MTF tutorial, and 100-400mm vs. 400 prime

[drh681]

yeah,
 you have to look at the whole graph not just left side.
 that represents the center of the test image,
the right side represents the corners of the test image.
if you look at the graph for the 400 5.6 you will see the lines go fairly straight across.
on the graph for the 100-400 the lines drop off and spread out to the left of the the graph. (that is why we call it "lens fall off")

[drh681]

oh.
you want to know in terms of a sensor size he was not discussing in the article.

that is the good news about using a smaller sensor. yo are using "the sweet spot" of a lens that may have soft edges in a full frame image.

so look at where the lines cross the 10mm mark, that is out near the edge for the 22x15 sensor (for 1/2 the 26mm diagonal of the sensor) in fact, at that 13mm mark, the 400 5.6 is basically a flat line MTF! while the 100-400 is well into the fall of zone.

[pixelcounter]

In the excellent tutorial on MTF,
http://www.luminous-landscape.com/tutoria.....shtml,
the following summary points are made:

The closer the black and blue lines are, the better the lens is wide open
When the solid and dashed lines are closer to each other, the better the bokeh.

The higher up the chart the thick lines (10 LP/mm), the higher the contrast.
The higher up the chart the thin lines (30 LPmm), the higher the resolving power/sharpness
When thick lines are above .8, the lens will have excellent quality

Then, in the comparison of the Canon 400mm f/4 and the 100-400mm f/4.5-5.6
http://luminous-landscape.com/reviews/lens...otten-400.shtml
examples show that the 100-400mm is a good deal softer.  

What troubles me is that the MTF curves
included in the review for the 100-400 show that the thick lines are together at .98 at 20 which means, according to the tutorial, that the lens provides “high contrast” and is “better” wide open and, being greater than .8, of excellent quality.  The thin lines being “close” at about .8 also favor quality at wide open and “better” sharpness.  

Further, the solid and thin lines for the 100-400mm are higher than those of the 400mm prime.  Except for the higher positions of the dashed line for the prime, one would think the zoom is superior.

Can anyone provide any perspective on this?  Might  the softness that appears in the test shots be the result of focusing?  Considering the differences in dashed line positions, is the pictured softness a bokeh issue?

[pixelcounter]

I am asking about the right side.  As I noted in the original post, I'm discussing the curve values with the x-axis at 20, i.e. beyond the width that would be appropriate for 35mm.  For a 1.6 crop such as in the 20D the portion of interest is only 0-14.  At 20, the thick lines are at about .98 which is supposedly "high contrast" and "better sharpness".  Still need some insight here.

[Ray]

There's too much variation in general to be precise about such matters.
As a rule, the higher up the chart and the flatter the lines, the better the lens, whatever resolution and f stop the lines represent. If the MTF curve representing 10 lp/mm is close to 1, indicating virtually no loss of contrast at that resolution, then the lens could be considered a contrasty lens. But it might not be a high resolving lens if the curve for 30 and 40 lp/mm is quite low down the y axis.

I believe the charts in Michael's tutorial are computer generated theoretical curves published by Canon. They are probably more representative of how Canon envisaged the design of the lens than of any real lens. My copy of the 400/5.6 prime proved to be less sharp than my 100-400 IS.