This one's for the pixel-pokers. Pentax vs Canon L! What's going on here?

[Bifurcator]

It's not a conclusive test or revelation but one does learn something from it. It would be awesome if you could test both cameras with both of those lenses. That would be much more conclusive. It would also be better as some people are pointing out, if you would adjust everything to it's lowest common denominator. ISO, S, A, Distance, Lighting, RAW, etc. I would also recommend downloading a resolving chart and printing that out for one of the tests. If you click on my sig link there's a free chart at the bottom of the link list.

 

[Dubious Drewski]

Well, I suppose it wasn't a 1m distance. What I did was put the 70-200 into manual, check the closest focal distance, back up an inch, then put it into auto for the shot. I judged distance wrong. But it doesn't matter, I took shots from all distances.

I need to head to work, but when I come back, I'll have some RAW shots instead, tripoded, etc.

By the way, I shot each lens at differing aperture and focal length because each lens' peak performance is at slightly different spots.

 

[Big Mike]

By the way, I shot each lens at differing aperture and focal length because each lens' peak performance is at slightly different spots.

What reference are you using to find out what the peak performance settings are?

 

[prodigy2k7]

I think the 70-200's peak is at 100mm F/5.6 but im not 100% sure

 

[ksmattfish]

One bonus is most (I think not all) L lenses are weathersealed. Couple that with a weathersealed camera and its a very good thing for those of us who shoot out and about

Yeah, that'd be great, if I could afford those bodies. Canon may be the only company out there that doesn't think a $3000 DSLR body is worth weather sealing. Those rubber gaskets must be expensive!

 

[prodigy2k7]

Yeah, that'd be great, if I could afford those bodies. Canon may be the only company out there that doesn't think a $3000 DSLR body is worth weather sealing. Those rubber gaskets must be expensive!

Im sorry what model does canon make that is $3000?

 

[Overread]

lol - well I think canon were trying to boost their FF department in the middleweights area, now with the new nikon releases just announced I think canon will up the stakes a little - possibly going as far as to challenge thier current flagship top line cameras with new middle level ones - and then releasing quickly a new line of flagship cameras

 

[Joves]

Im sorry what model does canon make that is $3000?

The 5D kit is that much. But the body is only $2400.
Your test does need to be set up a little better. Both at same distance/aperture/focal length/ISOs and, you need to vary the ISOs for a camparison at a given speed. Also any sharpening and noise reductions need to be off.

 

[usayit]

Also to add...

Most Canon DSLR users will notice that RAW straight out of the camera is flat thus also "seems" unsharp. The data is all there but stored in a VERY neutral way. This was a hot topic in the early days when Canon DSLRs were slowly hitting the market at less than stellar prices. When the typical photographer was still making an adjustment from a film workflow to digital workflow many ended up confused. This is/was an intentional design with the basic assumption that every RAW capture is intended as an intermediate step towards final print and not the final step towards final print. Canon's RAW (perhaps Nikon and others as well) file are intended to be post process since the final intention (contrast, saturation, size of print etc) is completely unknown to the designers of the camera. In contrast, a film photographer would choose the format as well as the film knowing full well the intention of the final print.

As such, its difficult to make an assessment of two different lenses when the sensor from which the data is gathered are from different cameras that leverage different processing techniques. A Canon photographer accustomed to Canon RAW files is going to be able to create a much better final print from a Canon RAW file than a Nikon RAW file. A Nikon photographer accustomed to Nikon RAW files is going to be able to create a much better final print from a Nikon RAW file than a Canon RAW file.

This is the reason why the sensor and processing being performed by the camera has to be eliminated if your true intention is an assessment of lenses.

 

[Dubious Drewski]

So that last test wasn't well thought-out. This time, I decided to exclude the results of the lens tests from any other brand and just show you guys the Canon results.

I addressed the issues people brought up with the results I posted last time. This time, I tested all lenses on one body: An XSi, shooting Raw, on a tripod with a 2 second shutter delay, with each lens at 100mm and f5.6 (Except for the F4 L lens), focusing on objects 2m away. The lenses had image stabilization turned off while sitting on the tripod, if they had that feature. I took each picture 3 times, and picked the sharpest of the three to show you.

Here are the results from 5 telephoto lenses. 200% crops:
Animated Gif
And for the people who can't be bothered to stare at a 15 second animation:
Side-by-side layout

Looks like that 70-200 2.8 doesn't perform so bad when compared to the other similarly- ranged telephotos.
(Though it's still not as sharp as the Pentax, for some odd, strange reason)

What do you guys think of these results? Are they fair?

 

[Bifurcator]

is #3 in focus?

 

[Dubious Drewski]

I know it seems very out of focus, but that's what I got 3 times in a row. I was using the autofocus.

 

[Bifurcator]

Wow.

And those are all on the canon body?

 

[Overread]

quick question - are these RAWs processed past RAW conversoin - ie has sharpening been applied?

 

[Alex_B]

So that last test wasn't well thought-out. This time, I decided to exclude the results of the lens tests from any other brand and just show you guys the Canon results.

I addressed the issues people brought up with the results I posted last time. This time, I tested all lenses on one body: An XSi, shooting Raw, on a tripod with a 2 second shutter delay, with each lens at 100mm and f5.6 (Except for the F4 L lens), focusing on objects 2m away. The lenses had image stabilization turned off while sitting on the tripod, if they had that feature. I took each picture 3 times, and picked the sharpest of the three to show you.

Here are the results from 5 telephoto lenses. 200% crops:
Animated Gif
And for the people who can't be bothered to stare at a 15 second animation:
Side-by-side layout

Looks like that 70-200 2.8 doesn't perform so bad when compared to the other similarly- ranged telephotos.
(Though it's still not as sharp as the Pentax, for some odd, strange reason)

What do you guys think of these results? Are they fair?

slowly getting there. however, you should not be using autofocus I think. also, you need the centre of the frame and the edge of the frame to really draw any conclusions. 200% crops will look very different at the centre, and far away from the centre. in particular with respect to CA, but also general sharpness.