[lol101]

I see a couple reasons (though this is speculation of course)

• constant f/4 lens has no reason to get a * except for Tele lens like 60-250 (IMO)
• Mechanical construction is not comparable to a DA*
• Weather sealing may not be as good as DA*
• Performance probably isn't as good ad DA* either
• DA* may imply dual AF

I have read the whole test and have some small reservations on a few comments by Chasseur d'Images on this lens:

1/ They state that there is no Quick Shift focus (thus no way to get directly to MF when AF is engaged) and I would be VERY surprised if that was true. I can't see Pentax downgrading the lens that way by not including a feature that has been present on almost every lens in the last 5 years, SDM or not.

2/ I am also not so sure about the weathersealing with this kind of barel design but I hope they are right on this.

And on a side note:

3/ Have you noticed their insistance at mentioning that all the new Pentax lenses are designed and manufactured by Tokina? Are these informations reliable?

Other than that, it's obvious that Pentax is in the lead when it comes to primes and presently the only brand promoting new, high quality and innovative primes for DSLRs. Good job Pentax!

[falconeye]

I still don't have my copy of Chasseur d'Images.

However, I've found a scan of 2 pages of those tests elsewhere on the Internet and after some consideration I decided to post the piece about the new 17-70mm here. I think that this is still considered a quote not infringing their copyright.

I think that Chasseur d'Images provides the absolute best information about this lens to date and therefore must be shared with the rest of the world.

Note:
The tiny unreadable font says: 17/25/36/70mm and f/f/5.6/8/11/16/22

[ogl]

What does it mean - bottom's graphs - if it's resolution - I don't understand - 8, 9, 10...
And it's very low results at 70 mm...No resolution at all...))

I understand LW/PH like photozone tests. And where did they get DA17-70 if it would be planned to launch on August, 2008.

[Talisker]

I don't know what they are measuring either, but the error bars on the first three measurements at 70mm suggest that that particuar test is worthless...

[falconeye]

Chasseur d'Images does extremely elaborate lens tests. I know that the error bars result from differences between corners/borders. So, large error bars indicate that a lens wasn't centered well in the fab. The opposite of worthless a test.

The scale 0 ... 10 is called "Piqué". Chasseur d'Images readers have to tell us how it is defined. All I can say that it is jointly defined and measured by French Ci-lab and DxO-Analyzer.


So, Chasseur d'Images obviously got an early copy of the DA 17-70mm from Pentax. I can only applaud them for this achievement.

BTW, the page is online now at http://www.chassimages.com/numeros/numero.xml on "aller page" 22.

[Geekybiker]

Those graphs look alot like MTF graphs from other sites. Not sure what the units are, but I suspect its very similar. Maybe 10 being the diffusion limit?

[Geekybiker]

"pique: the level of pique is overall exceptional until 50mm approximately then it degraded and needed to stop down two stops to find a good level on all the field. All in all, for a semi-economical lens, the quality is homogeneous and of good level but the minis-photosites of the sensor 14.6 mpix do not let pass any defect! Quality total and potential are much higher than those of a standard econo-lens 18-55mm. The motorization is integrated, same in its version simplified, brings silence and quick reaction. Protection antiruissellement n' is not oubliee. Very a " good choice; expert"."

My attempt with the help of babel fish to translate the pique part.

[lol101]

I don't know what they are measuring either, but the error bars on the first three measurements at 70mm suggest that that particuar test is worthless...

These bars denote something to the effect of IQ on 30x40cm print on a scale of 1 to 10. 10 is of course considered outstanding.

This is not really an MTF graph as it take into account the sensor, the AF precision and all other camera characteristics.

These graphs are cross comparable between cameras and the best lenses on a !DsMkIII would probably get a 9 mark at the best apertures.

The "error bars" show the dispersion across the frame, the smaller the bars, the better the homogeneity.

The 17-70 shows excellent results until 50mm and a noticeable fall in quality by 70mm. Nevertheless, it also shows considerable CA, noticeable distortion and vignetting.

A good zoom but the recent Pentax primes (200mm, 300mm and 35f2.8 Macro) are showing almost perfect notes on all accounts!

[bdery]

Geekybiker, this translation isn't bad at all for an automated system... The last two sentences should read something like "weatherproof protection isn't forgotten. A very good "expert" choice."

I hope it helps. I could go over the whole of it, but I think Babelfish did it sufficiently well.

[Geekybiker]

Geekybiker, this translation isn't bad at all for an automated system... The last two sentences should read something like "weatherproof protection isn't forgotten. A very good "expert" choice."

I hope it helps. I could go over the whole of it, but I think Babelfish did it sufficiently well.

Well, I looked at and fixed some syntax, etc. Best guess at mistranslated words based on context, etc.

[falconeye]

These bars denote something to the effect of IQ on 30x40cm print on a scale of 1 to 10.

@lol101: Do you know how the scale actually is defined? I only found one copy of Chasseur d'Images in my bookshelf (#284 from June 2006) and they already used this scale -- w/o explaining it. It is somehow related to DxO Analyzer software, though. I am puzzled too that P&S cameras got "Piqué"-ratings of up to 9. Is this really comparable then?

[ogl]

I don't understand web-sites, magazines or photo resourses which review not yet manufactured lenses. Ned Bunnell - Pentax USA posted several photos from 17-70, but said that it's not final version of lens and deleted it.
If I trust this review - 17-70 is really not very good lens. 70 mm is useless. It's 17-70/4-11.
New lens which can't work good at 70 mm with new camera K20D?

[theedudenator]

I don't get why pentax would have two lenses that are so similiar?
16-50mm/2.8
17-70mm/4.0

Unless there is a big price difference??

Maybe they should work on something in the tele end instead....

[falconeye]

If I trust this review - 17-70 is really not very good lens. 70 mm is useless.

ogl, you really shouldn't draw conclusions from test results which you haven't understood yet (like everybody else here, including myself). We haven't yet found out what a corner Piqué rating of even 0 fully open would actually mean. All we know for sure is that the people who do understand the test results gave it a four star rating. And this means very good; not excellent, but very good.

[ftpaddict]

I don't get why pentax would have two lenses that are so similiar?
16-50mm/2.8
17-70mm/4.0

Unless there is a big price difference??

Maybe they should work on something in the tele end instead....

Can you tell the difference between constant f/4 and constant f/2.8 along the focal range? Most can, and they know it is a VERY big deal.