New Olympus 17mm 1.8

Started Mar 7, 2014 | Discussions thread
Anders W Forum Pro • Posts: 21,468
Re: New Olympus 17mm 1.8

Jolly Oly wrote:

Anders W wrote:

Jolly Oly wrote:

Anders W wrote:

Noone is saying that the lens can't take decent images but user reports like this doesn't tell us how it compares to the alternatives.

and when they do, you choose to ignore them.

What user report showing good and directly comparable evidence from the two lenses did I ignore?

Like in this case here with Pixnat2 and many other users, mostly enthusiast photographers who know what they're doing, who had both 20/1.7 and 17/1.8

Could you please link me to the evidence from the systematic tests performed by Pixnat2.

What I actually got is a one of the sharpest lenses in my collection - ever. I never had an eqiv. 35mm prime on digitalSLR but this lens is by far sharper than all of my previous zooms - 12-60mm and 14-54mm Zuikos included - already at f/1.8.. The rest of the alleged lens faults were in fact equally minor.

As you might be aware, a number of fairly well reputed test sites have shown it to be less sharp than the 12-35/2.8 as well as the 12-40/2.8 at the same FL (or thereabout).

and I'm sure SHG 14-35mm/f2 Zuiko is sharper from all of them, so what ?

It's normally expected that a prime does at least as well as a zoom, even a good zoom.

You mean at least like 20/1.7 is doing against 12-35mm and 12-40mm zooms ?

Judging by the same (your) standards the 20/1.7 sucks in that regard. Or not ?

No it doesn't suck by my standards in that regard. Try to look for comparable data next time. The 20/1.7 was tested on an E-P1 (12 MP, strong AA filter) whereas the two zooms were tested on the GX1 (16 MP, fairly weak AA filter).

look at the size and price difference. m.zuiko 17/1.8 is like a glass compared to a bottle against 12-40mm, it was $400 when I bought it (compared to a $1000) and is f1.8. My sample is tack sharp already at 1.8.

What we are talking about is how it compares to the 20. Got any evidence you'd like to show me in that regard?

Again you choose to ignore that it was you who pulled a comparison with micro zooms. But never mind.

It wasn't. You brought in a comparison with FT zooms. There are no test data whereby we can compare how they do relative to the 17/1.8. So I turned to MFT zooms instead, where it is possible to find such data.

By your standards this is very unscientific and shallow I guess.

I am glad you point to Ming Thein's review. The directly comparable sample images he provides gives us very clear evidence of the optical superiority of the 20/1.7. So I encourage everyone to have a look at them. However, his comments and conclusions are out of line in a number of important respects. Here's the real story:

Ming says: "The 20/1.7 has the highest overall scene contrast, but the 17/1.8 wins out in microcontrast and reproduction of fine detail structures – personally, I prefer this as it gives me more latitude for processing before the shadows and highlights block up."

My comment: It's clearly evident from Ming's samples that the 20 does better with respect to global contrast and microcontrast (the contrast between smaller details) alike. When Ming says that he has "more latitude for processing [with the 17] before the shadows and highlights block up" he means that he can apply more sharpening. But the point here is that the 20 simply needs less sharpening by virtue of having more microcontrast in the first place. And not having to sharpen (much) is preferable on all counts. It doesn't increase the noise level and yields a cleaner, more natural look.

Ming says: "What will affect resolution (and perceived acuity) far more is lateral chromatic aberration."

My comment: It will not, at the magnitude shown here, affect resolution or acuity to any measureable degree once corrected. And this particular aberration can be corrected easily/automatically and without any downsides in PP. It is also corrected for the 20/1.7 by all Pany bodies and for both lenses by the E-M1 (and probably all Oly bodys following the E-M1).

Ming says (still talking about lateral CA): "We’re now seeing CA from all three lenses, with the 17/2.8 once again faring the worst. The 17/1.8 is slightly better than the 20/1.7. Interestingly, not much changes even when you stop down."

My comment: Why is it interesting (surprising) that not much changes when stopping down? It is well known that lateral CA (unlike longitudinal CA) never changes as you stop down.

Ming says: "The 17/1.8 renders out-of-focus areas with a rounded softness and lack of hard/ bright edges or double images, even against complex background textures."

My comment: A close inspection of the blur discs (in the problematic transition zone) shows them to have properties very similar to that of the 12/2, which (for good reason) came in last in the test/poll of bokeh among MFT WAs reported here. With the 12/2, the poor bokeh doesn't matter much, at least for my usage. With the 17/1.8, it is of somewhat greater concern since I am more likely to end up with at least parts of the frame outside the DoF whether I like it or not (although I'd rarely go for subject isolation by means of background blur with what is effectively a 35/3.6). The 20/1.7 is better in this department (although it is certainly no bokeh king).

Ming says: "I would not buy the 17/2.8 unless size is a critical priority, or you know that you’re going to be shooting only static objects stopped down; otherwise the slow AF speed will drive you crazy. The Panasonic 20/1.7 is in a similar boat; it’s faster to focus than the 17/2.8 and optically better, but nowhere near as fast as the 17/1.8."

My comment: See here.

In spite of the above, I appreciate Ming's review very much because of the evidence he puts at my disposal, which I can judge for myself, without relying on his comments.

So what about lenstip credibility (in my eyes went down completely)

Lenstip is just one of several sources showing the same evidence.

based on, obviously, not a representative copy of the lens. That's ok, but it would be fair from such a sites to do little research and to at least mention a possibility of sample variations.

What I pointed out is that Lenstip is not at all alone. The story is pretty much the same regardless of where you look.

You mean whereever you look. Plenty of proof of the opposite, in my case - my personal images included.

I mean wherever we look. Your personal images are not included since you haven't offered any directly comparable test images.

All of that (sample variations) is based on assumption that "good" lens sample is not accidentally good but designed that way and "bad" lens sample is accidentally bad - not designed to underperform. So if we have two different copies of the same lens I believe that good copy is the reference, not a bad copy.

And there are, of course, other sources with different conclusions (than lenstip).

What sources reach a significantly different conclusion than Lenstip when it comes to how the lens compares to the 20 in terms of sharpness?

or what did I learned from that ? Something more about sample variations I guess (mine is black like many 17/1.8 with stellar reports) and one more thing: for sure I will never try to unconditionally support someone's view of something I never had a chance to use (like you're doing now).

How do we know that the alternatives you tested it against aren't worse than copies tested by others? Unlike you, I don't fall into the trap of thinking that my own copies, test procedures, and perceptions are all infallible and those of anyone else just crap.

How do you know that radioactivity and its effects are for real? Ever tested that yourself? Do you still believe it? If so why?

good point from your side and an easy answer for me: I trust myself.

So you tested that radioactivity is for real on yourself?

With every lens I got I run a few test shots in various conditions, like a starting point just to see what to expect. Then the real life usage will further show and fine tune my opinion. Using the database of my lens collection (in my brain and lightroom archive) with sufficiently large sample size (20-30 lenses) it's easy to rank the specific lens performance. After all, it's not a rocket science.

That sounds like a very unreliable and unsystematic way of going about it. No wonder you end up with strange conclusions.

Sounds like ? How could you systematically measure sound out of pixels ? Or what is your reliable method to systematically define a beautiful image ? No wonder you end up with strange conclusions.

It sounds like you are somewhat unfamiliar with the meaning of the idiomatic phrase "sounds like".

So Acrill is right. I mean you have never used it, right?

No I haven't. But for reasons already spelled out, that does not prevent me from passing verdict on it. There's plenty of relevant evidence out there.

Your "problem" is that you like to ignore the relevant evidence which doesn't fit with your story.

What directly comparable evidence did I ignore?

As I'm aware the majority of users who had both Panasonic 20/1.7 and Olympus 17/1.8 kept the Olympus. If for you this is not relevant evidence that something is wrong with Lenstip alike reviews, so be it.

Here are three reasons why arguments of that kind don't work. First, neither you nor I know how many of those who had both did what.

In other words real user perspective, even in significant numbers, is completely useless if it goes against big review site and/or your opinion.

With regard to facts yes. The majority is sometimes completely wrong about them. That's why I gave you the flat-earth example. The mistake you are making is more generally known as "appeal to the people" (or "argumentum ad populum" in latin). See the link below for further information:

With regard to preferences, no. I have no trouble with you or anyone else preferring the 17 to the 20. That's entirely up to you. All we can meaningfully discuss are the facts.

Second, those who had the 20 and ordered the 17 were more attracted to the 17 in the first place than those who had the 20 and didn't order the 17.

Judging by your standards, again ? Just for the sake of the argument I could easily say everybody were attracted to the 17 but only some had extra money/needs/curiosity to play with both and bite the bullet. Except you, of course.

I know, based on discussions here, that I am not alone. But again, it doesn't matter for reasons already mentioned above.

Third, why do you think most people once thought that the earth is flat?

Most but not Galileo. He choose to look at the Moon and the horizon and trust his eyes against all the major review sites of that time who claimed the opposite.

You are doubly wrong here.

First, Galileo was certainly not the first to claim that the earth was round. This was well known among scholars since the ancient Greeks, a couple of thousand years earlier. Galileo was among the first (but not the very first) to argue in favor of the heliocentric view, which is something else.

Second, Galileo certainly didn't ignore the evidence provided by "the major review sites" (of the day and in the relevant area). Rather, he looked critically at the evidence provided by them and drew the correct conclusion. Those who confronted him were those who refused to take the objective evidence seriously and just "followed the crowd".

 Anders W's gear list:Anders W's gear list
Panasonic Lumix DMC-G1 Olympus OM-D E-M5 Olympus E-M1 Panasonic Lumix G Vario 14-45mm F3.5-5.6 ASPH OIS Panasonic Lumix G Vario 7-14mm F4 ASPH +28 more
Post (hide subjects) Posted by
(unknown member)
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow