Wide Angle Lens

The Pentax 10-17mm has a wider angle of view, in fact 180 degrees across the diagonal, but is not corrected for distortion. There were originally scientific applications that required this and indeed other different fisheye projections but for our purposes the effect is a creative one.

The Sigma 10-20mm is corrected for sitortion, but the Pentax 12-24mm is much better in this respect, almost perfect in fact.

Not according to Photozone's tests - unless I am misunderstanding them.

http://tinyurl.com/ywgx28

http://tinyurl.com/2h4luh

Thanks for the links.

The Sigma 10-20mm is corrected, i.e. it is not a fisheye, but it is true that it might be considered to have not been corrected very well.

The Pentax 12-24mm in my experience is corrected very well, and this was supported by Geoffrey Crawley's lens test in AP. I use this lens routinely for architecture shots and there is no problem. I can't really comment on Photozone's results - they say what they say so we can take that on board. Other tests seem to have been rather different.

Perhaps that's an argument for using tests as a guidance only but checking things out for ourselves.

Pentax 10-17mm is a fisheye lens, meaning it is of curvilinear design. The Sigma 10-20mm and Pentax 12-24mm lenses are both of rectilinear design. The latter keep straight lines straight (as best they can anyway) no matter where they are in the frame. Fisheye lenses will only produce straight lines across the centre of the lens.

Software is available to correct the fisheye effect making it more like the results from a 'proper' lens. (Let's see who bites . . . fisheye . . . bites . . . get it? Oh never mind.)

I have a Sigma 10-20 and it seems astonishingly well corrected to me. So far I've only noticed lines being curved when the lens is ridiculously close to the subject.

The Pentax 12-24 may be excellent and better in some ways, but it is more expensive.

It's a matter of degree and also of the sort of subject matter undertaken.

In any event I would say a small degree of barrel distortion would be acceptable in an ultra-wide lens. It's what we expect. An ultra-wide with pincushion distortion of any degree would be totally unacceptable though and would be a strange beast indeed.

It should be noted that Photozone are focusing on a chart of 120x80cm . which with a 10mm (or 12mm) lens means they are pretty close to it - and WA lenses often 'barrel' more at closer focus.

Thanks, that's a very good point. These sort of lenses are normally used at normal veryday distances, not close up, so that probably explains the discepancy. And probably means their test is meaningless as well.

Which makes the results of both even more impressive. And they do both look very good indeed compared to "standard zooms" such as the 18-55 and 16-45mm lenses at the wide end.

My ony point was that the Sigma 10-20mm is not far worse than the Pentax 12-24 according to Photozone. If anything, the Sigma is better at 10mm than the Pentax is at 12mm (except at the extreme corners) - according to Photozone.

I would expect overall the Pentax to be superior to the Sigma, but not by much. Therefore the cost might be the final decider for most people. It was for me, but then I'm not rich

Being rich is a state that can quickly meet its end when shopping new gear

As i know only too well, i keep reaching for my back pocket to check if wallet is still there, Its so thin these days

Thanks for everyone's comments and links to other threads.

I ordered a SIGMA 10-20mm from SRS.

I'll post some photo's when I get a chance to get out - a lot harder now I've got a baby son...

Cheers
M

They're each designed that way. The Pentax is for those that want the fish-eye effect but the Sigma is a normal wide angle zoom ( though it does seem to exhibit a wee bit of distortion at the extreme edges.