114 replies to this topic

[Alex_Mustard]

I have just begun experimenting with an optical corrector port for wide angle underwater photography. My intention is to use it on my full frame SLR to give me improved image quality - especially corner sharpness when using non-fisheye wide angle lenses. The corrector was build by Carl Zeiss and adapted for my housing by Pete Ladell. 

 

Here it is on my Subal ND4S housing:

 

I thought I’d start a thread to present and share my findings as I go along.

 

In theory the adaptor is afocal and can be used to correct any lens that will fit behind it, which is an enticing prospect to be able to use many different lenses underwater, including zooms, with theoretically a much better correction than a dome port. Basically turning any lens into a corrected underwater lens. Autofocus and everything else working as normal.

 

Of course, that is just the theory. I took it to the pool yesterday to try it underwater. I am not necessarily the biggest fan on pool tests and much prefer to evaluate optics with real world subjects. But this was a massive step into the unknown - so the first aims were to check that it didn’t leak and to check it even worked at all. This corrector was designed for a camera without a viewfinder - so nobody had ever looked through it before underwater! 

 

Hooray, it seems to be keeping the water out - test shots of tiles:

 

The small port size offers possibilities for CFWA:

 

And yes, it works, the photo I was taking, above:

 

Alex

[Alex_Mustard]

May aim of going to the pool was to experiment with setup and learn about how port extensions and different internal lenses impacted on performance of the optical corrector. But at the end I did shoot some comparison tests with my best dome (9” super dome)

 

The early tests are very promising. My tentative conclusions are -

1) corner sharpness from the optical corrector port is at least 1 stop better than my best dome. So not day and night different. Not one useless the other brilliant. But the optical corrector is obviously better. Examples below. I expect to revise this one stop figure upwards with more testing and more work on optimum setup.

2) definitely less CA compared with dome. Out of interest, when the correct used out of the water the CA is bad, which I guess shows what it is taking out when it is placed underwater.

3) lens to port distance effects distortion most obviously. Too close and pincushion distortion occurs (as with a dome port). Need more work to get optimum port spacing, then might even get better optical performance

4) was able to use 20mm lens, 17-35mm @ 20mm and 16-35mm through the full zoom range of 16-35mm. All focal lengths tested worked - although I might loose the full range of the 16-35mm (at the 16mm end) when I get the final spacings figured out for best optical performance.

5) the optical corrector is also supposed to enhance depth of field. I didn’t try and measure this.

 

I did not do any tests to evaluate depth of field. But this shot with the 20mm and the optical corrector does show excellent depth of field (everything is in focus - which is not something that is easy to achieve with a full frame SLR). 

 

Alex

[Alex_Mustard]

And here are some photos of pool tiles! 

 

I shot this scene with both the Optical Corrector Port and my 9”/230mm dome port (my best dome). With the same lens (20mm) at the same apertures. I was interested to compare corner sharpness of the port setups.  

 

 

The pictures below are 1:2 (50%) crops of my original camera file, showing the bottom left corner of frame (the area is the white box in the photo above).

 

This is from the 9” dome port, shot with the aperture very open at f/4. 

 

And this is the Optical Corrector port at the same aperture (I was slightly closer to the wall of the pool, which is why the tiles are bigger, but the crop represents exactly the same area of the photo):

 

The corner sharpness is very noticeably improved. To quantify how improved, below are a series of shots at different apertures (1 stop apart) from the dome port:

 

f/5.6:

 

f/8:

 

f/11:

 

If you think the optical performance of the Corrector Port shot at f/4 is most similar to f/5.6 on the dome, then it has a 1 stop advantage. If you think it is most similar to f/8, then it has a two stop and if you think it is most similar to f/11 then it has a three stop.

 

For now, I am content seeing there is a clear advantage and I am looking forward to getting it into the ocean to put it through its paces on real subjects. I am planning some more rigorous controlled conditions testing to try and quantify what the advantage is. As I do more testing I will add the results to this thread.

 

Alex

[Tom_Kline]

There was a correction lens I (hence the II) that went along with your port. I am gussing it was just a diopter, but maybe not. Since it was series VIII size (63mm) you can't use it on most of you lenses. Are you using it without any internal correction lens?

[adamhanlon]

This is very interesting Alex!

 

It mirrors some developments that are underway with a few manufacturers.

 

Great to see that image quality (and extracting as much from our cameras as we can) is now becoming a hot topic!

 

Adam

[Alex_Mustard]

Very good points, Tom.

 

The internal correction lens is essential and you can see it in the top picture. I had it measured in the summer and it is a non-standard lens, matched specifically to the port. I expect it has opposite colour dispersion properties too - but I don't know how to measure that. But I think I should get all the specs so it could be reproduced (if needed).

 

I have tested against similarly powered standard diopter lenses and they do not offer similar performance. 

 

Fortunately, while the thread is smaller than many current lenses - the front element on most Nikon lenses doesn't fill the whole lens. For example on the 16-35m, the filter thread is 77mm, but the front element is 47mm! And the 20mm f/1.8 is similar in this respect with the front element not filling that much of the front of the lens:

 

 

One frustration are the Zeiss 18mm and 21mm lenses are both 82mm filter threads.

 

Alex

Alex

[Paul Kay]

Hi Alex

 

So what is the design type (Ivanoff corrector)?.

 

If so from my reading its probably limited to about 90 degree fov anyway (20mm on FF).

 

Or is it aspheric (if so I'd have expected it to be matched to a specific lens).

 

Or is it a compound optic (doublets)?

 

I'm interested to hear anything more (when was it designed and what was its original purpose.

 

FWIW Leica (Elcan) designed some interesting ports but all were prototype or very low volume  (military) production, and being Leica are ludicrously price collectables today). One was even incorporated into a 35mm camera housing and marked as '90 degrees' rather than the focal length. The technology is probably quite old (I have some papers showing complex correctors for any lens up to about 90 degree FOV) but mass production of a viable system has never seemed to have been economically possible.

[AlexDawson]

Hi Alex,

I assume all shots are with the focus point set in the middle? I get a feeling that i get better corner sharpness wile maintaining the center sharpness by using the outer focus points with the 16-35 and the 14-24 at wide apertures. What is your opinion on that? 

 

 kind regards // Alex Dawson

[Alex_Mustard]

Here is a video clip of me during the tests.

 

 

 

This is very interesting Alex!

 

It mirrors some developments that are underway with a few manufacturers.

 

 

I think much of it has been driven by higher resolution video, Adam. And there will hopefully be a hand-me-down benefit to us stills shooters. The distortion of fisheyes and poorly performing rectilinear lenses are very off-putting in panning shots in video - certainly not acceptable at the highest professional level. And this has driven a demand for better rectilinear wide angle optical solutions. Plus the productions that these guys work for have the budgets to get these things commissioned!

 

 

If so from my reading its probably limited to about 90 degree fov anyway (20mm on FF).

 

FWIW Leica (Elcan) designed some interesting ports but all were prototype or very low volume  (military) production, and being Leica are ludicrously price collectables today). 

 

Hi Paul, 

 

The widest lens I have tried so far is the Nikon 16-35mm @ 16mm. I have seen this quoted as 107˚ field of view. This is the shot:

 

The system should be afocal and therefore should work with any lens that I put behind it, as long as it does not cut off. But I am particularly interested in zooms and some of the new wider aperture wide angles: Sigma 20mm f/1.8 and 28mm f/1.8 macro etc.

 

I fear that much of Ernst Leitz knowledge has been lost, I know E.L. Canada are still running, but I think it is a long time since they have done any serious underwater stuff. I really hope that knowledge can be unlocked again.

 

Alex 

 

 

[Alex_Mustard]

Hi Alex,

I assume all shots are with the focus point set in the middle? I get a feeling that i get better corner sharpness wile maintaining the center sharpness by using the outer focus points with the 16-35 and the 14-24 at wide apertures. What is your opinion on that? 

 

 kind regards // Alex Dawson

 

These tests were all done with single point focus, yes. That is not the mode I use in the field. I use AF-C Auto Area - as this does the same as you suggest and looks for the closest focus point, which given the field curvature with a dome is away from the middle of the frame. 

In the 1970s BSoUP members in the UK used a re-focus technique when shooting with manual focus lenses behind domes, in that they would focus on the subject and then adjust the focus slightly closer for the same reason to help improve corner sharpness. 

 

When Nikon introduced the D3, the first camera with lens specific tuneable autofocus, I tried setting the offset as far in front of the subject as I could with my wide rectilinear lens to see if it had any effect. I noted in my D3 review here on Wetpixel that it did not. I think the fine tune was just too fine. 

 

Alex

[Tom_Kline]

The internal correction lens is essential and you can see it in the top picture. I had it measured in the summer and it is a non-standard lens, matched specifically to the port. I expect it has opposite colour dispersion properties too - but I don't know how to measure that. But I think I should get all the specs so it could be reproduced (if needed).

 

One frustration are the Zeiss 18mm and 21mm lenses are both 82mm filter threads.

 

 

Sounds like the internal lens was designed to screw directly onto the Compur type 38mm Biogon lens so ignore the series VIII bit I wrote above. 67mm filters sort of will fit the mount on the lens but have a different pitch so it is important to only do a partial twist of a filter (probably not advisable for an underwater application). The male threads on your internal correction lens, however, are more problematic. There are Hasselblad series 63 filter adapters that may use the same series VIII retaining rings as the lenses and so would have the same thread. I seem to recall this for the adapter to Bayonet 50 but I have not handled one in decades. One could be cannibalized....

 

The Sigma 20/1.8 also uses 82mm filters but focuses so close I have not had to use diopters - I have had good luck using this lens for both APSC and FF and the smaller Seacam wideport.

 

Tom

[vaiski79]

Very interesting, can't wait to see what kind of price point this will be.

[tobyone]

This type of correction lens is discussed in Lawrence Mertens book "IN-water photography" published in 1970. By this time Ivanoff had a corrected wide angle lens"I think in a nikonos mount" The book  goes deep into the physics of the port science.  It is one of the definitive books on underwater photography physics. It scan be found used on amazon.

[Alex_Mustard]

Yes, I am a big fan. I agree that Mertens book is an interesting one - not least because he takes such an original approach compared with other authors - his refusal to call underwater photography "underwater photography" and instead insist on calling it 'in-water photography" is a wonderful example of this and his pedantry! 

 

His writings on corrector ports are one of the best. Certainly better than anything I have from Alex and Dimitry. But like many others he lacks practical knowledge of these systems - and I don't think he ever tried one. So with just one day with the port I have already found that one of his main conclusions does not hold up in the real world.

 

But his writing has been very valuable. And is guiding the development work I am doing before I take the port to Grand Cayman in January.

 

Alex

[RogerC]

This type of correction lens is discussed in Lawrence Mertens book "IN-water photography" ....   It scan be found used on amazon.

 

Also from half dot com.

 

http://product.half....79&cpid=3514334

[Larry C]

I got mine from Alibris for $5.10 including shipping.

[Alex_Mustard]

There was also a copy in the British National Oceanographic Library - which was the one that used to distract me from work!

I am pleased that this experiment is generating lots of interest. 

 

Thanks, Adam for posting it on the front page. I must absolutely stress this is a work in progress. In fact, I am very early in the process, having only done one day of testing. I have commissioned a few more optical parts to test. At present I am having them made in acrylic - and when I see how they perform I will have them made up in glass. 

 

As I mentioned above, for the first time for a few years there seems more general interest in this topic - normally it is just me and Pete Atkinson moaning to each other! I think a lot is driven by filmmakers shooting at higher resolutions and become more demanding about image quality. Despite the higher resolution of stills, moving pictures make poor optics more obvious, particularly when the camera pans. 

Most people are looking at adapting water contact Nikonos optics. Which have advantages and disadvantages

 

 

I have received a number of emails and phone calls about the optical corrector - and interesting ground has been covered in them, which is worth summarising:

 

The system I am working on is different from the water contact lenses that people have been adapting to digital cameras, particularly ones from the Nikonos V and Nikonos RS. Those lenses were designed to work specifically underwater with water contact optics. The system above is one that aims to completely correct the water air interface so that any lens performs exactly as it does in air. The downsides are that it is more bulky and can’t be used with really wide angle lenses - I am yet to determine a wide limit through cut off. 

It certainly can’t be used with fisheyes. Furthermore the system above consists of two parts, and the simpler, rear part must be attached to the camera’s lens. This further limits lens choice to optics that can accept  them, the Nikon 14-24mm, for example, cannot take a supplementary lens. Zooms that change barrel length are also out. 

 

The advantage is that many lenses can be used, including new and forthcoming optics. Early tests show I can use lenses that are wider and focus closer than rectilinear Nikonos lenses. I also have AF-S autofocus, etc etc. 

The corrector should also allow me to shoot open aperture wide angle, for both technical reasons (it is really dark!) and artistic effect (shallow depth of field wide angle, without the problems of simply seeing the blur from virtual image created by a dome port, which is what happens if I try that now). Perhaps also to be able to realise the advantages of PC lenses properly underwater etc. Getting it working is really the doorway into a world of underwater image experimenting.

 

While the lens can still focus on land, it cannot be used out of the water because optically the image is stretched and suffers strong optical aberrations (the opposite of what is actually being corrected). 

 

Alex

[StephenFrink]

Alex - how does the field of view compare to dome, all else being equal? Size of tiles in comparison shots suggests to me the corrected port may reduce FOV. I know you said distances were variable, but I'd be surprised not to see some FOV compression with your test port. Depth of field and corner performance look promising though. Good luck with your project.

[Alex_Mustard]

That was a concern of mine before testing. As I feared that a flat front element could only narrow the angle of view. What I have learned is that he rest of the optics actually increase the angle of view to compensate - so that if you shoot them port in air it stretches the angle of view of the lens behind it (and creates CA), all of which disappears in the water and angle of view is maintained. 

 

From my tests I think that the field of view seems the same, although I have not done a controlled, tripod based test, yet to confirm.

 

The tests I posted above the Optical Port tests I happened to be slightly closer to the wall of the pool than with the Dome. Which makes it look like the view is restricted. I actually shot two series of tests. And by coincidence in the others (with the test charts) that I posted on my website, it was the other way round an the Dome was closer, so it looks like the optical port is giving a wider view.

 

I decided to post these ones here. As I thought if I shared the test chart ones, where the dome was closer, than people would say that the dome performance was slightly compromised by being closer to the wall.

 

This link will allow you to view this comparison of the Optical Corrector (left) and Dome (right) as a screen grab of them both in Lightroom from the other tests, where the dome was closer. This was slightly better controlled as I had a mark on the bottom of the pool (and ankle weight visible in the photo at the top of the page) to help maintain my position: 

 

So my feeling is that angle of view is maintained. And the impression to the contrary, in the images above, comes from my inconsistencies in testing. To be honest, the first pool session was to see if the system even worked and didn't leak. I only put the dome in at the last minute! 

 

I already have a new acrylic optic to test with the port (a Christmas present) and have decided to go ahead and order a glass one to a slightly revised spec, hopeful that it will be ready for testing in Grand Cayman in January.

 

Alex

[StephenFrink]

If you look at the very small type on the lower left of the res chart it reveals the optical corrector being significantly sharper. This is not corner noodling, as this would be an important part of most any composition. Looks like you are on to something, Alex. Great initiative and a promising direction.

Edited by StephenFrink, 29 December 2014 - 06:10 PM.