Corner softness with Ikelite dome port?

I'm pretty new to the underwater game. This is my 4th roll and I'm having a bit of a problem with sharpness.

I'm using an Ikelite housing with an 18-35 EDIF Nikkor, a B&W No. 4 diopter and an Ikelite 5503.50 dome port. I'm not sure what the cause is, but I'm finding that I'm getting consistently soft corners with this combination (on an F80).

I'm not really sure what part of the system to blame (it could be the photographer ). I find the 18-35 to be very sharp when used at the same apertures topside, so I assume it must be either the diopter or the dome.

Is anyone using a similar combination and do you have the same issues?

Has anyone got a suggestion on how this might be solved?

The attached photo may not be the best example because the soft corners are closer to the lens, but even when I'm shooting a relatively flat subject parallel to the film plane at small apetures (f11 and smaller ), I'm not getting results as sharp as I woud expect.

Photo is from St Leonards Pier, Port Phillip Bay, Melbourne, Australia in case anyone was interested.

Cheers

BTW, in case anyone suggests this - I don't think it is ambient light (and slow shutter speeds) causing the softness, because it only occurs in the corners of the image and it is still present even in shots with no ambient light, ie., flash key-light shots.

anthp,

I got the same problem as you with Tokina 19-35 and same dome port. I tried f8, f11 and f16 without good results. Does anyone knows more about this problem ?

Best Regards

Fabio

You are right anthp, that's not shake, it looks like aberration to me. I would guess that you don't have the entry pupil of the lens at the centre of the port. It may require addition of an extension behind the port.

I do wonder if the diopter moves the entry pupil forward - can't do the optics theory off the top of my head.

Ike may be able to shed light on the issue.

Martyn

Hi,

I've used this exact same combination. I even have a B+W +4...

Sometimes the blurring you see is simply because the depth of field is not enough to extend from the dome port out to infinity. Some things near the dome may not be within the depth of field.

Other times, corner softness underwater is due to a bad dome/diopter/lens combination. Some lenses simply don't work well with diopters, above or below water. The 18-35 is not a "pro" lens so that contributes a bit. I also don't think it particularly likes diopters. But based on my experience, I agree with the above that you may need a different length "tube" on your dome port.

Here's a shot I took with the 18-35 that shows corner softness. It's both close to the lens and far from the lens...

Sorry, but I don't have the Exif for this one - I think it was at f5.6

Cheers

James

The famorim photo is simply beyond the depth of field capability of the lens. If everything could be in focus from one foot to infinity, we would not need to focus above water or below. Basic photography rules. James' photo could be fish motion, but I am just as mystified by the dark lower corners of the photo. And we have seen lenses that simply do not work well behind a dome. You might try a +2 diopter rather than +4 on the lens and let us know if that helps.

Thanks everyone for your suggestions and assistance. Great photo James! I'm interested to see that you are still getting corner softness even with the 1.5 crop.

I will try the No. 2 diopter. Has anyone else used this combination?? Any brand recommendations? Who make double element No. 2. diopters in the 77mm filter size?

Thanks again.

Hi Ike,

I forgot about this information, but I took this pict using a Roya +4 diopter. The Tokina 19-35 is far from a pro lenses and is very cheap, probably this lense doesn't work well with dome ports. I will get this set undewater next saturday and will try with +2 dipter too...

I saw some good results with sigma FE 15mm housed by ikelite... maybe it will be my next buy.

Thanks

Fabio Amorim

It looks very much as if the port/dioptre/lens combination that is being used is resulting in a curved image at the sensor plane. This tends to show up in the closest portion of images in the corners and looks very much as though the corners are motion blurring away from the centre of the image. All you can do is to try a different dioptre (type of that is or strength) and shifting the port with extenders which may help things. You will probably find that it is only possible to optimise the combination at one focal length (the widest hopefully) as the 18~35 optics shift quite a bit during zooming, if my memory serves me right. This is a very common problem and there are numerous published examples of it out there!

I am just as mystified by the dark lower corners of the photo.  

The horizontal surface in that shot is a cross-member of a rig, which is cylindrical. The substrate in the bottom right corner is curving away from the camera, and the strobe is lighting it at a tangent...

The bottom left corner I'm not so sure about, I think it may just be strobe falloff.

Can I add to my previous post be saying that using a 'high quality' dioptre (a two element acromat) may not be as effective as using a single flat surface dioptre which tends (apparently) to induce the reverse effect in terms of curvature of the image and whilst this may produce more chromatic abberation (which may be improved by using appropriate software) it may correct corner unsharpness better. All this said, some lens/port/dioptre combinations can produce an overall optical system which simply won't operate perfectly.

Yeah. I tried to use a trick non-parabolic diopter, on the nikon 24-120, in a special ike port, and it gave all kinds of bizarre reality bending effects. Went to a plain cheap diopter, and it worked fine. Same power: +4.

Don

Re "Can I add to my previous post be saying that using a 'high quality' dioptre (a two element acromat) may not be as effective as using a single flat surface dioptre which tends (apparently) to induce the reverse effect in terms of curvature of the image and whilst this may produce more chromatic abberation (which may be improved by using appropriate software) it may correct corner unsharpness better."

That's fascinating. Anyone have any similar experiences?

I'm querying the actual theory of this with a friend who is a lens designer. If I get an understandable answer I'll post it here!

By the way, no one has mentioned - ty shooting at a large f stop, to get the most dof that you can.

What aperatures have you used?

Don

Could anyone post the pict took with Sigma 15mm FE + ikelite dome port ?

Regards

---------------------------------------------------------------------------------

EOS Elan7e + Ikelite housing+SB105 strobes and Powershot A80+Canon housing

EOS20D + waiting for aquatica housing

Canon Lenses for UW use: 20mm, 50mm+life size converter

Tokina 19-35mm

Well I seem to have solved the problem.

As previously mentioned - the 18-35 was not performing well behind the Ike dome.

I've since purchased a Nikkor 12-24 to use on my F80/N80 - yep that's right! - I got a digital lens :shock:

Remarkably, it works without a filter from 16-24 mm on the F80 film camera - making it a perfect UW lens. Also a bonus is that it will work as an 18-35 when I finally buy a DSLR to house.

I've attached the 1st photo I took on my 1st roll with the new setup.

Thanks to everyone who offered suggestions earlier in the post (sorry famorim I didn't get the 15 mm).

Nikkor 12-24 behind Ikelite Dome - on a film camera!! (F80)

Hi Anthp,

I missed this thread 1st time around.

I've posted elsewhere about concerns re: corner softness with ike dome with another zoom lens (17-40L).

It seems you had to change lenses to ressolve the problem - which is a fairly radical solution!

As far as i can ascertain it's a dome port/entrance pupil issue. I now know that the entrance pupil moves as you zoom (on my lens, anyway), so the dome is only going to be at its sharpest at one particular focal length anyway. With my lens the entrance pupil doesn't seem to be well positioned for wide angle. We've discussed elsewhere that a smaller diameter dome may be more prone to such issues.

I have not yet worked out if using a diopter makes any difference, if the lens can already focus close enough without one. I've noticed the problem both with and without the +4, but haven't tried other strengths.

Hi whitey,

Yep, it was a radical solution, but i really wanted an excuse to get the 12-24 anyway

I actually read your post about the 17-40 before i posted my question, in the hope that you might pop up and offer some solutions

As far as i can ascertain it's a dome port/entrance pupil issue. I now know that the entrance pupil moves as you zoom (on my lens, anyway), so the dome is only going to be at its sharpest at one particular focal length anyway.

One of the nice things about the 12-24 is that the entrance pupil moves very little as the lens zooms (less than 1 cm) - so this may be the reason for the consistent performance across the range. It also seems to focus close enough that a diopter is not necessary. It seems to focus about as close as the 18-35 did with a diopter.

Anyway, I'm very pleased Now I just have to take some good pictures with it.

Corner unsharpness is caused by a combination of factors such as correct alignment of the entrance pupil and centre od dome, etc. BUT by far the most significant factor seems to be that a dome port produces a CURVED virtual image at a distance of 3 x the radius of the port in front of the port (or would do if it were of negligible thickness as the theory goes). If you think about this carefully, it becomes clear that the bigger the dome the further away is the virtual image so the more that depth of field (ish - there is more to it than this) is likely to allow the curved image to be more in focus. Using a dioptre allows the lens' focussing range to be realigned so that infinity on the lens is at 3 x radius of the dome in front of the port so that the lens benefits from its maximum ability to focus. Some reverse curvature of the virtual image may be induced by a 'poor' dioptre design (ie flat rear surface) probably at the expense of other aberrations creeping in (such as chromatic) but won't otherwise help sharpen up corners. In the real world there will be some shift in the entrance pupil's position due to focus shift or zooming. Generally it is possible to optimise a zoom at a specific focal length only, which is what I did when running an S2Pro with the 12~24 - a combination which worked very well at 12mm behind a 160mm diameter port (Seacam's Wide Port) at fairly long distances (say ~1m away and further). At anything closerr than a meter I found that corner softness appeared at any wider aperture than ~f/11. The effects of image curvature can be seen in numerous published pictures and is often of little consequence if there is not much of interest in the corners. If you want to optimise on lenses of over 90 degrees view (20mm on 35mm format) then generally there will be a degree of corner softness. I am trying to carry out both calculations and practical testing to establish where the optimum settings occur (aperture/lens/port) but this is a long winded process.