Every Sigma exposure is darker than the Canon's. It would appear that the Sigma's letting in less light at every f-stop compared to the Canon. So apparently the Sigma's not only less than 85mm in focal length, but also less than f/1.4 wide open....more like f/1.6.
I stacked the Sigma and the Canon images in Photoshop and then applied a levels layer to the Sigma moving the center arrow in the levels window to the left so that the white areas of the newsprint matched the Canon. When the white areas of the newspaper article were matched, the Canon's dark areas (the article's photos) were darker (had better contrast) then the Sigma's.
I dont doubt the Sigma will be a great performer. We can just hope these 85 siggy's will be consistent regardless of distance. I recall hearing wondeful reviews on the initial production of the Sigma 50mm's. Purchased one and appeared to work well under controlled tests until I used it at a very important event. Fortunately I didn't throw that sigmadud against a wall.
As we are so critical I think we'd have a hard time to tell the difference between the two lenses on print. I think in many cases I stop down to f/2 to f/2.8 and I've always relied on my Canon 85 f/1.8 to land the shot with the fast USM. Portraits I'll bring my 85L along as a different tool. 85 f/1.8 isn't bad for CA or PF when stopped down. Both my Canon 85mm primes have been rock solid in AF.
It's a bit difficult to compare the best a lens can do, unless they ae shot from the exact same point with the exact same settings, manually focused to the best possible.
Moving the tripod to compensate for diffenent focal lengths just makes the results suspect. Its best to present results exactly as they are.
Processing is another area where differences can pop up. Make certain that the software is not recognizing the 85 L and adjusting something. I'd particularly avoid DPP, since it does recognize Canon lenses and will tweak results depending on your settings.
Can you just crop the tire drawing?(whenever you get some time and will) I would like to see the difference in the textures for this exact little portion of the picture.
I'm curious about one thing; Is this lens showing the onion bokeh too? My Sigma 50 mm 1.4 (sold already) was showing this a lot.
Aaron, I deleted the CR2s. I'm going to redo that test, starting out far enough away that I won't have to move the tripod back for the L. And I'll post 100% crops from the center and corners. I've got things to do, so it may be a few days, but I'll do it.
I checked the EXIF, just to make sure that the exposure time is the same for both lenses at the same aperture.
Given how much darker the sigma samples are, the light gathering ability is certainly nowhere near f/1.4 I can't tell if this is a T-stop difference or an F-stop difference.
Hard to tell from JPG's exactly how much it amounts to, but I guess the Sigma is in reality an 82 mm f/1.6 lens. Game over for me. I'll keem my Canon 85/1.8.
These tests run counter to the ones outside, as you reported seeing more contrast on the 85L, so perhaps it is wise to MF with LV to exclude any focus issues, since you are testing IQ.
I checked the EXIF, just to make sure that the exposure time is the same for both lenses at the same aperture.
Given how much darker the sigma samples are, the light gathering ability is certainly nowhere near f/1.4 I can't tell if this is a T-stop difference or an F-stop difference.
Hard to tell from JPG's exactly how much it amounts to, but I guess the Sigma is in reality an 82 mm f/1.6 lens. Game over for me. I'll keem my Canon 85/1.8.
...Actually, the "Mighty L" is only 83F/1.3 (1.265) at infinity, so the Sigma being an 81F/1.44 (not confirmed by measurement yet, but the drawings indicate this) is no big deal... The problem is more likely how much the aperture action works. Either the Sigma stops down to much, or the L stops down too little - or both. This is actually the case with most lenses, compare two EF 50F/1.4 lenses at F/8.0 and they will NOT give the exact same exposure.
So, if you really want to know the T-stop relative to the L, shoot manual shutterspeed and wide open. And DON'T use DPP that contains a lot of correction stuff hidden behind the scenes for native lenses...
My guess is that the Sigma then will show almost exactly one sixth (1/6) Ev lower exposure - the same as the Zeiss, Pentax, Sony and Nikkor compared to the 85L (they are all ~ 84mm F/1.44).
theSuede wrote:
My guess is that the Sigma then will show almost exactly one sixth (1/6) Ev lower exposure -
I find that guess optimistic, given that in this test the Sigma wide open (no aperture action) seems to be more than 1/6 Ev darker than the L stopped down to nominal 1.4, and the L has to be at least a litte darker stopped down than wide open...
alundeb wrote: theSuede wroteMy guess is that the Sigma then will show almost exactly one sixth (1/6) Ev lower exposure -I find that guess optimistic, given that in this test the Sigma wide open (no aperture action) seems to be more than 1/6 Ev darker than the L stopped down to nominal 1.4, and the L has to be at least a litte darker stopped down than wide open...
But as theSuede said, the OP used DPP, which may have done auto corrections for vignetting etc. Open them in ACR or other non-Canon browser and turn off all lens correction options, then compare them.