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

[Dubious Drewski]

So we all know and revere the L glass. I do. But a recent test with one of these lenses here at work has shown me something I don't understand.

I was just goofing off trying to find out just how much this 70-200 2.8 L lens would kick that cheap Pentax 55-300. (At a $1800 higher cost, it had BETTER!) I photographed a camcorder which was sitting about 1m from me. Here's what I saw:

Animated Gif

Yes, the Pentax has some real noise issues which I've never liked, but look at the clarity of the image. With the L lens, there's chromatic aberration and general softness while the Pentax has none. Yes, I retook this shot again and again with both cameras to confirm. Yes, I shot different things. Here it is again:

Animated Gif

What's going on?

By the way, I chose 100-120mm and f5.6-f6.7 for each lens because that's where they're sharpest.

 

[Steph]

How did you do the test? It seems to me a bit dubious to compare 2 lenses on 2 different bodies.

 

[Dubious Drewski]

Dubious? Touché!

But I completely disagree that it's unfair to do the test on two body types. Even if I could somehow get both lenses to fit onto one body, the results would mean nothing because in the real world, Canon lenses are used on Canon bodies, like Pentax only fits on Pentax. So the practical results are all that matter. Hence the pixel poking.

And for what it's worth, half of the Canon shots were taken with the 40D. It had the same results.

 

[tirediron]

Interesting - I certainly don't have an answer for you, but it looks like it bears looking into further.

 

[Garbz]

Extra sharpening applied by the Pentax? No it really is dubious. You gain no knowledge of a lens without knowing the fine details of the equipment you are measuring with. It's like trying to measure volume using a microphone you don't know any sensitivity specifications for.

All you have shown is that the Pentax lens on the pentax body looks noisier and has slightly more edge definition than the Canon lens on the canon body. This is far from a lens comparison.

 

[Dubious Drewski]

Well I definitely agree this is far from a solid, empirical comparison, but I took literally dozens of shots with the same result each and every time. The cameras had all of their settings on default, so that's not an issue. If there any variables I missed, I can't think of them.

And Garbz: look at the first Gif, that's chromatic aberration there. No setting in the camera will increase or reduce that. That's all in the lens.

 

[Big Mike]

I didn't think that the 70-200 F2.8L was capable of focusing at only 1m...try it again with the subject 3m away.

 

[tirediron]

I didn't think that the 70-200 F2.8L was capable of focusing at only 1m...try it again with the subject 3m away.

Well, that would certainly have an effect!

 

[usayit]

* Why are the apertures different?
* Why are you shooting with two different sensors?
* Are processing completely turned off?
* Why are you focusing at 1m? The 70-200 f2.8 is not capable at focusing at that close distance?

If lenses are a focus the best way is to take the sensor out of the picture. Shoot two FILM camera bodies with their lenses at the exact same distance using the same roll of film (midroll change). Take that roll and scan it with the same scanner in a single run. Even then.. after all that effort... the test is pretty useless. It ignores so many other cases.... and situations.

A valid test comparison of equipment (of any type... photo, auto, anything) is a WHOLE lot harder and effort intensive as you might expect.

 

[ksmattfish]

So we all know and revere the L glass. I do. But a recent test with one of these lenses here at work has shown me something I don't understand.

What's going on?

It's like learning Santa isn't real isn't it? You spend 4x the moola on a lens, and you'd think it would just kick-ass!

I've rented a few Ls, and I own a 24-70 f/2.8. Optically (sharpness, color, contrast) none of them have significantly outperformed the similarly speced Sigmas and Tamrons I've used/owned. Build quality of the Ls seems better, but how how bash proof can a lens get anyway? AF is often better, although the Tamrons and Sigmas usually work fine. The Ls are typically bigger, heavier, and much more expensive (x2 to x6 times similarly speced lenses). IS is definitely a very nice feature.

I love my Tamron 28-75 f/2.8 ($350ish). It's slightly sharper and flares less than my Canon L 24-70 f/2.8 ($1200ish). It seems half the weight too; that's the big difference. If the Canon had IS I'd probably prefer it (but then I would have paid $1600).

On the other hand I've experienced a few equipment failures (breaking with normal use) with Sigma and Tamron lenses, and have never had a Canon lens (L or not) fail. Considering the price I could buy 3 Tamrons and still come out ahead.

My advice is to ignore the hype, test the lens you are thinking of buying yourself, and see if you can spot the $800 to $1500 difference. If so then it's worth it. If not then you can afford 4 new lenses! Yippee!!!

 

[Battou]

* Why are the apertures different?
* Why are you shooting with two different sensors?
* Are processing completely turned off?
* Why are you focusing at 1m? The 70-200 f2.8 is not capable at focusing at that close distance?

If lenses are a focus the best way is to take the sensor out of the picture. Shoot two FILM camera bodies with their lenses at the exact same distance using the same roll of film (midroll change). Take that roll and scan it with the same scanner in a single run. Even then.. after all that effort... the test is pretty useless. It ignores so many other cases.... and situations.

A valid test comparison of equipment (of any type... photo, auto, anything) is a WHOLE lot harder and effort intensive as you might expect.

I agree with this completely, even if you can't put the two lenses on the one body there are still technical discrepency in sensor tech that completely moot the entire testing session if the lens is the focus.

 

[Overread]

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

 

[Alex_B]

Well, redo the thing at ISO 100

in camera sharpening will be different with both cameras, definitely.

on the second gif, exposure seems to be quite different.
how about focus? did you focus manually, do focus series and only use the best?
did you check the centre of the frame, and the four corners (or at least one of them)? The latter point is extremely important! no serious comparison can be done by looking at one region of the frame only.

as stated before, comparing lenses in a system consisting of more different components than just the lens, is not at all easy. you have the lens, the different body, probably different focus, different exposure, different, translation of sensor data into the image (the colours are different!), and different sharpening.

 

[Alex_B]

* Why are you focusing at 1m? The 70-200 f2.8 is not capable at focusing at that close distance?

totally missed that one, but true! try a 10 m focus or similar.

 

[Village Idiot]

If you're shooting JPG, that can make a huge difference. Just because a camera is set at it's default settings doesn't mean that it's going to be the same amount of sharpness and contrast adjustment as every other camera.