Zeiss 24mm f/2 review now posted

[Jasper_D]

Thank you David, for that review.
Have the lens for some time already, but it´s nice to read a pro´s take on it once in a while.

[agorabasta]

The more I peer into David's samples from that 24mm Zeiss, the less I actually like them.

And that comparison against the 16mm SEL on APS is not really telling how good the 16mm pancake is. It only tells how mediocre, at best, the FF Zeiss is.

So I would like to add another aspect to the comparison. The best lens I have to do 16mm on APS is the Samyang 14mm, there's only some x1.14 crop And its f-number then also scales by the same x1.14 factor. So then to get at comparable FF aperture for the effective crop turns the f/8 into f/14. And there is a sample in David's article at exactly f/14.
So I have a good sample shot with the 14mm Samyang with a Nex5. Now to arrive at the same pixel density for the crop, I have to upscale the 14.2Mp by 150% linear to 32Mp with further crop by 1.14 linear to 24.6Mp. (Linear values must be squared to arrive at area values)

And below is that sample full reduced followed with a 100% crop from a 150% upscaled image having the same exact pixel density as 24.6Mp APS sensor - the most fair comparison possible -

DSC00803-2.jpg

(225.05 KiB) Downloaded 2756 times

DSC00803.jpg

(237.56 KiB) Downloaded 2756 times

Frankly, the Zeiss on real 24Mp FF camera doesn't come close to the quality on an upscaled and cropped 14Mp APS.

So if you need a truly good 24mm equivalent, you are much better off buying a Samyang together with an a65. That combo is considerably cheaper than the 24mm Zeiss alone. And you also get far superior image quality. And more importantly, that quality is absolutely the same across all the APS frame - the Samyang is at least that good.

[agorabasta]

Here's one more 150% upscaled crop, this time quite sharpened.

[01af]

David, are you aware of photoscala's review of fast 24 mm lenses they did in May 2011 (in German language)? The test included the following lenses:

• Canon EF 24 mm 1:1.4 L USM II
• Nikon AF-S Nikkor 24 mm 1:1.4 G ED
• Sigma AF 24 mm 1:1.8 EX DG
• Sony SAL Distagon T* 24 mm 1:2 SSM ZA
• Leica Summilux-M 24 mm 1:1.4 Asph

The Sony ZA lens came out on a disappointing last place. Acceptable at medium aperture but poor at full aperture. The reviewer's resumé was this:

Meiner Meinung nach liefert Sonys 24-mm-Objektiv abgeblendet ganz gute Ergebnisse, lässt aber bei größeren Blendenöffnungen doch einiges zu wünschen übrig. Dem großen Namen Zeiss wird es nicht gerecht. Das könnte, so legen unsere Tests und Recherchen nahe, nicht nur ein einzelner Ausreißer sein, sondern an den vielleicht aus Kostengründen etwas großzügigeren Produktionstoleranzen und der nicht so akribischen Endjustage liegen. Konnten doch auch die zeitgleich getesteten Sony-Objektive 1,4/35 mm G (SAL35F14G) und 1,4/50 mm (SAL50F14) nicht völlig überzeugen. Canon und Nikon beispielsweise machen es jeweils sichtlich besser.

Translation:
In my opinion, Sony's 24 mm lens delivers acceptable results when stopped down but leaves a lot to be desired at the wider apertures. It does not live up to the great Zeiss name (emphasis added by the translator). According to our tests and research, this obviously is not a single runaway case but due to generous tolerances, to keep costs down. After all, the Sony lenses SAL 35 mm 1:1.4 G and SAL 50 mm 1:1.4 also weren't able to fully satisfy our testers. Both Canon and Nikon, for example, clearly are doing better.

So according to photoscale.com, the Sony 24 mm ZA lens is, umm, not really bad—but not really good either. David's test, on the other hand, seems to indicate this lens was just excellent. So this means that either David isn't able to do a meaningful lens test, or photoscala.com isn't able to do a meaningful lens test, or there's an awful lot of sample variation in the ZA lenses.

At the end of the day, I am afraid all this confirms the old adage: Do not trust a lens review unless you did it yourself—and even if you did, triple-check that you didn't fool yourself.

[David Kilpatrick]

Completely baffled by agorabasta - the image shown doesn't look to have any particularly fine detail at all. No stone texture or anything. It certainly doesn't look better than the Sony 24mm.

The Sony 24mm may be not as good as the MF Zeiss 25mm f/2, and it may be either not as good as or only equal to Nikon and Canon - but I've only used f/1.4 designs from them, there is no f/2, and there's nothing comparable in price.

And the entire APS-C issue is irrelevant for judging the lens itself as far as I'm concerned - that is a secondary use for the lens, mostly for video. It's interesting to check out, but actually quite difficult to test as there is only one Sony APS-C body which can focus the 24mm at full aperture (PD) well enough to evaluate the lens - the A77, with micro AF adjustment.

I'll stand by what I have said - that no general market (consumer) full frame 24mm f/2 AF or closely comparable lens exists at this price point and with its qualities. I have my criticisms, I'm aware of conditions where flare can occur, where 1 pixel CA appears, the degree to which corners can be slightly soft; but I'm also aware of the value of a lens with this low level of simple distortion over the 84° angle.

There are just many issues involved. Ken Rockwell says the Nikon is 'the world's best 24mm f/1.4 lens' - but the Nikon only focuses to 25cm, as does the Canon mkII 24mm f/1.4, not to 18cm like the Sony - and if you use wide lenses like I do, that is almost a deal-breaker.

As I have said in my review, it would difficult for the Sony to focus any closer - the lens shade is only 1 or 2 cm from subjects.

Something else I have been studying recently has been how the minimum focus of so many lenses has been affected by the change to SSM/HSM (Sigma, especially). So it's great to find an SSM lens which does not have its range restricted - the same applies to the 24mm f/1.8 for NEX, focusing down to 16cm (though there's actually a bit more clear space between lens and subject).

One weakness of Sony has been the inaccuracy of the AF. A strength of Canon and Nikon (at least in pro bodies) is the accuracy of AF. Many tests - dPreview included - have, I think, been blighted by poor focus accuracy and no real attempt to get the bottom of the problem.

David

[agorabasta]

David, you can't be serious!

It's exactly your 24mm f/14 sample that shows no stone texture. And then some very bad haze halos at strong bright-to-dark transitions. So no sharp strong edges at all.

Mind also, that the church building in my sample is brand new - less than one year old.

Here's another crop from my sample, this time without luma NR -

[David Kilpatrick]

Well, I suggest anyone interested can read the review and see this image:

http://www.pbase.com/davidkilpatrick/im ... iginal.jpg

Maybe others can help point out to me exactly where you see the damning faults you point out, and why your example just posted is superior. My image contains a few other useful reference points, such as the detailing of the leadwork on the cupola, the hair on the people, the textures of clothing.

David

[pakodominguez]

as we say in Spanish, le estan buscando tres pies al gato...

[agorabasta]

Well, Pako, that Spanish saying is used to discount the other's opinions as unimportant and/or meaningless. The problem is that, whenever someone considers something as unimportant, there are always two possible causes for such attitude - one is against the others, another against the sayer...

Now more to the real point. 'Resolution' as it is normally getting 'measured' is not always meaningful. The more realistic assessment is to measure the resolution in some meaningfully recognisable things.
The readily available thing to recognise is a printed text. If you can get you system to resolve a 6x8 pixel letter characters so that the text is generally legible, your system is at least adequate. If you can get 5x6.5 pixel letter text legible, your system is good. But only if you can guess the characters as small as 4x5.5 pixel well enough to actually read the unfamiliar text, you system resolution is truly awesome (for the nominal pixelcount).

BTW, it means that real resolution also depends on the quality of processing...

[pakodominguez]

Well, Pako, that Spanish saying is used to discount the other's opinions as unimportant and/or meaningless. The problem is that, whenever someone considers something as unimportant, there are always two possible causes for such attitude - one is against the others, another against the sayer...

Nope, that spanish saying mean that you (not you, Agorabasta, you in general) are lookin for problems that don't really exist. Like a bizantine discussion that is more valuable as rhetoric than the real meaning of the subject (like the sex of angels...)

I think that David's review is right and his test accurate. And I think Agorabaste, your position and your test is also accurate. But even if the math you use in order to validate your comparation, it doesn't works. Your 14mm is probably good, or better than the Zeiss (...) but you can not compare it with a 24 on FF. It is not the same kind of lens, the constraints are not the same. Then, one lens is manual while the other AF and then, and then... I would love to compare the Zeiss against the Sigma Super Wide 24mm f2.8 Macro, a lens from the 80s that I find surprisingly good, definitely better than the Maxxum 24mm f2.8

But I won't buy this lens, and I'm still wandering about the SEL 24mm f1.8 (I'm really tempted, speciall if I can find a good discount) that anyway I won't try to challenge against any 24mm or 35mm on FF. I will only unload the lenses it will replace.

(I'll show you samples of the sel 55-210 on the other forum, during this weekend)

regards

[artington]

But I won't buy this lens, and I'm still wandering about the SEL 24mm f1.8 (I'm really tempted, speciall if I can find a good discount) that anyway I won't try to challenge against any 24mm or 35mm on FF. I will only unload the lenses it will replace.

Excellent post, Pako. Like you, I am pondering the Zeiss E 24/1.8 but have not decided yet because (1) it is quite big and (2) it is dedicated E-mount. The second reason might sound a little odd but - amazing camera that the NEX-x is - these are still early days for a new system and paucity of lenses may yet sink it and while I dont mind spending £200 or so on a lens which might end up being redundant in time I am much less inclined to put £800 down. Furthermore, in the Zeiss ZM 25/2.8 there is a viable alternative (given the ease of focusing manually on the NEX-7) which is also married to a system which is likely to hold value. Of course, this might be wrong - Leica was on the verge of extinction before it came up with the digital Ms and S2 and technology marches on and favours companies with big balance sheets who can afford the R&D. In this context, the well leaked Fuji X Pro-1 looks to be a real shot across Sony's bows in a war tbat may only just have begun but which, it would not surprisr me, might blow away MFT, particularly now that Olympus is walking wounded at best.

[agorabasta]

Nope, that spanish saying mean that you (not you, Agorabasta, you in general) are lookin for problems that don't really exist. Like a bizantine discussion that is more valuable as rhetoric than the real meaning of the subject (like the sex of angels...)

But that's the same as I said - you simply dismiss the matter. One may dismiss the things considered by the others as important for two basic reasons - having a superior knowledge of the matter, or having too little knowledge to understand.

So let me elaborate. A 24mm on FF and a 14mm on APS are not too far apart. The overlapping effective range in FF terms starts at f/5 24mm. So all depends of your use of the lens. If you shoot your 24mm FF wide open and often need AF for close objects, then the comparison is impossible, being outside the overlapping range.
Then if you use that 24mm for street views, architecture and landscapes, the comparison is very much possible. And in the latter case the AF turns from advantage into nuisance. And exactly in that overlapping range the Samyang 14/2.8 used on APS beats that 24/2 ZA handily in every single respect.
The Samyang used on APS has lower distortion and much lower CA. It has better resolution that is also very even across all of the cropped frame. And if you use a dense 24Mp APS sensor with a slight 114% upscaling to move from 14mm APS to 16mm, you actually get less aliasing and even slightly higher real resolution at 24Mp output.

Now you see - it all depends.

[pakodominguez]

But I won't buy this lens, and I'm still wandering about the SEL 24mm f1.8 (I'm really tempted, speciall if I can find a good discount) that anyway I won't try to challenge against any 24mm or 35mm on FF. I will only unload the lenses it will replace.

Excellent post, Pako. Like you, I am pondering the Zeiss E 24/1.8 but have not decided yet because (1) it is quite big and (2) it is dedicated E-mount. The second reason might sound a little odd but - amazing camera that the NEX-x is - these are still early days for a new system and paucity of lenses may yet sink it and while I dont mind spending £200 or so on a lens which might end up being redundant in time I am much less inclined to put £800 down. Furthermore, in the Zeiss ZM 25/2.8 there is a viable alternative (given the ease of focusing manually on the NEX-7) which is also married to a system which is likely to hold value...

First at all, I don't think the NEX system risk anything. Sony made a dent with it. We will probably see a slow development, lens wise. But I believe it will become for Sony, as important as the Alpha system.

But I thought about this issue (which lens to buy), thinking that a Zeiss on Leica M mount will have e better and easier re-sale price. But the Minolta MD 24mm f2.8 is good, and the Sigma Super Wide 24 f2.8 Macro is better, and both are on the 150 $, while the Zeiss or any other Leica M mount will be 1100$, but I find hard to believe it is 7 times better... So I thought about buying the SAL 24 f2.0 and using it on the NEX with the adapter, but there you are going to have a huge and heavy combo. Then, I don't really need another 24mm on Alpha mount.

At the end, if I want to spend 1000 $ on a 24mm, it will be on the SEL 24 F 1.8 that is not that big, definitely not bigger than the SEL 50f1.8 but light, AF and we expect great quality from it.

Now: do I want to spend 1000$ on a 24 mm?

[artington]

Now: do I want to spend 1000$ on a 24 mm?

For me the question is: Which is the better investment of $1000, the Zeiss E-mount only, APS-C only 24mm or the freely afaptable Zeiss ZM full frame 25mm? And I'm pretty certain the answer is the latter. However, since I already have a Minolta MD 24/2.8 I may do neither.

[pakodominguez]

Nope, that spanish saying mean that you (not you, Agorabasta, you in general) are lookin for problems that don't really exist. Like a bizantine discussion that is more valuable as rhetoric than the real meaning of the subject (like the sex of angels...)

But that's the same as I said - you simply dismiss the matter. One may dismiss the things considered by the others as important for two basic reasons - having a superior knowledge of the matter, or having too little knowledge to understand.

One more time: Nope.
Your position and David position are both right -if we look from your or David's point of view. But in your discussion, both of you are only enforcing your point of view, that's why your are looking for 3 legs of the Cat (that we know have 4): it is not because the 14mm is an excellent lens that the 24mm perform poorly...

So let me elaborate. A 24mm on FF and a 14mm on APS are not too far apart....

This lenses are not similar just because on APS the crop give you a similar angle of view. The 14mm will accentuate the Lilliput point of view of using an extreme (and really good quality) wide angle against a moderate wide angle lens.

BTW, how good the 14mm performs on FF? I was thinking about buying one of this lenses for a job I had, photographing interior design, until now I'm doing OK with the KM 17-35 F2.8-4...

Regards