Camera Labs

[Thomas]

Don't stumble across all the Nikonian abbreviations on this one. Just wanted to make sure you know what I'm talking about!

I had the opportunity today to test this lens against the Sigma 10-20mm 1/4-5.6, because I was searching for the ideal lens to complement my standard 18-200mm VR Nikkor. Well, I think I've found it!
For my struggle with the Sigma lens, read here: http://www.cameralabs.com/forum/viewtopic.php?t=182

I will update you on my impressions up to March 2nd, then take the lens to New Zealand and report on my findings from that challenge. So, stay tuned...

1. At first let me state that this is a fisheye so almost all lines are reproduced as curves. BUT if you like , you can correct for this with the CaptureNX software. But ONLY if you shoot in RAW-format
2. The lens focuses up to 4cm from the front lens Can you imagine what this can do to the perspective of your picture?
3. It is an AF lens, not a AF-S. That is, the focussing is done by the cam-body, not the lens. Meaning: this lens is not for the Nikon D40!

See a pic of another fisheye-owner that is one of my favourites: http://www.flickr.com/photo_zoom.gne?id=264170988&size=o
If anyone else has personal experience with this fisheye, please report here!
_________________
Thomas (beware: Nikon-fanboy and moderator!) My Pictures
D300/18-200VR/SB800+other lenses from 10 to 500mm, Manfrotto055ProB/808RC4, CanonPixmaPro9000...

[Thomas]

Now here's the meat:
You'll find two indoor photos at flickr shot with the fisheye 10.5mm Nikon lens at 1/30s, f4, ISO 1600 (so don't complain about the grain )
normal:


defished:

There are several things to note:
- contrast is unbelievable: You see the windows totally overblown from the outside light @ (255,255,255) R,G,B , but still there are pixels with (5,5,5) at the front of the loudspeaker.
- not all fisheye-perspectives are radically curved: I found this pic in its normal fishyeyed version not too extreme! And if you shoot outdoors in the wild (no architecture!) the curvature shows up even less.
- correction through CaptureNX (defishing) is great: All the lines straightened out and quite sharp (look at the tiles)

I've included another examples of a "low curved" indoor fisheye-shot here:
as is: http://farm1.static.flickr.com/166/407084088_d9d0305444_b.jpg
defished: http://farm1.static.flickr.com/135/407084488_f615ff739b_b.jpg
In this case I even find the defished perspective less "natural" than the fisheye-perspective. This clearly comes from the disadvantages of tilting the lens down (or up) from the horizon.

In both cases you cannot really see a loss of sharpness in the defished corners. If you could see the pics in original resolution, you can certainly see some loss in a 100% crop. But only so slightly in these cases.
There are other cases though, where CaptureNX has a lot more stretching to do and that shows in the corners.

To be continued...
_________________
Thomas (beware: Nikon-fanboy and moderator!) My Pictures
D300/18-200VR/SB800+other lenses from 10 to 500mm, Manfrotto055ProB/808RC4, CanonPixmaPro9000...

[Phil Lee]

Those are really interesting Thomas. It would also be interesting to see what a specialist piece of software such as DXO Optics Pro makes of the fish eye correction.
_________________
Nikon D80, Nikkor lenses: 50mm f/1.8 AF D, AF-S VR Micro 105mm f/2.8 G ED, 18-70mm f/3.5-4.5G ED-IF AF-S DX, 70-300mm f/4.5-5.6G IF-ED AF-S VR, SB800 flash
My photos on SmugMug

[Thomas]

I don't have DXO, Phil!
If you have, I can send you the pics and you can make a comparison between CaptureNX and DXO.
In CaptureNX you can only toggle between a straightened normal version as above and a maximum width version that tries to cover almost the same angle vertically than the fisheye, but that shows missing picture information for a rectangular pic like the following:

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Pictures
D300/18-200VR/SB800+other lenses from 10 to 500mm, Manfrotto055ProB/808RC4, CanonPixmaPro9000...

[Gordon Laing]

Hi Thomas, nice examples!

Are there any tools in Adobe Camera RAW which could help? I suspect Nikon's own Capture NX would do the best job though with one of its own lenses.

So the big question is whether it's a keeper? Has the Sigma gone back and will the Nikkor be coming with you to NZ?

Gordon

[Phil Lee]

[Thomas]

Yep! The Sigma is back (to square one). I can't stand a lens that doesn't deliver consistent results. Sometimes it was sharp, sometimes not. You were never sure what the outcome was
The Nikkor will come to NZ
I will not try DXO - have enough to do anyway
I'll keep ye'all posted...
_________________
Thomas (beware: Nikon-fanboy and moderator!) My Pictures
D300/18-200VR/SB800+other lenses from 10 to 500mm, Manfrotto055ProB/808RC4, CanonPixmaPro9000...

[Thomas]

B.t.w. you can even shoot perfectly normal looking photos with a fisheye as long as you adhere to two rules:
- keep the horizon strictly in the centre of your pic
- don't shoot architecture...

The proof is here!

_________________
Thomas (beware: Nikon-fanboy and moderator!) My Pictures
D300/18-200VR/SB800+other lenses from 10 to 500mm, Manfrotto055ProB/808RC4, CanonPixmaPro9000...

[zorro]

Hi Thomas.

As you probably know I have the 18-135mm kit lens. I bought that lens because I had a holiday in Corsica coming up and due to a lack of funds I needed something that would do it all. Anyway, I only really shoot landscapes and so I was looking at the Sigma 10-20mm. Having read your opinions - and those of others - I think I'll give it a miss.

I'm having a hard time imagining what 15mm (effective) lens actually behaves like, even with these pictures you've supplied. Could you maybe link to a couple of landscapes you've shot with the 10.5mm lens? I'd really appreciate it. Do you think it would serve as an all-around landscape lens? Or do you think it's maybe too wide?

Zorro.
_________________
http://www.casazorro.org/index.html

http://zorrofox4.deviantart.com/gallery/



Various glass, SB800 & Manfrotto 190 with 460MG head

[Bob Andersson]

Hi Tom,

I started another thread asking about fisheye lenses on cropped sensor bodies without realising that you had done a review of the Nikon offering. Can you confirm my suspicion that such lenses don't give the expected 180° diagonal view on such cameras?

If so then I wonder why neither Canon nor Nikon has produced a true fisheye for their cropped sensor bodies.

Bob.
_________________
Canon EOS 5D Mk II, EF 16-35mm f/2.8L II USM, EF 24-105mm f/4L IS USM, EF 70-200mm f/4L IS USM, EF 85mm f/1.2L II USM
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head, PowerShot G10 and some other stuff.

[Thomas]

Oh boy, I nearly missed your entries.
So here it is - a link to all my flickr 10.5mm fisheye shots:
http://www.flickr.com/photos/thomasrubach/sets/72157600086869958/
You can also do a search on "nikon fisheye" at flickr to see other people's results.

As to Bob's question: I didn't measure, but when you tilt the cam to diagonal/upright you might easily capture your feet!
See some nice test on it here:
http://www.fotoverkstan.se/fisheyetest/

And some nice photos here: http://www.pbase.com/northqueenslandphotos/image/43490026
_________________
Thomas (beware: Nikon-fanboy and moderator!) My Pictures
D300/18-200VR/SB800+other lenses from 10 to 500mm, Manfrotto055ProB/808RC4, CanonPixmaPro9000...

[Bob Andersson]

[Thomas]

Hi Bob, but sorry, I didn't get that!
Do you imply that the 10.5mm is not a fisheye?
And what do you mean by "angular cut-off"?
_________________
Thomas (beware: Nikon-fanboy and moderator!) My Pictures
D300/18-200VR/SB800+other lenses from 10 to 500mm, Manfrotto055ProB/808RC4, CanonPixmaPro9000...

[Bob Andersson]

Hi Tom,

As you know, if you strap a 50mm lens on a cropped sensor body it behaves as though it is the equivalent of an 80mm lens (roughly) on a full frame body. Another way of looking at this (sorry, bad pun) is that the field of view or angular coverage of the 50mm lens has been reduced to (roughly) 5/8ths of its full frame capability when used on a cropped frame body.

The Canon fisheye, and I assume the Nikon one is the same, is designed to give a full 180 degree diagonal angular coverage on a full frame sensor. On a cropped sensor the fisheye is obviously not focussing the image onto the sensor over that full 180 degree range because a cropped frame sensor isn't big enough.

The thrust of my question was how much of that angular coverage is lost. Because of the geometry of the barrel distortion it may not be as simple as the 3/8ths indicated above. It seems from your own experience that less than this is lost.

Note, I am not suggesting that the Nikon fisheye doesn't behave like a fisheye on a cropped sensor body. It obviously does because of the barrel distortion. I am suggesting that there really should be dedicated fisheye lenses for cropped sensor bodies if one wants to get full 180 degree coverage.

Bob.

P.S. I've probably made a meal of trying to explain what I meant. Hope I haven't been too opaque.
_________________
Canon EOS 5D Mk II, EF 16-35mm f/2.8L II USM, EF 24-105mm f/4L IS USM, EF 70-200mm f/4L IS USM, EF 85mm f/1.2L II USM
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head, PowerShot G10 and some other stuff.

[Bob Andersson]

Thanks to both Gordon and Tom for clearing up my question in this thread.

Nikon wins this round as the Canon EF 15mm f2.8 Fisheye is designed for full frame sensors and so may not provide the expected effect on cropped sensor bodies.

Bob.
_________________
Canon EOS 5D Mk II, EF 16-35mm f/2.8L II USM, EF 24-105mm f/4L IS USM, EF 70-200mm f/4L IS USM, EF 85mm f/1.2L II USM
Gitzo GT1541T, Arca-Swiss Z1 DP ball-head, PowerShot G10 and some other stuff.