The Big Bad wrote:
The sigma 24-70 doesnt look to bad but you can see what the extra $800 or so buys you in the 24-70L, (aside from USM, weather sealing etc)
Don't mean to pick on ya, Jeff, but what am I missing here? Where's the $800 difference in these images? They look pretty close, to me. Certainly not an $800 difference!
20Dshooter wrote:
I looked at the originals, but, no. These are the corect order.
BTW, f/4 looks sharper to me. 5.6 is probably an AF error.
Well, the top Canon also looks sharper at f/2.8 than at f/4. Whether it's a focusing error or handhold blur - there is not much confidence in the results.
That's the main problem with tests - they either have to be done carefully or not done at all. If I can't compare shots taken with the same lens how can I compare shots taken with different lenses?
JCDoss wrote:
I'm still waiting on a Sigma EX lens of this quality and range to sport an HSM focus motor.
like some poster said, the speed difference is minimal between
L USM and the non HSM ,,,, in fact , a friend of mine who had the L
was surprised about the speed on the Sigma
So you dont really feel the difference
but back to back test, yes, there is that slight difference
but doesnt stop you from doing what you want..................
that's why you bought the lens in the first place, right ?
gfiksel wrote:
That's the main problem with tests - they either have to be done carefully or not done at all. If I can't compare shots taken with the same lens how can I compare shots taken with different lenses?
Absolutely. Which is why I take all of this stuff with a heavy grain of salt. Ultimately, I've learned you can't trust any of these kinds of tests unless A) you do it yourself or B) it's done by someone with a proven track record for setting up tests effectively.
The sigma 24-70 certainly looks like a good lens. This may seem silly, but the one thing that keeps me from getting one is the 82mm filter size! I just hate having to buy another polarizer. I realize that I can still buy the lens and a nice polarizer and still be way under what the canon costs. Why did the need to make it so big?
Neither one of the two lenses looks significantly better or worse in these examples.
However, I think, if you do a test like this, the subject should contain some primary colors, it should be in a stable environment, and it should be done from a tripod (mirror lock, remote release etc.), otherwise many of the differences could be accredited to any of these factors.
I want to see corner sharpness, because I have a feeling thats where the game is. Only until then will I say anything about the sigma's performance vs the Canon.
I was kinda curious what was going on with the soft focus of the bricks on the right in some of these images. I have highlighted those areas with a red arrow... I don't know if this lmage link will work...
http://parmley.smugmug.com/photos/31028559-L.jpg
Ok it works wooohooo! First time I got that bad boy to work! Anywho. You can see where my arrows are pointing, note those locations and go back to the original and see if you can also see the soft focus/lack of sharpnes in those bricks as well...
Clayphish wrote:
I want to see corner sharpness, because I have a feeling thats where the game is. Only until then will I say anything about the sigma's performance vs the Canon.
This will only matter on a full-frame camera. Otherwise, you're imaging through the center of the lens and cannot investigate the usual suspects that hide in the corners.
That being said, if the center of an image through one lens is significantly worse than on the other, it doesn't matter what the corners are like.
ishmaiel wrote:
This will only matter on a full-frame camera. Otherwise, you're imaging through the center of the lens and cannot investigate the usual suspects that hide in the corners.
That being said, if the center of an image through one lens is significantly worse than on the other, it doesn't matter what the corners are like.
I'm not so sure about that. Even if you're using a 1.6x camera you can see the weaknesses of a lens as you gradually move further away from the center. I've witnessed this with the Canon 50mm 1.8 when I compared it to the Takumar 50mm 1.4. In that test, even on my D60, did you see a major fall off in the corners of the frame. I can only imagine that it would be even worse on a full frame camera. I've also noticed this with the 20-35mm 2.8, and the 17-40 -- though its more understandable since they are wide angle zooms.
So when it comes to this test, I think seeing the corners is extremely important to dictate how well one performs, even if its not on a full frame camera.
I guess you are right. I was being too extreme when I said that corners didn't matter on a cropped FOV. I was giving both lenses the benefit of the doubt. If you're starting to see corner-type failings near the cropped center, then perhaps it's good you don't see the horrors of the edges
ishmaiel wrote:
I guess you are right. I was being too extreme when I said that corners didn't matter on a cropped FOV. I was giving both lenses the benefit of the doubt. If you're starting to see corner-type failings near the cropped center, then perhaps it's good you don't see the horrors of the edges
Yea, its a bit startling when you see enough of it on a 1.6x camera, but you're totally right, full frame is the real test. I remember using a 17-40 on my film A2e, and boy I was suprised. Even through the viewfinder I could see the soft edges. It kind of makes 1.6x seem like a good thing, if you disregard the extra reach that is.
Sam Bennett wrote:
Absolutely. Which is why I take all of this stuff with a heavy grain of salt. Ultimately, I've learned you can't trust any of these kinds of tests unless A) you do it yourself or B) it's done by someone with a proven track record for setting up tests effectively.
True. However, as I said originally, I never intended to provide scientific tests. They were done informally and quickly. That is why I slipped up, I guess, in the 24-70 tests. The EX f/5.6 shot is not sharp, and neither is the f/4 from the L. Probably focusing error. So, yeah, take it with some grains of salt, but I was trying to point out how close these lenses are to each other. The differences, if discernable, are slight, which is astonishing.
Clayphish wrote:
I want to see corner sharpness, because I have a feeling thats where the game is. Only until then will I say anything about the sigma's performance vs the Canon.
The 20D corners, or full-frame? The 18-50 will vignette significantly on a 1DmkII - a 1Ds will be much worse. It is intended for a 1.6x crop, so the corners will likely not be as good as the 17-40.
Parmley, you are right. This was already noted. Probably a focusing error. Slight as it is, it is only visible in a 100% crop. Egg on my face! I'll do a better job next time!
I'll do a better job next time!