Tokina 14-24 FX fisheye announced

You wish! OK, so I had to get your attention somehow. Following all the great testing by Steve in the last few days, I got to thinking how complex it is to get these wide angle rectilinears to work consistently across the frame, and how many variables there are using them.

As a recent convert from DX to FX, I have mourned the loss of my Tokina 10-17 as well. So the obvious solution is to add a Kenko 1.4 TC and voila, we have a Tokina/Kenko 14(15)-24 Fisheye zoom.

So what do I know so far. Yes it can be done. It vignettes at 10mm, but by 11mm the vignette is gone.

So I started some tests.

Corner to corner on the Tokina is pretty acceptable. Shot at f8.

So we're done topside; Ryan has a gear on the way, and this baby's going in the pool today, 50 degree weather or not. Set up so far. D700, Subal FE2 dome with 30mm extension (I would maybe prefer a 20-25 to put the lens plane at the FE2 opening.) If all goes well I'm taking it diving at a Caribbean location next week.

Stay tuned. 14-24 comparisons at least topside coming as well.

Oh, and one last thing, the Tokina focuses right on my domeport as before, something the Nikon cannot match.

What an excellent idea!

Presumably if you used a 1.5 teleconverter instead you would avoid the vignette at 10mm.

I look forward to the underwater results.

Colin

Jeff,

I wondered with all this hullabaloo on rectilinear lenses when someone with a FX camera would try this.........

A couple of things......The Tokina MAY auto-focus a tad slower in less than bright light, but on modern dSLR bodies might not be a problem.

Another thing I'd try since it can focus right up to the dome is simpy pre-setting the focus manually at like 10" - 11". Letting the depth of field of the fisheye do the job!

You could still zoom a bit and I'll bet not lose that much detail from 11 -17mm as you say it works with no vignetting. having to zoom from 10-11mm is no big deal in my book......Especially if the images are sharper edge to edge.

Larry Oberlander shooting I think a D200 used the preset focus very effectively on one of my Shearwater trip with the dolphins. He won a great prize with one shot, too! So maximizing the Fisheye lense's DOF capability can work in addition to solving edge problems.

Just thinking out of the box here. If this works maybe Canon 5D MK II shooters can do it with a 1.4X TC too!

Keep us appraised of your progress!

dhaas

I'll meet you in Grand Cayman and then you can have a backup D700.

You wish! OK, so I had to get your attention somehow. Following all the great testing by Steve in the last few days, I got to thinking how complex it is to get these wide angle rectilinears to work consistently across the frame, and how many variables there are using them.

As a recent convert from DX to FX, I have mourned the loss of my Tokina 10-17 as well. So the obvious solution is to add a Kenko 1.4 TC and voila, we have a Tokina/Kenko 14(15)-24 Fisheye zoom.

So what do I know so far. Yes it can be done. It vignettes at 10mm, but by 11mm the vignette is gone.

So I started some tests.

Corner to corner on the Tokina is pretty acceptable. Shot at f8.

So we're done topside; Ryan has a gear on the way, and this baby's going in the pool today, 50 degree weather or not. Set up so far. D700, Subal FE2 dome with 20mm extension. If all goes well I'm taking it diving at a Caribbean location next week.

Stay tuned. 14-24 comparisons at least topside coming as well.

 

Oh, and one last thing, the Tokina focuses right on my domeport as before, something the Nikon cannot match.

Jees thats a good idea Loftus! Being a Canon man it made me think to try my Sigma 10mm FE with a 1.4 Teleconverter on my Canon 5D Mk II. Well it works beautifully and does not vignette!!!!!!!!. I have shot a few pictures like this and compared it to the full frame Sigma 15mm on the 5D2 and it appears to be just as good if not maybe even better. I am going to play around with it but it is still going to be a while before I can try this UW ( waiting for Subal housing). But it sure looks promising and for one thing it will allow me to carry only one FE lens for both my full frame and cropped sensor cameras. Thanks

OK, time to post.

First I took a couple of pics of my monitor, just to get an idea of FOV comparison. The Tokina set at 13-14 (around 18-20mm equivalent) gives a similar field of view to the 14-24 Nikon at it's widest. First the Nikon, then the Tokina.

Then it was time to get in the pool, or at least get the camera in the pool. Water temp's about 60 degrees, friggin cold for Florida.

I laminated a couple of focus test charts I have and sunk them to the bottom of the pool. I used the 30mm extension with the FE2 domeport. I think I could use a 20-25mm one if I had it..

I took 3 sets of pictures, 11mm, 14mm, 17mm on the Tokina ring. First at f8 then f11. All at ISO200

At the 11 and 14mm settings I show 100% crops of the outer chart and the fish tile on the upper periphery.

Overall pretty decent I think; this puppy's going 'unda da sea' next week.

11mm f8

11mm f11

I guess I need two more posts to allow all the pics; it's not allowing me to post the rest at the moment.

In February of last year I tested the close-focus abilities of the Tokina 10-17 alone and with a Kenko 1.4x TC on the 1Ds Mark III (topside). I'll see if I can dig up the images...

Tokina 10-17 (and w/ 1.4x TC) on 1Ds Mark III (Warning - long post)

Shot some newspaper classifieds - I really hadn't realized the wild amount of image plane curvature with a fisheye lens prior to this.

Distance from the newspaper was approximately 3-4". Tripod was on carpet so I used 2sec timer, but I did bump it forward once or twice - it doesn't really impact the actual image much, just a micro-change in near/far perspective. All files converted from RAW in DPP with no adjustments, sharpness level of 5. (No additional PP)

I've grouped equivalent focal lengths together for relatively quick comparison (i.e., Tokina 10-17 @ 14mm and Tokina 10-17+1.4xTC @ 10mm (14mm equiv.) are grouped)

Tokina 10-17 @ 10mm

Overall field of view sample:

100% Center and corner crops:

f/3.5:

f/4:

f/5.6:

f/8:

f/11:

f/16:

Tokina 10-17 @ 14mm

Overall field of view sample:

100% Center and corner crops:

f/4:

f/5.6:

f/8:

f/11:

f/16:

Tokina 10-17+1.4x Kenko TC @ 10mm (14mm equiv.)

Overall field of view sample:

100% Center and corner crops:

f/3.5 (f/5 equiv.):

f/4 (f/5.6 equiv.):

f/5.6 (f/8 equiv.):

f/8 (f/11 equiv.):

f/11 (f/16 equiv.):

Tokina 10-17 @ 15mm

Overall field of view sample:

100% Center and corner crops:

f/4:

f/5.6:

f/8:

f/11:

f/16:

Tokina 10-17+1.4x Kenko TC @ 11mm (15.5mm equiv.)

Overall field of view sample:

100% Center and corner crops:

f/3.5 (f/5 equiv.):

f/4 (f/5.6 equiv.):

f/5.6 (f/8 equiv.):

f/8 (f/11 equiv.):

f/11 (f/16 equiv.):

Tokina 10-17 @ 17mm

Overall field of view sample:

100% Center and corner crops:

f/4.5:

f/5.6:

f/8:

f/11:

f/16:

Tokina 10-17+1.4x Kenko TC @ 13mm (18.2mm equiv.)

Overall field of view sample:

100% Center and corner crops:

f/4 (f/5.6 equiv.):

f/5.6 (f/8 equiv.):

f/8 (f/11 equiv.):

f/11 (f/16 equiv.):

Tokina 10-17+1.4x Kenko TC @ 15mm (21mm equiv.)

Overall field of view sample:

100% Center and corner crops:

f/4 (f/5.6 equiv.):

f/5.6 (f/8 equiv.):

f/8 (f/11 equiv.):

f/11 (f/16 equiv.):

Tokina 10-17+1.4x Kenko TC @ 17mm (24mm equiv.)

Overall field of view sample:

100% Center and corner crops:

f/4.5 (f/6.3 equiv.):

f/5.6 (f/8 equiv.):

f/8 (f/11 equiv.):

f/11 (f/16 equiv.):

Pretty reasonable I think. Ever think of taking it underwater?

Now I'll add the rest of mine.

14mm f8 - Probably the most interesting for comparison to the Nikon 14-24 at 14

14mm f11

17mm f8 - Probably the most interesting for comparison to the Nikon 17-35 at 17

17mm f11

Thanks for throwing yet another monkey wrench in my "pick the next setup" sweepstakes!

Seriously, that looks pretty good ...

Jees thats a good idea Loftus! Being a Canon man it made me think to try my Sigma 10mm FE with a 1.4 Teleconverter on my Canon 5D Mk II. Well it works beautifully and does not vignette!!!!!!!!. I have shot a few pictures like this and compared it to the full frame Sigma 15mm on the 5D2 and it appears to be just as good if not maybe even better. I am going to play around with it but it is still going to be a while before I can try this UW ( waiting for Subal housing). But it sure looks promising and for one thing it will allow me to carry only one FE lens for both my full frame and cropped sensor cameras. Thanks

Sorry just one more for the Canonites:

The Sigma 10-20mm with a 1.4 Teleconverter also works just great on the 5D Mk II without vignetting

I should also edit my post above in that on very close inspection the Sigma 10mm FE/ Teleconverter 1.4/ 5D2 combo DOES show a TINY amt of vignetting ( which could easily be corrected without cropping)

Pretty reasonable I think. Ever think of taking it underwater?

I sold it .

With the teleconverter, I felt it too slow (f/5-6.3) and lacking the quality I was looking for. The usual benefit of using a fisheye lens underwater is the ability to obtain sharp corners nearly wide-open, or without stopping down as much as a typical rectilinear. Unfortunately, the lens appears to perform more poorly (from center crops) once the teleconverter is added until it is significantly stopped down, defeating that benefit.

Your results may vary!

I sold it .

 

With the teleconverter, I felt it too slow (f/5-6.3) and lacking the quality I was looking for. The usual benefit of using a fisheye lens underwater is the ability to obtain sharp corners nearly wide-open, or without stopping down as much as a typical rectilinear. Unfortunately, the lens appears to perform more poorly (from center crops) once the teleconverter is added until it is significantly stopped down, defeating that benefit.

 

Your results may vary!

I was about to sell mine; and I'm glad (at least till I try it in open water) that I didn't.

A few things:

Your tests are done topside, testing close focus abilities, and not behind a dome at average shooting distances. Just from my own observations of my own tests, my lighting with the natural light in the pool is more even and with less noticeable drop off towards the periphery. DOF will also be greater across the frame at average shooting distances.

I would not consider this combination of course for topside, but I do think putting a fisheye behind a dome is likely to make a relatively better match (at least with the FE2 designed for the 16mm) than any rectilinear. Putting a rectilinear behind present dome designs is always a little like putting a square peg in a round hole.

My main interest here was to compare corner sharpness to two lenses that I own: The 14-24 Nikon at 14, with the Tokina at similar FOV(13-14 on the ring), and the 17-35 Nikon at 17 with the Tokina at similar FOV (16-17 on the ring). I suspect Canon's wide angles to be similar to the 17-35. I use other peoples' tests for these two lenses for my comparisons, which of course is not ideal.

My impression so far is that at f8 this setup compares favorably and may even be better than either of these two rectilinears at the corners, and definitely better than the 17-35 without a diopter. One limitation of my tests is the glare on the peripheral chart, which I think degrades the quality a little. I will try to repeat sometime with better more even lighting. I have yet to see any Canon rectilinear tests that convince me they are better as well.

The other concern I have about all the rectilinear tests is that to some degree they are very specific for the exact setup of each specific test. With a rectilinear it is only possible to say that for this exact, lens, port extension, dioptre, port, and distance to the test chart at the periphery, we have achieved a certain degree of sharpness at the periphery. Change any one of those parameters, particularly peripheral test chart distance by even a few inches, and the conclusions may be different.

At least for the FE2 domeport, being designed for a fisheye, knowing the DOF characteristics of the Tokina, I am certain I will have a far greater range of focus at the periphery than I am likely to have with any rectilinear. In practice, this should allow a much larger range of conditions and subject planes that are likely to be in focus across the whole image. The close focus ability of this setup is also unmatched by any rectilinear.

The slowness of the setup is much less of a concern to me now with the D700 compared to my D200, knowing I can shoot at 800 or 1600 if need be and still get quality as good or better than I could get with 200 ISO on the D200. If someone ever actually makes an FX fisheye zoom, of course I would like it to be faster particularly for autofocus, and I'm sure it would be. Come to think of it, this combination seemed to autofocus OK when I used it with DX. No question this is a 'poor mans' setup, but I have not seen any wide rectilinear underwater tests yet that blow me away as being significantly sharper. My 14-24 is a spectacular lens topside, and if I saw anything that made me think it can be easily and repeatably replicated underwater, I would adopt it in a heartbeat.

I think one should also keep in mind that considering the popularity of the 10-17 on DX, a 14-24 FX Fisheye zoom which this setup mimics, is a very desirable lens. A fisheye zoom is considered by many as one of the 'deficiencies' currently in full frame lens availability for underwater use.

A few more tests today. The light was a little more flat so less glare.

Today I only tested at setting of 14(19.6 with TC), about the width of the 14-24 at it's widest. The 17 (23.8 with convertor) setting only gets better as I have already shown.

These are from f4 through f11 with 100% crops of the peripheral chart.

One pleasant surprise was the sharpness of the pool light on the left, particularly at f8 and f11. The pool light is half way up the wall, whereas the charts are on the bottom. This I think demonstrates the field of curvature of this lens, which is probably beneficial in many underwater situations. I also think the distortion at the 14 setting is quite acceptable. One of the things that is a little different as well from rectilinears, is that even though there is loss of sharpness towards the periphery particularly at low apertures, one does not seem to see the 'smearing' loss of sharpness that one sees with the rectilinears in the corners. And of course, over under possibilities made more difficult with a rectilinear.

Anyone see any problems I am missing?

f4

f5.6

f8

f11

f11 image showing significant sharpness of pool light and acceptable distortion.

Can you zoom her in the housing with the TC attached? How did you do this if yes?

Can you zoom her in the housing with the TC attached? How did you do this if yes?

Yes; Ryan from Reef Photo sent me the Canon EF28-105 gear (Subal Code 4ZC057)

Hi Loftus,

I have read through all the posts on using the Tokina with an extender on a FF body. Very interesting and since I will be shooting a Subal FE dome on my first (2nd hand) rig soon, I was wondering if the 30mm extension is still what you use for this lens+extender combo?

All best,

David

Hi Loftus,

I have read through all the posts on using the Tokina with an extender on a FF body. Very interesting and since I will be shooting a Subal FE dome on my first (2nd hand) rig soon, I was wondering if the 30mm extension is still what you use for this lens+extender combo?

All best,

David

Yes David; the 30mm extension ring is my choice, probably could use a fraction shorter if I had it, but 30mm works fine.

Interesting. Thanks for posting.

Of course you could always just use it straight without the TC in crop mode. I'm not sure which is worse--taking the performance hit optically with a TC or just taking the resolution hit by cropping.

On the one hand you have to stop down, focus is slower, and the pixels you get are less sharp. You also need special port extensions and gears.

On the other hand you still get the benifit of high ISO, fast focus, normal apertures (for subject isolation) and simplicity but only ~6mp.

6mp is enough and that's what I get with my D70. So all things considered until there is a true FX replacement I'd just do the simple thing and use this lens on crop mode.

Interesting. Thanks for posting.

 

Of course you could always just use it straight without the TC in crop mode. I'm not sure which is worse--taking the performance hit optically with a TC or just taking the resolution hit by cropping.

 

On the one hand you have to stop down, focus is slower, and the pixels you get are less sharp. You also need special port extensions and gears.

 

On the other hand you still get the benifit of high ISO, fast focus, normal apertures (for subject isolation) and simplicity but only ~6mp.

 

6mp is enough and that's what I get with my D70. So all things considered until there is a true FX replacement I'd just do the simple thing and use this lens on crop mode.

I completely disagree on that one.

I disagree that the pixels you get are less sharp.....I simply don't see a quality loss with my setup. I've posted a number of pics that show this setup to be pretty sharp. I've also not heard this criticism when folks like Alex posted CFWA pics using the TC and FE on DX before. My edges seem to be similar to what they were on DX, and still better than my rectilinear options.

Yes you loose a stop, but I have simply not found focusing speed to be an issue. Focuses fine and fast, at least in normal lighting conditions.

In the big scheme of things a port extension and gear are hardly big ticket items. One already has to use a short port extension and ring with the 10-17 on DX, so no more complicated.

6MP really is a limiting factor for anything over 8x10 enlargements which I routinely do. 6MP is OK for web, but not much else for my purposes.

Having used this setup quite a bit now, I can say that there is only one disadvantage in my opinion....in low light (poor viz), the viewfinder is darker than I would like because of the one stop loss in maximum aperture, otherwise it performs in every other way as it did when I used the 10-17 on my D200.