[civiletti]

Zoom lenses continue to be improved. At the ultra-wide end, there are zooms [DA12-24mm, Sigma 8-16mm, and Nikkor 14-24mm] that outperform available primes. I was impressed by the MTF results in Photozone's test of the relatively new Sigma AF 17-50mm f2.8 on a Canon EOS 50D, and wonder if zooms are now matching the performance of primes at longer focal lengths. By comparing pairs of lenses tested on the same camera with one of the same lenses tested on another camera and repeating this process, it appears that this Sigma zoom betters the MTF of any Pentax zoom and equals the MTF of Pentax primes and available third party primes, including the Zeiss 35mm f2 ZF.

I did not do comparisons of primes of 40-50mm, where I suspect some primes, especially macro lenses, would outperform the zoom lens.

So, I wonder if users of the Sigma 17-50mm f2.8 have found its performance from 17-35mm matches the best prime lenses they have used. That is in resolution and contrast, not speed, bokeh, distortion, or abberrations.

[Zafar Iqbal]

No idea about your question but - the charts can't be compared across different systems. It took me a while before I noticed this myself. Before then I was like wtf!?

The MFT charts have this bar along with them - notice the difference:

[civiletti]

No idea about your question but - the charts can't be compared across different systems. It took me a while before I noticed this myself. Before then I was like wtf!?

The MFT charts have this bar along with them - notice the difference:

Direct comparisons are not valid, but comparisons can be standardized by using a lens tested on both cameras involved. For instance, the DA 17-70 and Sigma 17-70 are both tested on the Pentax k-10. The Sigma 17-70 is also tested on the Nikon D200. I can use this to compare the DA 17-70 to the Nikkor 18-70 or either of the Nikkor 35mm primes tested on the D200. The math is simple proportion. This is what I did to compare the Sigma 17-50 f2.8 to Pentax zooms, wide primes, and the Zeiss 35 f2. I should note that I looked at performance at optimum aperture, since I am concerned with landscape images shot on a tripod. Folks who want good perfomance at wide apertures would likely find prime lenses preferable.

[Zafar Iqbal]

Thats good. I was just trying to make sure you didn't make the same mistake as I did. I hope someone can step up and say zooms suck and primes rule because I like using primes the data on the Sigma does look very intriguing however.

[asdf]

By comparing pairs of lenses tested on the same camera with one of the same lenses tested on another camera and repeating this process, it appears that this Sigma zoom betters the MTF of any Pentax zoom and equals the MTF of Pentax primes and available third party primes, including the Zeiss 35mm f2 ZF.

There is a reason why they include corner and border performance. Take a look at MTF charts released by manufacturers. Also take a look at slrgear.com. Having high performance in a tiny part of the centre of the image isn't useful is it?

[pcarfan]

Did you check the MTF figures for the Tamron 17-50/2.8 SP XR Di II LD Aspherical [IF] (version I without VC and the only one available in Pentax mount), which has better MTF than the Sigma especially at f2.8?

Also, I don't think photozone MTF accounts for contrast, but I could be wrong?

Also, just as a side note, I have long concluded MTF figures doesn't mean much when it comes to evaluating lens perfromance, in fact the entire lens "test" doesn't tell it all and at times gets it wrong. These tests to me are just guidelines.

[asdf]

Did you check the MTF figures for the Tamron 17-50/2.8 SP XR Di II LD Aspherical [IF] (version I without VC and the only one available in Pentax mount), which has better MTF than the Sigma especially at f2.8?

Also, I don't think photozone MTF accounts for contrast, but I could be wrong?

Also, just as a side note, I have long concluded MTF figures doesn't mean much when it comes to evaluating lens perfromance, in fact the entire lens "test" doesn't tell it all and at times gets it wrong. These tests to me are just guidelines.

Manufacturers have always released charts with MTF on the y-axis and image position on the x-axis. Curves would correspond to a fixed LW/mm frequency value (and either meridional or sagittal lines). A smaller frequency would correspond to macro-contrast, whereas a larger frequency would correspond to micro-contrast. The height of the curve would indicate how much of either one you should expect at the given point on the image.

The stuff that's on the web is pretty lazily done, IMHO.

[deejjjaaaa]

Manufacturers have always released charts with MTF

except Pentax

[normhead]

Just to jump in here, my DA*60-250 at 90mm has MTF numbers very similar to the Tamron 90 macro, and compared with 2 other lenses, these lenses standout, but the Tamron gets the nod based on, the image just looks a little sharper, maybe colour cast, maybe contrast, you can only guess why you like an image more a lot of the time. So, I'm not going to agree that MTF numbers mean nothing, the 4 lenses I looked at performed exactly as their photozone MTF numbers said they would.. only when the numbers were very close did intangibles come in to play. That being said, those intangibles seem to favour prime lenses. You can disagree if you want, however, my primes out-perform my zooms, and that DA*60-250 is the most expensive lens in production at the moment.. I used to argue the opposite side of this. The Tamron 90 and 21 ltd changed my mind.

[Edgar_in_Indy]

There seems to be very few owners of the new Sigma 17-50mm in Pentax-mount so far. There are only four customer reviews on B&H's site, and none whatsoever on Adorama or Abe's. I imagine that the higher price over the Tamron 17-50mm is keeping most people away. And the lack of response on this thread seems to support that notion. But from what I've heard, it's a superior lens and the price is justified.

I hope to be able to add it to my collection alongside the Tamron 28-75mm f/2.8 and the Sigma 50-150mm f/2.8. With those two lenses, I've never felt the need to buy a prime for better image quality.

[DanielT74]

From my limited experience zooms can be very sharp, hell, cheap Korean lenses can be very sharp (like my Samyang 14/2.8) but there is a lot more to image quality than sharpness.

So I certainly wouldn't rush to write off primes as yet.

DA12-24 which you mention, while being sharp has pretty significant CA and distortion. I used Tamron 17-50/2.8 and at least my copy had a pretty shocking AF and consistently underexposed. So there is a lot more to it.

Even with sharpness, the MTF graphs don't tell the whole story. Nice contrast can make an image look sharper, for instance. Most tests don't show my old K 135/2.5 to be a particularly sharp lens but the images from it certainly don't look soft (check the Guess the lens thread at the end).

That said, zooms can be very handy and produce great images for sure.

[DaveHolmes]

I used Tamron 17-50/2.8 and at least my copy had a pretty shocking AF and consistently underexposed

I've been using the same lens for a while but my copy definately does not exhibit these problems... AF is quick (but bloody noisy) and exposes perfectly (unless I screw it up!)

[civiletti]

Even with sharpness, the MTF graphs don't tell the whole story. Nice contrast can make an image look sharper, for instance. Most tests don't show my old K 135/2.5 to be a particularly sharp lens but the images from it certainly don't look soft (check the Guess the lens thread at the end).

This is true, but not as true as in the past. Post processing can minimize chromatic abberration and linear distortion. Contrast can be increased at the macro and micro level. Color balance can be tweeked and saturation adjusted. Post processing cannot increase resolution, though. Sharpness can give the impression of more resolution, but that's not the same. It's somewhat analogous to the basketball quip, "you can't teach seven foot."

[civiletti]

Manufacturers have always released charts with MTF on the y-axis and image position on the x-axis. Curves would correspond to a fixed LW/mm frequency value (and either meridional or sagittal lines). A smaller frequency would correspond to macro-contrast, whereas a larger frequency would correspond to micro-contrast. The height of the curve would indicate how much of either one you should expect at the given point on the image.

The stuff that's on the web is pretty lazily done, IMHO.

Given the capabilities of digital post processing, I wonder if good old resolution testing would be more useful than MTF charts. Overall and micro-contrast can be increased effectively. Resolution cannot.

[civiletti]

Did you check the MTF figures for the Tamron 17-50/2.8 SP XR Di II LD Aspherical [IF] (version I without VC and the only one available in Pentax mount), which has better MTF than the Sigma especially at f2.8?

Good suggestion. The Tamron 17-50 f2.8 SP XR Di II LD Aspherical [IF] betters the Sigma 17-50 f2.8 at optimum aperture everywhere but the center at 50mm. [My comparison is based on PZ tests of the two zooms tested on Canon APS-C cameras, adjusted by a factor of 1.23 when comparing 8mp results to 15mp results.] I wonder how much the curved plane of focus of the Tamron would detract from results at the wide end. It would take some extended testing to settle that, I fear.