17-55mm F2.8 vs 50mm 1.8 sharpness

Like musical instruments, lenses are subject to a certain about of variance. So your results are valid for your 2 samples but another 2 samples will be slightly different, sometimes vastly different. My experience with these two lenses is somewhat different: from 2 meters to infinity the 17-55 bested my 50 1.8 MK I from F2.8 to 5.8 in terms of both contrast and sharpness. By F8 they were basically indistinguishable. Close focus is a weak point of the 17-55 and, although it is decent, the 50 1.8 is better. My EF 50 2.5 CM pretty much wipes the floor with both other optics at both near and far focus. I'd say the 17-55 is a darn fine zoom and certainly approaches top prime quality from normal to infinity at all apertures.

I would think that asking a zoom, however good, at the end of its FL range to outperform a 50mm operating at or near its optimal aperture would be placing extremely high expectations on the zoom. Equaling a 50/1.8 across the field at f/8 is a pretty high bar.

I'm not surprised that your manually focused images were better than the AF; I get similar results if I'm paying attention to what I'm doing.

Thanks for the reply. I should have added that I'm pretty sure my 50mm is a good copy as I used to buy and sell 'MK1' 50 1.8s on eBay and the one I kept was the best I ever tested. I'm just trying to get an idea of whether my 17-55mm is performing 'normally' - a few times I've not been impressed with the results from it, although as always, user error is the most likely explanation!

Focus distance on the test was 1.6m.

Thanks again.

At f8 most decent lenses are very similar. By that aperture the spherical aberration should be negligible and the lens may be diffraction limited in the center.

The full frame lens would be expected to be better in the corners of a cropped frame simply because you're not at the edge of the image circle.

At (link) there is a review with comparisons: choose ISO 12233 Crops then pick the other lens to compare it to at various f/stops. The consensus is that the 50/1.8 @ f/2.8 should beat the stuffing out of that EF-S zoom at 50 mm wide open.

if you did not have IS disengaged on the 17-55 you may want to try the test again. this may affect results, especially in the corners.

my 17-55 matches my 50s in sharpness in the center at all f stops (i don't test for corners). the zoom outdoes the primes in brightness, dynamic range, color saturation, color palette, and contrast. the primes have less CA. if there is good light i will use the zoom. i only use a prime if i want better bokeh or if the light is bad. in good light the zoom produces images with more pop, while the primes produce a murky image by comparison. if i'm shooting in bad light and i'm using a tripod (and i don't need bokeh) i will use the zoom.

i have gotten bad results from the 17-55, though. it is harder to use than a prime. zooms are so much heavier and longer than primes, especially 50s.

Thanks for all responses so far.

IS was disengaged.

I was probably expecting too much of the zoom. I like primes (I have 15mm/28mm/50mm/85mm) and was using a Sigma 18-50mm F/2.8 as a 'walk-around' until it developed a focus fault (as Sigmas often seem to). After reading so many good things about the 17-55mm and wanting an IS lens *badly*, I splashed the cash.

I was a bit suprised by the size of the thing - with the hood attached it's bl**dy huge! That and the weight is likely to be the reason for the unsharp pics I've had and I could do to concentrate more on my hand-holding technique generally. Still, the Sigma is tiny is comparison and I don't see that adding USM and IS to it would make it as big as the Canon. Maybe I'm wrong.

Cheers

I have both, and the 17-55 is hands down overall much sharper.. at least mine. I have been unhappy with the clarity of my 50 1.8 as a matter of face. Maybe I have a bum one.

Shawn

terry

i don't think it's the usm and IS that makes the canon so much bigger. it's the focal range. 50 vs. 55 isn't that big a deal but 17 vs. 18 is. so put together 17-55 f2.8 should have a substantially bigger front element than a 18-50 f2.8

IS is cool but it's no magic bullet. sometimes, i get soft results at slower shutter speeds with 17-55. my 'in-the-field' fix is to motor through several frames hoping to get one that's sharp. it works. with the new 50d you should also be able to blast the ISO and still get usable results.

Interesting point, Alan. I didn't realise that 1mm made such a difference to element size at that focal length. Seeing as I have the 10-22mm, I wish they'd made it an 18-55mm!

I suppose there's an argument for saying the 24-105mm would have fitted into my kit better, but the 28-88mm range (in 35mm terms) is so useful when you only want to carry one lens. Plus I like the F/2.8......

I've just done a quick test using my 28mm F/1.8 and the 17-55mm @ 28mm. At F/2.8 the prime has a slight edge in the centre, but the zoom destroys it in the corners. By F/8 there is no visible difference in the centre but the zoom is still much sharper in the corners.

Looks like my 50mm and 17-55mm are fine and the 28mm is not so good - considering it's a 'full-frame' lens it's gets pretty crappy pretty quickly. Still, corner sharpness is not why I have the lens!

We're due some nice weather here (Lancashire) tomorrow, so hopefully I won't be taking any more pictures of newspapers taped to walls...

Cheers

yeah, the 28 1.8 is not a great lens. it's a holdover from the 80s or 90s, based, i'm sure, on a design from the 50s or 60s. canon has a lot of work to do on its short primes. they've rolled out the new 24 1.4. hopefully that marks the start of a trend. current short primes will not do on the bodies which will be coming out in next couple years (or, really, on the ones we have now). they simply don't deliver.

right now, imo, the best canon (short) prime is the ef-s 60. stunning. the rest are a pretty sad lot. i haven't seen results from the 24 1.4 yet, though. hoping for good stuff.

I agree. Problem is that Canon has shown very little interest (apart from the EF-S 60) in making new short primes that mere mortals can afford. The 28mm is great in everyway apart from the actual glass - it's small, fast and has USM. A new one (EF-S if necessary) that was razor shap would be cool, plus, say, an EF-S 12mm F/2.8 and an EF-S 15mm F/2.

I doubt it'll happen though.

yes. i've been wanting some ef-s primes, like a 10, 15, and 30. if they made two versions of each -- a super-fast 1.4 or 1.8, and a normal 2.8 -- that would be fantastic. or ef lenses in the appropriate ranges with two versions each.

agree. don't really need ultra-fast wide angles usually. the 24 1.4 is nice, but really, how often do you need that type of performance.

they have to do something eventually. with nikon and now sigma, red, and others breathing down their necks. whatever happens, we win. whoever makes the equipment, we'll have sizzling bodies and very clean primes.

I own and use both lenses, and my results are largely similar to Puppy Face's. I tested both lenses at close focusing distances on sheets of text, while making sure they were parralel to prevent focusing errors. I then tested in more realistic outdoors settings. The 17-55 was better than the 50mm prime till about F8, at which point it was impossible to tell the diffrence. In the corners the prime was sharer till I stopped down the 17-55 to F4, at which point it was equal with the prime.

The 17-55 had better contrast and saturation than the prime, actually it wiped the floor with it! My 50mm F1.2 Nikkor (EOS adapted) was softer than both lenses at all apertures. Unless I really need F1.8, I generally reach for the zoom instead of the prime. I swapped the 50mm F1.8 a few times since I figured I had a lemon, no luck, results are largely similar. My current copy is the sharpest one I ever got, but it still trails the 17-55 in almost all areas except absolute speed.

These threads always fascinate me. I haven't printed and framed a picture of a newspaper taped to a wall, so I'm curious what practical application this has. In the old days - not too far in the distant past - folks would lay their slides out on the light table, study them through a loupe, and then reach their conclusion on sharpness, color saturation, etc. Now that had practical value - I just question whether this exercise does.

Kind of similiar for me also when i test my lens. My 50 1.8 was better than my 24-70 2.8 on sharpness, contrast, and color.

Greg, I don't think he printed and framed a picture of a newspaper. He taped an actual newspaper to the wall to use as a test target.

-- "but I wondered if the 17-55mm (which is VERY highly regarded) might be expected to have an overall advantage by F8? "

Why? I believe this is the point where your expectaion was wrong. There is absolutely nothing in the (relatively simple) construction of the 50/1.8 that would make it bad at f/8.

Greg is saying what is the point of testing a lens to the nth degree. People do not usually framed and enlarged their lens test result and showing them to their friend about their superior lens(es)

My take on it is If people want to do lens testing and having fun doing it, why not. If you share their result here, it's even better. Different people has different way of enjoying photography. For me personally, my skill is nowhere near my lens quality even the sub-par one, so I don't worry too much about lens testing result for now.

The Zoom is sharper with a better bokeh.

It 'fascinates' me the way some people react anyone talking about the technical performance of characteristics of photography equipment . Of course I didn't buy the lens in order to take pictures of newspapers taped to walls, but when I've been a little disappointed with some of the results of something I've paid a lot of money for (and it's pouring down outside), I see no problem with conducting a test to see if my poor technique is solely to blame. Thanks to the test and the helpful comments on this thread, I now know that my lens is fine and I have only myself to blame for any unsharp results. That can only help me improve.

It may be that people used to judge sharpness by looking at slides through a loop, but with digital we can quickly do a test and look at the results on a monitor. Is that such a bad thing? I realise that pixel-peeping and measurbating are all some people seem to be interested in, but that doesn't mean that any technical assessment of a lens or camera is 'wrong'. It's not 'all or nothing'.

my 35/2 and 50/1.8 are both sharper than my 24-70L (though the 24-70L was more lovely in other ways). The corners from my cheap little 135SF were sharper than my 70-200/4L @f4 and 135mm and even with a Kenko Pro 1.4, the SF was sharper than the zoom in the corners at 190mm (the 70-200/4L was also lovely though)...

bottomline is that I like small light weight primes so I've sold the zooms.

About the fact that manual is better than autofocus : you haven't said which camera you are using. My 20D has horrible autofocus performance... it might be the back focus issues that everyone talks about.

Apparently the newer Canon crop-sensor cameras have fixed this with profiles for individual lenses. I'll be grateful if users of this feature can comment.

ps - ulp! I reread the OP and saw it was a 40d.

terry

i'll never understand why people take pictures of real locations when they could just as easily test a lens at home using a newspaper taped to a wall.

seriously, though...looks good. is that the 17-55?

It's not actually, it's the 10-22. But, I did use the 17-55 yesterday and it performed well.

Cheers

I did a similar test.

http://www.fredmiranda.com/forum/topic/684237

Happy shooting, Yakim.

I don't see whats so insane about testing lenses that you have spent your hard earned money on.

Why was the picture I posted removed from the thread?