Question about 9-18 - how wide is it really?

zettapixel

Mu-43 Veteran
Joined
Aug 12, 2010
Messages
470
Location
NY
As many, I was amazed when Olympus announced 9-18 lens. This is what m43 is all about - a tiny wide lens. Doesn't hurt that, apparently, it's pretty sharp. Although I don't really need it right now, my next trip to Europe will certainly trigger the purchase.

But one thing bothers me a bit and I haven't seen any discussions about it. To quote photozone, "The amount of barrel distortion is indeed very extreme at 9mm (4.8%) altough the problem eases towards the long end of the range". I know it's automatically corrected, but this correction involves cropping the center of the image, which, basically, makes it, not 9mm anymore... correct me if I'm wrong... pun intended...

So how wide is it really on the wide end? Or, let me refrase the question: which FL would you need to set Panny 7-14 to achieve the same field of view in the resulting image as you would get with 9-18 set to 9mm?

Thank you in advance
 

pictor

Mu-43 Top Veteran
Joined
Jul 17, 2010
Messages
637
I know these tests, but I don't care about this detail too much. Every super wide angle zoom I know has a more or less strong distortion somewhere. I am no expert in lens design, but it should be really difficult to do otherwise and even if there were such a lens, it would be extremely expensive.

I own this lens and I really love it. I love it not only because of its optical quality, which I am very impressed of, but also because of its versatility, which the much more expensive and excellent Panasonic 7-14mm has not. Your mileage may vary, but I love to take only one lens with me on a photo walk, which I would not do, if I had only 14mm at the long end of the zoom.
 

dwig

Mu-43 Top Veteran
Joined
Jun 26, 2010
Messages
717
Location
Key West FL
...I know it's automatically corrected, but this correction involves cropping the center of the image, which, basically, makes it, not 9mm anymore... correct me if I'm wrong...
You are wrong.

You are basing your assumption on the incorrect idea that that a lens rated at 9mm but with a significant degree of distiontion has a 9mm focal length across the whole field. The distortion is the result of the fact that the focal length varies across the field. In the case of barrel distortion, the focal length decreases toward the edges. The rated focal length applies to the center portion of the image only.

Also, the "crop the center" concept is inaccurate, or at least inaccurately stated. You only crop after the stretching involved in the distortion correction. You'll find that there is no significant cropping along the horizontal and vertical center axis. This is easy to see when you use an application like Photoshop to correct the distortion after slight expanding the "canvas". The correction doesn't involve any stretching along either the horizontal or vertical centerlines. (yes, this is a bad way to correct lens distortion as expanding the canvas before correction will make the programed corrections less accurate. I mention it only as a way of seeing what is actually happening.)

You can't get overly precise when comparing lenses as the actual focal length and the marked focal length can easily differ by 5%, and sometimes as much as 10%, with zooms. Primes are usually marked somewhat more accurately, with a maximum error of roughly 5%. Also the lens' distortions can be rather complex in nature and software correction often leaves some small spatial distortion even when it straightens lines well. In general, though, you will find that any two lenses with differing optical distortions will yield the same FOV when set to the same FL and after both have had their distortion corrected in software, at least within the plus or minus 5-10% accuracy range of their FL markings.
 

Amin Sabet

Administrator
Joined
Apr 10, 2009
Messages
10,879
Location
Boston, MA (USA)
You'll find that there is no significant cropping along the horizontal and vertical center axis.
Dwig, I guess it depends what you mean by "significant". If you look at any of these cameras' output in dcraw or Raw Developer (which don't correct barrel distortion) and compare to the in-camera JPEG or RAW conversions in C1, Lightroom, Silkypix, or Oly Master (all of which do correct barrel distortion), you find that there is a decrement in both horizontal and vertical angle of view. Same holds true for manual correction in Photoshop or using PTLens.

Roman, in my experience all of these wide angle lenses with corrected barrel distortion have a shorter actual focal at the wide end than they are labeled. For example prior to barrel distortion correction, my Panasonic 14-45mm zoom at 14mm gives a wider diagonal angle of view than it should for 14mm, whereas it gives the expected angle of view (for 14mm) after barrel correction. Same holds true for my Lumix 7-14mm, which is a fair bit wider (based on diagonal angle of view) than expected for 7mm before correcting the barrel distortion and just about right after correction. Having tested the Oly mZD 9-18mm at 9mm, I'm 98% sure it gives the expected angle of view for 9mm after barrel distortion correction.
 

zettapixel

Mu-43 Veteran
Joined
Aug 12, 2010
Messages
470
Location
NY
You are wrong.
You are basing your assumption on the incorrect idea that that a lens rated at 9mm but with a significant degree of distiontion has a 9mm focal length across the whole field. The distortion is the result of the fact that the focal length varies across the field. In the case of barrel distortion, the focal length decreases toward the edges. The rated focal length applies to the center portion of the image only.
I see. That's good to know, thank you.

Also, the "crop the center" concept is inaccurate, or at least inaccurately stated. You only crop after the stretching involved in the distortion correction.
Well, obviously... I agree my description is not scientifically correct, that's not the point, you understood me perfectly.

You can't get overly precise when comparing lenses as the actual focal length and the marked focal length can easily differ by 5%, and sometimes as much as 10%, with zooms. Primes are usually marked somewhat more accurately, with a maximum error of roughly 5%.
I'm familiar with the phenomena at the long end of telephoto zooms at close focusing distances, but I didn't know it applies to wide angle and to what degree.
 

zettapixel

Mu-43 Veteran
Joined
Aug 12, 2010
Messages
470
Location
NY
Roman, in my experience all of these wide angle lenses with corrected barrel distortion have a shorter actual focal at the wide end than they are labeled. For example prior to barrel distortion correction, my Panasonic 14-45mm zoom at 14mm gives a wider diagonal angle of view than it should for 14mm, whereas it gives the expected angle of view (for 14mm) after barrel correction. Same holds true for my Lumix 7-14mm, which is a fair bit wider (based on diagonal angle of view) than expected for 7mm before correcting the barrel distortion and just about right after correction. Having tested the Oly mZD 9-18mm at 9mm, I'm 98% sure it gives the expected angle of view for 9mm after barrel distortion correction.
Amin, thank you very much for the explanation. That cements 9-18 as the travel lens to have imho.:2thumbs:
 

dansenor

New to Mu-43
Joined
Jul 27, 2010
Messages
9
Links on this page may be to our affiliates. Sales through affiliate links may benefit this site.
Mu-43 is a fan site and not associated with Olympus, Panasonic, or other manufacturers mentioned on this site.
Forum post reactions by Twemoji: https://github.com/twitter/twemoji
Copyright © 2009-2019 Amin Forums, LLC
Top Bottom