It's actually pretty easy to take a boring photo when you go so wide, as alluded to above... Look for threads of a first timer's UWA shots, there's always more than a few that feature 1/3rd of a frame of huge empty sky, 1/3rd of featureless ground, and a tiny 1/3rd of the frame depicting some sorta far away looking landscape. A FE might not push the middle away quite so badly but it still has the same potential, letting the distortion overpower every shot etc.

What is better is obviously subjective, but I personally find fisheyes look more like what I see with my eyes for some reason.
I used to have a non zoom 15mm fisheye, and I loved it on a crop body. Great for close portraits in my opinion, as they don't distort like a typical wide.

But here's the thing, if you don't like the effect, you can adjust the amount of rectilinearity (real word?!) in post process. Lightroom makes this easy for example. You can do the same in reverse to create a fisheye effect from a normal wide, but it never looks quite right.

I'm with PhotKhan on this, though I'd stop short of calling it an "uninspiring tool". Perhaps the shooter just didn't find the right inspiration. My 14-24 can rest easy, as this FE zoom sure isn't going to displace it.

I'm sure he just took the one lens...

I find FE lenses to be kind of gimmicky. The shots on the long end look good to me, but the 14-24 would've been similar to those.

Just to make things clearer, I meant "uninspiring" in the sense of what was lost in those locations, with that light, by being forced to use this lens instead of a, say, Nikon 14-24. Fisheyes on landscapes are not very effective tools unless the "narrative" specifically makes use of their characteristics - like it so beautifully does in that bee shot.

And you don't think Fujifilm has quality issues?

This is from when I first had the lens
http://www.northlight-images.co.uk/canon-ef-8-15mm-f4-l-usm-review/

Here's a more recent fisheye specific review
http://www.northlight-images.co.uk/samyang-12mm-fisheye-lens-review/

Hope that gives a bit of a feel for how I use fisheye lenses?

Stick with it when you get one - they really can justify the time to get a feel for what can be done. Let others dismiss them as just for trick use ;-)

Agreed. Patagonia is a massive area. For the most part these photos were taken in Torres del Paine National Park in Chile and Los Glaciares in Argentina. There are also a few from the Alabama Hills in California.

No, distortion refers to linear objects not being linear. Most UW lenses do not result in distorted images because they are corrected (optical design and or software processing).

All lenses cause distortion. This is because they are smashing a spherical world onto a flat surface. Since long lenses are smashing only a tiny portion of that sphere, the distortion isn't very noticeable. Ultrawides always create noticeable distortion because they are capturing so much of the sphere.

This issue is, what type of distortion is preferable for a given shot?

Rectilinear ultrawides keep straight lines straight but turn circles into ovals and stretch the heck out of anything near the corners.

Fisheye ultrawides keep circles circular and don't stretch the corners but they turn curved lines that don't point at the center of the image frame into curves.

Personally, I find fisheye distortion preferable to rectilinear distortion on about 98% of my ultrawide shots, and I absolutely love my Canon 8-15 on my crop camera where it's wider than the 11-24 on full-frame and can be fisheye or defished, in part or in full, to rectilinear.

@lee
That's not true. The world (universe) is not spherical. The problem is with the lens, not the world.

@virtualreality
We can perceive the word as:
- keeping eyes still: 180°, but projected on an half sphere, and only the middle is sharp;
- moving the eyes: only the sharp middle is important, we see the world as a scan of a lot of pictures of a fov like a 200-300 mm, that merged are like a spherical projection.
I may say that the problem is not the lens, nor the world, but the way we have to see pictures, flat prints and screens. Seen from a VR, or printed inside a sphere they would look natural. That's the same exact problem of flat world maps.

From the point of view of the lens aperture, which is small relative to the world, the world is spherical. The aperture is a point looking around from the center of the sphere (let's ignore macro and telecentricity for this discussion). The mapping of that spherical world onto the flat plane of the sensor or film (called "projection") is the same problem as trying to map the spherical globe onto a flat map - something has to give.

And you're wrong, since rectilinear lenses have exactly the same problem (mapping a spherical world onto a flat plane). And as I already said, at least for me, most of the time rectilinear distortion is *worse* on ultrawides. To put it another way, when I got my fisheye, I nearly stopped using my rectilinear ultrawide.

Toselli above said it right - those 180 degree IMAX domes use a fisheye lens with the image projected onto the inside of a half-sphere. That's really the only way to do imagery without distortion.

A room with round tables - an ultra wide rectilinear makes the walls looks straight but makes the tables look badly shaped. The fisheye will give curved walls but the tables will look right. Fisheye large group shots looking down from a step ladder look far better for people because the people at the sides don't look double width. You can try and remove volume deformation from a rectilinear lens and you can defish a fisheye image either in both axis or just one.

The bigger point is simply that using a fisheye lens doesn't necessarily produce images that look fisheyed.

Take both at the same location/time/light. See, then. So far it is clear bold statement, sorry.

I'm always baffled how people can tell properties of lenses or cameras they never used.
They sure must be very experienced to be able to judge this well.

Actually, it doesn't.

Well he is not being truthful.

In comparison to a 8-15mm, you will stick with your 14-28mm and 15mm.

Thanks for adding to the convo.

"I think the Tamron 15-30mm is ok, but I will stick with my 28-70mm."

^ That's you.

Where's your M43 8mm equivalent? ;-)

You jumped into the convo on a lens that your format has absolutely no equivalent for, 4-7.5mm, to proclaim you prefer your format's options that are not at all the same or even able to cover the same focal lengths.

Do explain the utility in your contribution.

I made an original comment, you were the second person to reply, albeit unnecessarily negative when you could've kept you could have simply said nowt.
Get over it. I just did :-)

I've never seen anyone on DPR jump into the comments for a product that has nothing to do with their camera system to say, "muh unrelated camera system is cool!"

Nope, not once. Very original.

I tried once with the lightroom presets on my sigma 15/2.8 (diagonal fisheye), and the effect was incredibly unrealistic. It seems something that doesn't even exist like a 6-8 mm rectilinear on full frame! Things towards the margin look a lot bigger than the ones in the middle! And obviously you loose a lot of sharpness outside the central portion.

On a 3:2 aspect ratio sensor the horizontal and vertical FOVs are very approximately 150 and 100 degrees respectively.

If you take something like Nikon's current 16mm fisheye, hold it dead horizontal, it produces a fairly normal-looking but wide image. Pointing it downward or angling it upward makes the fisheye effect more apparent.