How to get sharper corners when using E-mount 12mm Zeiss in Ikelite housing and 6 inch dome?

Firstly, why are you worried so much about the corners. Your photo looks fine to me and unless you are pixel peeping why worry. Secondly you can shoot at smaller apertures (like f:16) but most likely won't help too much. Getting a bigger dome will help but at a cost and pain for travel factor. Using a diopter on the lens can often help and is quite inexpensive compared to a new dome. One thing to check on is the corner sharpness of the lens on land; test it in the garden; it can't be much better UW than it is on land.

Bill

Thanks Bill for the input. I don't think I am pixel peeping, but its quite clear that the corners are substantially less sharp. Just trying to improve the photos I get No, on land the corners are less sharp but very marginal, that is why I got this lens as it has pretty decent corner sharpness.

What aperture do you usually shoot at for your wide angle shots, and what size dome is it?

I have a similar issue with my sony a5100 and the 16mm F2.8 + Fisheye adapter. The photos are hardly useable at F2.8. So far I stop down to to F11 to get the best corner sharpness, but even then there's still distortion. I am using a small nauticam 4.33" dome though. Also I'm kinda new to underwater photos, but that's what I've done so far.

First of all - your shot is a great example of the classic problems of wide angle underwater. Solving these problems is something many have to go through. Please can I use your photo as an example in my presentation on this topic?

The cheapest fix you have to improve the situation is to close the aperture. You should try and be at least f/11 on an APS-C camera. On full frame that needs to be f/13-f/14. Take a series of shots at f/8, f/11 and f/16 on your next dive - and you will see how the corners get better.

Next up it is important that the dome is correctly positioned. Just because the lens fits behind the dome - does not mean it is correctly placed. This means using the correct port extension. Manufacturers should know this and give you the correct advise at purchase - but we so many many new lenses appearing - their advice is not always right. A poorly positioned dome port will ruin your image quality.

(A dome should be positioned so that the centre of the sphere that would be made by the dome is in the same place as the lens's nodal point (the point in the lens where the light rays converge) - however, in practice it is not always easy to find out the exact position of either! So we tend to learn by testing and then sharing what port extension works best).

A larger (more expensive) dome port will help too. But only if it is correctly positioned. A larger dome in the wrong position will be worse than a smaller dome in the correct position.

Alex

First of all - your shot is a great example of the classic problems of wide angle underwater. Solving these problems is something many have to go through. Please can I use your photo as an example in my presentation on this topic?

 

The cheapest fix you have to improve the situation is to close the aperture. You should try and be at least f/11 on an APS-C camera. On full frame that needs to be f/13-f/14. Take a series of shots at f/8, f/11 and f/16 on your next dive - and you will see how the corners get better.

 

Next up it is important that the dome is correctly positioned. Just because the lens fits behind the dome - does not mean it is correctly placed. This means using the correct port extension. Manufacturers should know this and give you the correct advise at purchase - but we so many many new lenses appearing - their advice is not always right. A poorly positioned dome port will ruin your image quality.

 

(A dome should be positioned so that the centre of the sphere that would be made by the dome is in the same place as the lens's nodal point (the point in the lens where the light rays converge) - however, in practice it is not always easy to find out the exact position of either! So we tend to learn by testing and then sharing what port extension works best).

 

A larger (more expensive) dome port will help too. But only if it is correctly positioned. A larger dome in the wrong position will be worse than a smaller dome in the correct position.

 

Alex

Thank you Alex, will try this on a dive/in a pool and test with different apertures as suggested, and see how corner sharpness varies. Good to also know that f8.0 is probably not stopped down enough! I guess with the APSC camera, if I stop down to f12 and beyond, I will probably end up having to bump up the ISO to above the ISO 400 that I am shooting now.

On dome positioning, it is one of the recommended domes in ikelite's port chart, but maybe I can email them to see if they can give some input on this. Also good to know that buying a bigger dome is probably not the first thing I should do. Money saved for now I guess!

And of course, please feel free to use the photo as an example in any of your presentations - I can even send you the high resolution JPG, just let me know if you want it. And can I say, its a very pleasant surprise to get a reply from you, I am a big fan and was just reading your "underwater photography masterclass" book at home yesterday! Its been a big inspiration to go out and try to take better photos!

Thank you! The resolution here is perfect for the talk.

Do you want me to credit you when I show it, or not? It is a great examples of the classic problems that many people find.

Alex

Can I also add a comment please? Ideally you should also use a modified version of what is called the 'hyperfocal setting'. On land this would be the point of focus which ensures that infinity remains in focus ('sharp') whilst maximising the amount of image in focus. As we rarely need infinity in focus underwater, in essence we need to try to focus on a point which ensures that we maximise depth of field and this may or may not be the central subject of your image. So what we are trying to do is to ensure that the furthest point in the photograph which needs to be 'sharp', is 'sharp', and to do so we need to adjust the point at which we focus to optimise the depth of field, so that this is the situation. To some extent selecting an appropriate point on which to focus comes with experience.

We tend to forget all this in our current days of pixel peeping where we demand absolute clarity of the point of focus, but in fact depending on how an image is to be used depth of field is still as it always was; the acceptably sharp parts of the image given the final output requirement of the image.

If this all sounds complex it is not, it just needs getting your head around and there is plenty of information on it on the web.

Thank you! The resolution here is perfect for the talk.

 

Do you want me to credit you when I show it, or not? It is a great examples of the classic problems that many people find.

 

Alex

Haha no need to credit me for it. Maybe one day when I have a great photo that you want to use for a show then i'll be keen

Can I also add a comment please? Ideally you should also use a modified version of what is called the 'hyperfocal setting'. On land this would be the point of focus which ensures that infinity remains in focus ('sharp') whilst maximising the amount of image in focus. As we rarely need infinity in focus underwater, in essence we need to try to focus on a point which ensures that we maximise depth of field and this may or may not be the central subject of your image. So what we are trying to do is to ensure that the furthest point in the photograph which needs to be 'sharp', is 'sharp', and to do so we need to adjust the point at which we focus to optimise the depth of field, so that this is the situation. To some extent selecting an appropriate point on which to focus comes with experience.

 

We tend to forget all this in our current days of pixel peeping where we demand absolute clarity of the point of focus, but in fact depending on how an image is to be used depth of field is still as it always was; the acceptably sharp parts of the image given the final output requirement of the image.

 

If this all sounds complex it is not, it just needs getting your head around and there is plenty of information on it on the web.

Thanks Paul for you input. So if I get this right you're saying maybe in my case, my point of focus should have been maybe something nearer, so that everything "near" would have been more in focus? Yeah I think the concept makes sense. But for this particular picture, it seems that the central object is kind of in the same focal plane as everything else. Will keep this in mind though

Yes. What you are trying to do is to extend the depth of field in front of the central subject as far as you can whilst still keeping the subject sharp too. So you can either focus on the subject which extends the depth of field roughly twice as much behind the subject as in front, or you can focus slightly in front of the subject to move the depth of field towards you, which will help with corners where they are closer to you (they often are). As with much in photography, its a compromise and it all depends on what's 'acceptably' sharp for however you intend to finally view the image. Today we often look at 100% on screen which shows up any discrepancy. In reality we should probably print and view at a reasonably viewing distance which is the idea on which depth of field scale were originally based.

If you have a static subject like the one posted, its worth 'focus bracketing' incrementally closer and then seeing if this is acceptable to your needs.

The Zeiss 12mm lens is absolutely fantastic underwater. It's known to be sharper than FF lens from both Canon and Nikon. Being a rectilinear design, corner sharpness is always a topic of conversation with most photographers suggesting a larger dome, plus some of the suggestions listed in this thread. I have a modified Zeiss 12 where I removed the bell shaped shade supporting structure and replaced it with 67mm filter rings (the lens doesn't need the shade structure to function) . This has allowed me to use the lens in both a Nauticam 4.33 and 7" single piece domes. For travel purposes, I wanted the 4.33 to work since it's tiny and well suited for CFWA photography. I bought the much larger 7"dome (yet small by some standards) that was designed for the Sony 10-18 WA to test and make comparisons. . Given the 7" is not a fisheye dome (like the 4.33), I expected it to have much sharper corners. In testing, it did have sharper corners, but not by as much as I expected. In fact, to my surprise, the 4.33 dome is almost as sharp and much, much better for CFWA. I can actually focus on objects very close, within an inch or two from the face of the dome! Obviously, not so with the much larger 7" dome.

I ended up using the 4.33 for travel purposes and for CFWA. To get even more use out of the dome, I actually configured a 67mm Nauticam swing mount (from my old port 72) to fit across the face of the dome. I use this configuration for my Sony 18-55mm and can swing a CMC-1 15+ close-up lens at the 55mm end for some great super-macro or 18mm WA on a single dive. To get the correct dome placement for the Z12 that Alex is referring to, I added a 20mm extension to the 4.33. This was also needed for the Sony 18-55 longer length at 55. It seemed to put the Z12 in the ideal spot for decent corner sharpness. I usually shoot f11-16, sometimes more, but too much and you end up with soft shoots from aperture refraction.

I've considered at some point of creating a graduated sharpening filter to match the 4.33 optics in Photoshop that will gradually sharpen the corners very slightly for all shots with the Z12 and and S18-55. The shots below are an untouched jpeg's made from a RAW file on my NEX7 and Z12. The lens is sharp and the corners look pretty good for a small 4.33" FE dome.

Looks like my photos didn't post? I need to study how to upload media to the site to use it? Hope the info helped though.

To expand upon what's been said about getting wide angle lenses to work properly, the issue is getting the corners sharp is twofold, first of all there is a depth of field issue, the virtual image is located about 3 dome radii out at infinity and gets closer as you focus closer and it is also curved at the same radius at infinity. So the corners are closer to the lens and the DOF needs to be sufficient to cover all of the curved image. Stopping down of course serves to bring that the corners into focus.

In addition the lens' principal plane needs to be at the centre of curvature of the dome.. This is about symmetry, if it is at the centre all light rays pass through the dome at right angles and are not refracted. If it is not in the right spot the rays must pass through at an angle and are refracted and this gets progressively worse towards the edge, the refraction introduces chromatic aberration and other issues. Stopping down is less likely to fix this. This online calculator shows the concept: https://oceanity.com.au/tools/dome-port-positioning/ Move the top slider to move the principal plane in and out from the centre and the bottom slider to how refraction varies from centre to edge of field.

Looking at the port charts it seems like the ikelite port is a one size fits all, for example the 8" dome is used with the same adapter for many lenses. Particularly the news that the lens performs so well in a small dome on another system shows a bigger dome is not your answer.