimes it is useful to vary 2 variables in different pairs, to see interaction. But 5 variables changed at the same time is not good for obtaining insight in the effect of one of the variables (like sensor size).

[Erik Gunst Lund]

Distance from sensor (image) center to left & right edge

Center to upper & lower edge

Center to tip of corner

M8

9 mm

13.5 mm

16 mm

M9/M240

12 mm

18 mm

21 mm

 

 

As most Leica M lenses before 2006 were designed for 35mm format full frame film (24mm x 36mm), when put on M8 body, the rays traverses through lense cover not only M8 sensor (18 mm x 32 mm, with a cropping factor of 1.33)

......

Just for the reference; In your drawing you have by mistake written 13.5 for M8 it should be 16.2mm also in drawing and text for M9: 21 should be 21.6mm, like this:

 

M8

9 mm

13.5 mm

16.2 mm

M9/M240

12 mm

18 mm

21.6 mm

[Thomas Chen]

In those figures you show images in which you varied

 

1] position of camera

 

2] orientation angles of camera

 

3] sensor size of camera

 

4] focal length of camera

 

5] aperture

 

When doing experimental comparisons, it is wise to only vary 1 variable to get conclusive results.

Sometimes it is useful to vary 2 variables in different pairs, to see interaction. But 5 variables changed at the same time is not good for obtaining insight in the effect of one of the variables (like sensor size).

 

Bert,

 

Thank you for the comments.

 

This is not a serious experiement. I just want to show the effect of crop factor to the "far corner".

 

I have a plan to do with 90/2 and 90/2.8 before M240 arrives, then I will do per your advices. Even with tripos and shutter release calbe. Just a little bit worry about the focus shift.......

 

All the best,

 

Thomas Chen

[Thomas Chen]

Just for the reference; In your drawing you have by mistake written 13.5 for M8 it should be 16.2mm also in drawing and text for M9: 21 should be 21.6mm, like this:

 

M8

9 mm

13.5 mm

16.2 mm

M9/M240

12 mm

18 mm

21.6 mm

 

Thank you so much, I will straighten it out.

[Lindolfi]

Thomas, I invite you to react to the experiment I did this morning (click), aimed straight at the heart of a long standing statement you made on sensor size and out of focus rendering in discussion with Erik and myself.

 

Have we resolved the issue now? Thanks.

[jaapv]

Link does not work.

[Lindolfi]

Advertisement (gone after registration)

Link does not work.

 

This is the direct link to the document click

[Thomas Chen]

Thomas, I invite you to react to the experiment I did this morning (click), aimed straight at the heart of a long standing statement you made on sensor size and out of focus rendering in discussion with Erik and myself.

 

Have we resolved the issue now? Thanks.

 

Bert,

 

I saw your presentation before I made the reply today, it's a good presentation.

 

The problem is that I mistook the "far corners (told by Eric)" or "missing bokeh presentation in the field due to cropping by sensor size (in my terms) " for "no bokeh" or "bad bokeh". Also an instance of miscommunication owing to "inappropriate termnology".

 

Now it's clarified. Thanks a lot.

 

Best Regards,

 

Thomas Chen

[Thomas Chen]

Please read up on this software first. Adobe offers excellent tutorials on their site, the books by Scott Kelby (for beginners) and Martin Evening (for advanced users) are mandatory literature.

Mr. Jaap,

 

Scott Kelby's CS6 book arrived and CS6 software has been installed in my computer.

 

Thanks for help.

 

All the best,

 

Thomas Chen

[jaapv]

Ok, learning curve ahead. Note that Scott Kelby is very good, but only scratches the surface of the possibilities this program has.

[Thomas Chen]

Some technical material may well be potential parts of his pro-Leica presentation, for our consideration and feedback.

 

Bokeh + MTF diagram = better 3D-like rendition? (4-1)

Ladies and Gentlemen,

Thanks to Eric’s indicating a mistake in Fig. E, it has been straightened out in this post.

Leica lense with 3D impression rendition that Leica pure M users seldom experience will be discussed here.

Fig. H, I, J, K are MTF diagrams for four respective Leica lenses.

The reference sources are identical from post (2), no repeat here.

The most upper curve in MTF diagram is for 5 lp/mm, while the lowest curve is for 40 lp/mm. Quote: the 5 lp/mm curve describes the reproduction of very coarse image details; 10 lp/mm for clearly visible details; 20 lp/mm for very fine details, and 40 lp/mm for the finest details of the subject [3]. Curves where the tangential and sagittal values are identical in the whole field are called “ideal MTF curve”, i.e. “bokeh” is particularly good [5].

Low contrast values for 5 and 10 lp/mm indicate a flat image; high value for 20 and 40 lp/mm indicates fine details reproduced cleanly and clearly separated. When tangential and sagittal curve are widely separated, this often means a blurry reproduction of subject details [3].

Digital sensors with 24M pixels in 35mm FF format having Nyquist frequencies of about 90lp/mm, its maximum resolutions are roughly comparable to the color negative film. For a A4 print from a distance of 24cm, the resolution human eyes can resolve corresponds to 66lp/mm in the senor image, thus, the 40lp/mm in MTF is important for human eye in large monitor[5]

When examine the MTF of Leica APO Summicron-M90 f2.0 ASPH, one finds that tangential and sagittal values are identical in the whole field, indicating “good bokeh”. However, the sense of sharpness against unsharpness is related to the field of depth, the subject’s contour and light condition, rather than the “depth perception” implied in the curves of MTF [3].

If one makes a portrait for a bady, it is likely that his face will be filled inside the circle of 9mm radius from the center of image and the background is in the area between circle of 12mm radius and 9mm radius. The depth from his nose to ears could be around 12mm.

to-be-continued

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Thomas Chen]

Hi Bert...

 

Some technical material may well be potential parts of his pro-Leica presentation, for our consideration and feedback.

 

Bokeh + MTF diagram = better 3D-like rendition? (4-2)

In this case using Leica Summilux R-35mm f1.4 lense set at f2.8, shooting distance 1 meter (giving a DoF of 12.3mm), the reproduction of his face within radius of 6mm from image center reach 65% contrast for 40lp/mm curve, 90% for 20lp/mm curve with almost overlapped sagittal and tangential curve, indicating very crisp and very fine details reproduction of his face. Starting from 6mm radius all the way to 9mm radius the reproduction contrast by 40 lp/mm rapidly drops to 20%, 20, 10 and 5 lp/mm curves also decreasingly drop, rendering a depth perception in the area from nose to chin and cheek by the smoothness of the unsharpness gradient, against the “bokeh” created by large aperture wide open. This justifies the statement by Mr. Olaf Stefanus:”….what we extract from the MTF curves is just as valid for the visual reproduction of any given three dimensional object.”[1]

Based on MTF of Summilux R-35mm f/1.4, it’s the best in the potential for depth perception, next is Summilux M-35mm f/1.4 ASPH, Summilux-M 24mm f/1.4 and Summicron R-35mm f/2.0. Some other R, M lenses also enables the depth conception, however, to a less extent.

I’m confident in the proposition mentioned above as I witnessed the 3D impression of Leica R 35/1.4 in film photography many years ago.

If a good raw-to-Jpeg conversion is in place, the M240 is entitled to be the best gift that Leica presents to so many Leica aficionados in the eve of 100-year anniversary of Leica 35mm format camera. It not only offers a great deal of innovations but also opens a opportunity for pure M users to appreciate the power of Leica R lenses in terms of “the best” 3D-like rendition they have yet experienced.

All the best,

Thomas Chen

Fig. H Summilux-R_35mm f1.4.pdf

Fig. I Apo-Summicron-M 90 mm f2.0 ASPH.pdf

Fig. J Summilux M 35mm f1.4 Asph.pdf

Fig. K Summilux-M 24 mm f1.4.pdf

[Lindolfi]

Thomas, there is a short paragraph on "bokeh" on page 33 in the paper by Zeiss on MTF curves, which can be found here: click

 

From that paragraph, you quote a single line, which is not the full story. The rest of that paragraph explains it well. If you look at the figure on the bottom of page 32, the x-axis is labelled, "defocusing". So this particular kind of MTF curve plots the modulation in the image space at a frequency of 20 cycles per milimeter perpendicular to the focal plane.

 

If you move outside of the focal plane for only 0.4 mm, you have lost modulation, so the detail in the image is gone.

 

Bokeh is the quality of the image outside the focal plane and is dealing with much lower spatial frequencies. In fact, the spatial frequency and its modulation is quite useless to describe it, which is very well explained in the paper you quoted from. Point spread functions are more revealing. However, also these functions tell only part of the story in a technical way.

 

So in your quest to get the best image out of a Leica M 240, while concentrating on the choice of lenses and while interested in 3D like rendering, you should not look at MTF curves, but look at photographs taken with those particular lenses.

[jaapv]

Let's quote just one sentence from Bert's link:

 

It is not possible to use MTF data to draw conclusions about the brightness distribution of the strongly defocused point spreads, however.

[Thomas Chen]

you should not look at MTF curves, but look at photographs taken with those particular lenses.

 

Bert,

 

Thank you very much for your comments.

 

You are right, the picture is the ultimate outcome.

 

What's I'm doing now is just a "proposition" that needs to be tested by "empirical" means, i.e. looking at photographs taken with those particular lenses . Or may I say I'm "preparing fro M240" ?.

 

I will explain my stance in next and next two post.

 

Best Regards,

 

 

Thomas Chen

[Thomas Chen]

Let's quote just one sentence from Bert's link:

 

Bokeh + MTF diagram = better 3D-like rendition? (5)

 

Ladies and Gentlemen,

 

Examining the MTF diagrams for Leica M & R lenses, one can find out that a large portion of lenses at f=2.8 feature the specific curve described by Mr. Erwin Puts (quoted from “Puts Column 28-35 mm Leica R Lenses_en.pdf”, Leica Website): ”The Leica-typical dip in the zone performance at image height of 12mm can be seen in the curve. It can be the case that the well-known smoothness of unsharpness gradient for which many Leica lenses are famous can be attributed to this characteristic.”

 

To make the best of this characteristic, for each image height positions from center: 6mm, 9mm, and 12mm at the horizontal X-axis, via 40 lp/mm curve in MTF to get respectively thecorresponding reproduction contrast % on the vertical Y-axis, marked with blue, green, and red colors (please see Fig. L & M).

 

And mark the gap between the maximal and minimal contrast % from image height positions 0 mm to 12mm at the top of chart, e.g. (76) in the case of Summilux-R 35mm f/1.4 lens.

 

I apply these two sets of procedure for all lenses in my possession. Thus, by the gap I can easily identify which lens is most suitable for a portraiture project with particular demand in certain degree of 3D impression. While the difference of contrast % presents the trend how the unsharpness gradient will be in between, triggering a mental simulation to speculate upon the rendition of the to-be image.

 

The choice for R or M lens of the same focal length leaves to one’s discretion, taking into account other factors such as lens weight, size, minimal focusing range or the smallest subject area, in addition to what is revealed in MTF diagram.

 

This is the way by taking advantage of MTF diagrams offered by Leica I enable myself to horn photography skill together with M240, aiming to acquire “the best” 3D image impression that Leica monopolizes and no others alike. However, only empirical outcome can validate my proposition.

 

All the best,

 

Thomas Chen

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Lindolfi]

Thomas, the fact that a certain M or R lens has a dip in the MTF curve of 20 linepairs per mm at a certain distance from the center of the image has nothing to do with 3D rendering of a portrait. If there is still some modulation at 20 linepairs per mm, it means that the lens is not extremely sharp, but the plane of sharpest focus of the portrait (for instance through the eye lashes) is of decise influence on the 3D impression when a wide aperture is used. Defocussing in the image is the way for the photographer to obtain 3D impressions, not some detail in an MTF curve.

 

Again: 3D rendering has nothing to do with MTF curves.

 

Thomas, before we continue, can we agree on that? Else we get in a situation that is a bit like a Decca record that is broken: the needle stays in the same groove over and over again

[Erik Gunst Lund]

Thomas, this is now getting ridicules IMHO... Sorry I fail too see any real facts in your statements, if you where on to something you would be able to select a lens from how the MTF curves look, not going to happen in the real world.

 

Why don't you look at some images from different lenses and select the ones you like, you could also mount them on your Leica M8 and see for yourself which lenses you like... you have almost all of them...

 

Strange quest your on....

[Lindolfi]

The quest of Thomas is based on three pillars (or what he calls propositions):

 

1] You need a fullframe sensor rather than an M8 crop sensor to get good bokeh

 

2] You need a good on-camera raw to jpeg converter to get the best images

 

3] From MTF diagrams you can determine which lenses produce the best 3D like images

 

 

Usually I have a weak spot for people with a quest. But in this case the three pillars of the quest of Thomas all tumble down and so it is very hard to support such a quest, especially when it is hard to agree on progress in insight made.

 

All the three propositions can be tested and have been tested.

 

It is up to Thomas how to continue. We have now seen 5 chapters of a monograph he is writing on "Managing Viewing the Best from M240".

 

It would be my suggestion to improve the foundations of that work, else any conclusions could be invalid.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[Thomas Chen]

Thomas, this is now getting ridicules IMHO... Sorry I fail too see any real facts in your statements,

quote]

 

Erik,

 

In the subject "Bokeh + MTF diagram = Better 3D rendering? "

 

I just quote literature from Leica and Zeiss publications.

 

As for what I'm going to do selecting lenses in the context of portraiture in last post, is the approach to test the ball park relationships among Bokeh, MTF and 3D-like impression.

 

It's not an ordinary way in photography, in this regard you're right.

 

Best Regards,

 

Thomas Chen