Who's at fault, camera or lens?

I

insertusernamehere

Messages
1,598
Name
Adam (if you couldn't guess...)
Edit My Images
No
So I am in the fortunate position to have two sigma 120-300mm f2.8 lenses.
One nikon, one canon.

The nikon is a D and I love it
I recently acquired the canon DG version and its not doing it for me, I'm not sure whats at fault the sharpness or the focus, so I thought I would do a test:

Unfortunately my 1D won't be back until tomorrow so I plan on retesting once I have that.

So I took the same shot with a D90 and a 40D, the only PP is auto colour, the results were surprising to me

see new photos below
 
Last edited:
I'm not surprised!

You have not used the same settings on both cameras so you can hardly expect similar results?
 
Chuckle, I am not that daft.
However I am daft enough not to have picked up the right images!!

Let me check which images i've picked up!!!
 
Grrrrrrrrrrrrrrrrrrrrr i've deleted the others.

I'll reconvene tomorrow because quite frankly I am just too lazy to set it all up again right now :-D
 
Ok so its bugging me.

I think these should be exactly the same settings, if not then please do tell me what I am missing.... I can only think that I must be missing SOMETHING on the 40D!!

iso 400
1/30"
f2.8
0 exposure comp
Auto focus
AWB
300mm
full metering
Tripod mounted, in exactly the same place
Single Centre focus point

NO PP now

Canon:


canoncrop2.jpg



Nikon:

nikoncrop2.jpg
 
Pop it into live view (I think the 40D has liveview doesn't it) and focus it there - see if the lens/camera combination is off in any way.
 
Although the Canon shot being a bit of a duffer could be down to the lens not liking f/2.8, it does look like there's a tad of camera shake.

How do both lenses compare when they're at f/4?
 
AF calibration could be off. If it is then the culprit could be either the camera, the lens or both. Can you improve on the focusing accuracy using Live View and 10X magnification? If you can then you possibly have an AF calibration mismatch. If you can't then maybe the lens is soft.

Were they shot raw or JPEG? If JPEG you can't assume much about what the in camera processing did or didn't do and I don't think you can easily make direct comparisons.

1/30 is in the sweet zone for highlighting mirror slap. Did you use MLU? Was the tripod on a hard floor or carpet? Did you use a self timer or remote release? Was anyone moving in the room (affecting the floor) when you took the shot?

Sometimes you get a focus glitch. Are the results from the Canon combo consistently duff? Is the Nikon combo consistently sharp?
 
Information overload!!

Hehe, the reason for the test is that in the 'real world' I was finding the results from the canon sub-par on my 5D and then the 40D on the one day I used it.
I didn't want to get too laboratory on it because I think these results pretty much reflect my experiences.

I'll wait til tomorrow and try the canon on my 1Ds ii and see what sort of results I get, taking into account some of the comments here!
 
That's an interesting comparison between lens resolution and contrast.

The first image looks to be close to the second in terms of resolution, in that most of the fine lines can be seen if you look closely (which also suggests it's focused, and reasonably shake-free). But the second image has loads more contrast, ie much higher MTF values.

That's what sharpness is all about - contrast (in a way that cannot be put back in post processing) rather than resolution or loads of pixels.
 
adamdavi3s said:
Hehe, the reason for the test is that in the 'real world' I was finding the results from the canon sub-par on my 5D and then the 40D on the one day I used it.
I didn't want to get too laboratory on it because I think these results pretty much reflect my experiences.
Click to expand...

Still, taking a carefully manually focussed shot and comparing the to an AF shot will reveal any focus calibration problems that will make AF images seem soft as the focus would be consistently missing by a small amount. It should be a pretty quick trip to sigma to fix if that is the problem :)
 
Last edited:
I did try several MF shots earlier but they were worse, I didn't realise that the 40 had live view though so I'll have a crack at that :-)
 
adamdavi3s said:
I did try several MF shots earlier but they were worse, I didn't realise that the 40 had live view though so I'll have a crack at that :-)
Click to expand...
Also get it to focus itself in live view. The focusing system when the camera does it in live view is using the data directly off the chip so you should get "ideal" and correct focus.
 
Something else that made me chuckle

When I opened the canon image in CS2 it was ok.
When I opened the nikon one, it gave me a warning that the print option was disabled for bank notes.

Even adobe agrees with my issue :-D
 
Yup, the Canon sample was so blurred it didn't even register as a banknote!
Anyway, I'll second the suggestions to live-view manual-focus and then compare to AF, it looks to me like AF is well off.
Also, focus midpoint along a slanted ruler ... in addition to the flat focus test.
 
Ulfric M Douglas said:
Anyway, I'll second the suggestions to live-view manual-focus and then compare to AF,
Click to expand...
There's also live view contrast detect AF too - which should be 100% accurate as it uses data off the sensor rather than a separate autofocus chain.
 
arad85 said:
Also get it to focus itself in live view. The focusing system when the camera does it in live view is using the data directly off the chip so you should get "ideal" and correct focus.
Click to expand...

Not on the 40D, it was a model before they introduced contrast detect AF in live view, the only way it'll AF is by flipping the mirror back down to use the phase detect.
 
squishy said:
Not on the 40D, it was a model before they introduced contrast detect AF in live view, the only way it'll AF is by flipping the mirror back down to use the phase detect.
Click to expand...
Ahh...fair enough then.... Just goes to show what making assumptions does ;)
 
Ok these are all shot manually focused on live view and with a remote cable, looks like it just doesn't like F/2.8
Guess which is 2.8 and guess which is 5.6..... :-D

canon28.JPG


canon56.JPG
 
Considering the results possible from the Nikon version at f/2.8 there is really no excuse for the Canon equivalent to be so far adrift. If it was my lens I'd absolutely want it fixed. I just hope it is still under warranty and you can get it made right without cost. Good luck. :)
 
You must log in or register to reply here.
Back
Top