[Blue]

having a larger element doesn't necessarily mean faster f-stops. e.g. telephoto lens versus standard lens. f-stop brightness or light gathering ability of the lens depend on the whole optical design.

Larger elements on the front was done in some cases to improve optical vignetting in wide angle lenses which is independent of f-stop, but can be related. A good example there would be the Auto Tak 35/2.3 lens.

[Dizono]

My pentax 28 3.5 shift has an ever bigger front element and has no vignetting so i think you're right

[whojammyflip]

I picked up a copy of the f3.5 and have done a quick comparison with the f2.8. Tipod mounted, mirror lock up etc. using f5.6 on both the f2.8 and f3.5 M series lenses, on my ist DS, about 1.5m from the Economist newspaper, which is fairly shiny and reflective, as well as being well printed (and was in the bin for being a little dull).

I can't see any difference in contrast, and the f2.8 may be better! Neither managed to resolve the small text. Maybe the focal length of the f3.5 is slightly greater, as these are 100% crops, measuring exactly 999 pixels wide, without moving the tripod and the f3.5 pic looks slightly larger.

To quantify contrast, I took a 3 pixel average (using the curves dropper in GIMP) for the black area under the tie in David Cameron's suit, vs the white next to the text on the RHS. The f3.5 scored (171-17)/(171+17)=81.9% and the f2.8 scored (173-16)/(173+16)=83.1%. (If someone could write a GIMP plugin to calculate the average accutance for a picture, it would be very handy for contrast testing!).

Maybe there is an advantage on full frame to this f3.5 lens, and I have not shot with it out doors yet, so maybe it handles flare better? Suggestions are welcome for further tests as I don't want to get rid of the f2.8 too soon. One weakness I can see to this kind of testing, is that its dependent on my manual focussing skills without a split screen on the DSLR and my ist DS only goes to ISO 200.

Lens M 2.8/28, iso 200, 1/125 at f5.6


Lens M 3.5/28, iso 200, 1/125 at f5.6


Quick test shot, M 3.5/28 with 20mm hood, f5.6, 1/125 handheld

[Marc Sabatella]

I can't see any difference in contrast, and the f2.8 may be better

Looks like you're setting up some decent tests and have a workable method for actually measuring things - good job there. Seems to confirm what I see more casually - whatever difference there is is barely worth mentioning. Although of course you'd want to test at different apertures too.

Maybe the focal length of the f3.5 is slightly greater, as these are 100% crops, measuring exactly 999 pixels wide, without moving the tripod and the f3.5 pic looks slightly larger.

Could also conceivably be a difference in the actual focal plane as opposed to length - in other words, the f/3.5 might be effectively forming the image from slightly closer due to the optical design? Complete speculation; I have no idea how plausible it is.

One weakness I can see to this kind of testing, is that its dependent on my manual focussing skills without a split screen

Sort of, but not really - the nice thing about focusing on text on an angle is that it doesn't matter if the point of focus is *exactly* the same, as long as you identify the most in focus portion of each image and compare those. Actually, by chimping the LCD immediately after shooting, you can with some trial and error get the same lines in focus - the poor man's live view :-).

It will be interesting to hear your impressions as you use the lens more. But now I'm actually thinking of trying to rig myself up one of the those stepping ring hoods so I don't have to keep sharing with my DA70.

[whojammyflip]

Thanks for the comments. A quick thought is that I am not really measuring anything other than very low frequency MTF with my calculations, as the areas are large where I have sampled the black and white values. I think the only way to really go about this is with a resolution test chart, so my conclusions on the difference in contrast really only apply to veiling glare and nothing else.