77 replies to this topic

[bmyates]

...If you add a diopter (+2 for a Superdome I would suggest) then adding 10mm of extra extension tube would be a reasonable starting point. . .Perhaps of higher benefit to more underwater photographers will be the inevitably suppressed prices of used MkI lenses!!!

I thought one of the benefits of the Superdome was not needing to use diopters with zoom lenses. Can you explain why a +2 plus ext. ring would improve the image vs. simply finding the "optimal" extension ring for the port?

As for depressing the prices of used MkI lenses, I can attest to that! I've got mine for sale for $850, barely over half what I paid for it!

[Paul Kay]

A diopter has two functions when used underwater behind a dome port:

First, it allows the lens in use to see the virtual image (created by using a concentric dome port underwater) of an underwater subject nominally at infinity (Ok well 10m ish or so is practical!) when the lens is set to infinity. In essence the diopter resets the lens's focus so that when it is set to infinity it is actually focussed on the virtual image of a subject at infinity. The distance of this virtual image of an underwater subject at infinity from the front of the dome is 3 x the radius of the dome (3R). The way the required dopter is calculated is from a simple little formula:

P = 1000/4R where P is the diopter required (usually +1, +2, +3, or +4) and R is the radius of the dome being used (in mm)

For a Seacam Superdome of 240mm diameter R is 120mm so P = 1000/4x120 = 1000/480 = 2.0833 etc which is near enough to +2. This said, the superdome is of large enough diameter so that the the maximium focus point needed is 3R infront of the dome or 3 x 120mm = 360mm - and this is in front of the dome so at least another 120mm (to the entrance pupil position at the centre of the dome can be added plus the distance from the entrance pupil to the image sensor position, say ~100mm) can be added to this giving around 580mm - so you don't actually have to use a diopter as many lenses will focus close enough for most subject without one, although not using one may limit the closest focus available, which brings me to:

Second, diopters to cause some image curvature because they are a simple (or most are) lenses. Fortunately, this curvature is apparently in the reverse direction to that created by the dome so it may have some positive effect in helping to minimise this curvature (although in all honesty its not going to be much and whatever it does do will depend on both the diopter (design) and lens it is used with - most diopters are designed to be optimised with short telephoto lenses rather than wide-angles).

The problem you may face is that using a diopter means that you can't shoot splits or anything out of water should you want to do so with the camera still in the housing. I am looking into various possibilities to help resolve image curvature but even if they work, they will only do so in very specific situations. A non-concentric dome might be an answer but I have a suspicion that these would have to be lens specific - can't see this being economic at all!

Yes this is all theory, but given the price of good wide-angle lenses and wide zooms, it makes sense to wring the full quality available out of them. Its a bit like earlier today. I shot some landscapes in glorious spring sunlight. As a matter of course I used a sturdy (and far too heavy) tripod, self timer and mirror lock to minimise vibration - is the irrelevant or best practice based on sound theory? Should we apply the same standards to underwater imagery? I don't like fisheyes much either!!!

[dhaas]

James (and Bruce!)

I hear you on using a rectilinear lens versus a fisheye to minimize curvature. Always happy to help out LOL......

But since you have a full frame sensor dSLR, I don't understand why you just don't shoot your 24mm F1.4 which should be great without having, as Alex calls it, much of a "Tadpole" effect. Bruce, dont you have one of those Canon 24mm F1.4 lenses, too? Just don't get too close to your models, as in your latest shoot, and don't aim extremely up or down....

In reading Pauls information, I kind of understand it, but have to confess it also makes my head hurt As in, can anyone really tell in looking at photos where a lot of this is going to make any difference in the original digital capture?

Paul and I DO agree on one thing.....I HATE putting diopters on ANY wide angle lens, fixed focal length or zoom. It just cramps my style of being able to shoot above and below at any moment. I don't care about no stinkin' water droplets on the dome above water, either, as that can be cleaned up later on the computer!

"And that's all I have to say about that" - Quote by Forrest Gump

dhaas


P.S. - Even the Tokina 10-17mm Fisheye, for cropped sensor dSLR cameras only isn't as "fishy" when zoomed in. I'm just starting to experiment with this end of the range (See photo taken at 17mm, with very slight curvature.)

[caminu]

the discussion gets too technical for me to understand. i think i need some time to digest all the theories behind

hopefully Stephen will come up with an optimal setting for this lens soon, so that I can just apply it without any hassle.

[bmyates]

First of all, thanks Paul for taking time to explain all that (although, like David, it kinda makes my old synapses overheat trying to understand it! )! That explanation DOES help, and in a situation where I was fairly sure I wouldn't want to do a split over/under, a diopter might be worth the trouble and cost.

In most cases, however, I'd probably be more likely to get as much as I could from the lens (i.e., after Stephen Frink has declared a certain ext. ring as being best for it in a Superdome), and if the edges are still soft, just crop them out. That's what I do now, and it's one of the advantages of having all these "extra" pixels on my 1Ds MkII; I can crop the daylights out of a pic and still have 8 or 9 megapixels left.

David wrote:
> Bruce, dont you have one of those Canon 24mm F1.4 lenses, too? Just don't get too close to your models, as in your latest shoot, and don't aim extremely up or down....<

Yes, like James, I do have a 24mm f/1.4 (that's what he and I were "pool testing" in the Bahamas in the photo I attached earlier in this thread), and I actually use it quite a bit. I actually find that in wide angle "situations" (e.g., big animals, plush reefs, etc.), I rotate between 3 lenses: 16-35mm, 15mm FE, and that 24mm. But your suggestion not to get too close to my "models" (which are usually non-human, so don't always cooperate) is not necessarily the most desirable option. In fact, it is contrary to one of the cardinal rules of uw photography (get as close as you possibly can to reduce the amount of water between you and your subject!). In general, I strive to get as close as I possibly can, and then use either a 15mm fisheye or the 16-35mm at the wide end. (I can't believe how many of my shots I look at and say, "Crap! I should have cut the distance between me and the subject in HALF!)

OTOH, since consensus seems to be that shooting the 16-35mm at 16mm is where the most "problems" are, and zooming in a little (say to around 24mm) yields better results, a person (especially one named Dave) might make the argument that it makes more sense to just use the 24mm f/1.4 for even those "close" shots. :glare:

But that gets back to the primary reason for using a zoom lens in the first place -- the flexibility to shoot a given subject/scene at the best focal length for THAT subject on THAT dive, rather than being forced to shoot it at a specific single focal length of a lens you happen to have on the camera (e.g., 15mm FE, 24mm, etc.) that dive. I like using the 16-35mm lens specifically because it lets me zoom in to 35mm for one shot, and then zoom out to 16mm when a huge beast swims right up to me. That's why I (and presumably so many others) occupy so many brain cells trying to get a lens like this to work as well as it possibly can uw!

[Paul Kay]

I know advertising photographers who use the 24/1.4 as they shoot models in pools and need:

1. Optical quality to be as high as possible as this is demanded by their agencies and clients (so use 24 + Superdome), and

2. Minimal distortion - fisheyes would be unacceptable

I also use the 24/1.4 although I tend to use it behind the Seacam's Wideport due to weight constraints when travelling and conditions constraints when diving on our west coasts. It is my 'standard' wideangle above (I used it today, wide-open on a shoot) and below water and I'm probably about to buy another (used and cheap) for my second housing. In poorer light conditions it also has the advantage of providing the brightest view available. The spider crab shot (left) was taken using a 24/1.4 with Canon's Achromatic close-up lens and does show some (hopefully acceptable) softness in the corners as the subject was very close (10~15cm maybe, I can't remember now) to the port where everything does finally go pear shaped.

I really rate and recommend this lens - you get what you pay for I'm afraid.

[Iggy]

Fisheye lenses with inherently sharper corners will outperform any rectilinear zoom, unless you are prepared to spend countless hours testing extension lengths and diopter / no-diopter combinations. And even then, you won't get edge to edge sharpness at all focal length settings....

Some people really enjoy investigating and sharing these kinds of things, and I enjoy benefiting from their industry. Except for specific circumstances or subjects I almost always like a good rectilinear setup over a fisheye. And since I have neither the time nor the place to do these tests myself, the people who "spend countless hours testing" are the people who allow me the latitude, in this case, of choosing FE or WA rather than simply defaulting to FE. And that makes my shooting more productive

[dhaas]

Iggy (and others),

I hear you on investigating optimum set ups through testing. I didn't mean to imply it isn't appreciated for those who enjoy it, or simply feel the need to do it.....I just think sometimes we lose sight of what all this technology has given us.

Photographic freedom........

Cameras, lenses, sensors, flashes, housings, etc. all literally developed in the last 5-6 years have provided us divers with a quantum leap in imaging capability. So I constantly ask myself (and fellow UW shooters) how GOOD is good enough?

I have always stated here (and on other forums) that a properly exposed Fine / Large JPEG or RAW file from even a modest cost dSLR when post processed can produce any image format 99.9999% of shooters could ever envision. From published images, to huge prints, even display murals for trade shows, etc.

Consequently, I think a lot of discussion is simply for discussing and debating specs, oohing and aahing over a miniscule product improvement. And that's fine since it's like sports or politics endlessly presented on TV and radio talk shows, on the web, discussed in bars, etc.......I just don't agree that a lot of this will make a discernable difference in pictures many (if not most of US) will create underwater. That said, certain things like iTTL and eTTL2 plus faster AF on cameras certainly helps us these days......

It is fun, and I often think of how in the "old days" i.e. shooting FILM, that this sharing didn't and couldn't have taken place....So I guess that's another advantage of forums.....

I never mean to post just to be argumentative.....Only sharing one lowly soul's opinion So keep testing and sharing.....Then go out and shoot maximizing the capability....That's when the testing will finally show its' value!

dhaas

[james]

Dave,

Think about this analagous scenario:

You are getting married next month.

You want to hire a wedding photographer.

You want a set of photos delivered to you that will last a lifetime.

Do you hire the photographer with the digital Rebel and one or two lenses because his camera gear is "good enough?"

Or do you hire the photographer w/ the high-end setup?

Many divers just don't get to make that many dive trips every year, so they want to take the best photos they can possibly get - to last them a lifetime. So these little details can be pretty important. Not everyone is like this of course as fortuntaly everyone has different priorities and viewpoints.

Cheers
James

[bmyates]

...I think a lot of discussion is simply for discussing and debating specs, oohing and aahing over a miniscule product improvement...

Don't forget the importance of being able to justify (rationalize?) spending obscene amounts of money so you can have the thrill of receiving a UPS package and unwrapping a shiny new piece of "the latest" gear (my new 16-35 II arrived yesterday)! B)

[dhaas]

Hey James !

I hear you on trying to maximize one's results when that precious dive time comes around each year.....I've been there myself many times. And what I've found is that simplifying one's rigs, number of lenses, etc. usually (at least for me) has produced pics I'm more satisfied with, and more consistent. Alex posted something recently about his own self imposed rule of getting one lens at time, and learning how to use it best through trial and error...Although in Alex's case, it's likely trial and immediate success

As I said, I don't post to be argumentative, just to add a different view! Serious, I mean SERIOUS photographers in photojournalism and art fields use equipment many hobbyists would scoff at....That's all I'm saying when some of the discussions get into details that will likely never be apparent in viewing a photograph.....That's all I was trying to say.

Viewing a wedding photographer's portfolio and sample prints, etc. is where one chooses (or should) select a photog for such a lifetime event Maybe some folks need to know what he / she is shooting with. I'd want to see if the photog can not just purchase equipment, but produce images.....And you're right, having certain lenses, etc. does make certain types of photographs easier to produce.....

But I still see people shooting with Oly C5050, Fuji E900, Canon A640 (my current favorite) and other rigs producing astounding UW pics. So this is someting I wrestle with myself all the time......

As I've always said, you can't buy a good eye

Later,

dhaas

P.S. - YES, I can relate to Bruce's current orgasmic delight of unwrapping a shiny box labeled "Canon" Whoo Hoo!!!

[bmyates]

As I've always said, you can't buy a good eye

Too bad...this guy could use one. (Sorry - off-topic and inappropriate!)

Attached Images

• 

[dhaas]

I'm cringing looking at that photo, even though I know it possible!

My wife's niece when younger somehow had a puffed rice piece from breakfast go up her nose, only to re-appear a day or so later through her eyelid.......

Eeeewwwwwwww......

dhaas

[Paul Kay]

I can think of many applications where soft corners could be a serious issue - photomosaicing for scientific/archaelogical use, etc. Underwater photographers come in many flavours - not all are interested in pretty pictures and underwater photography as an end in itself. Ultra-wides and soft corneras are the current Achilles heel of underwater photography and whilst stopping a lens down with a port like a Superdome and diopter will give very good results, shooting with fast apertures and smaller ports is still a no-no if ypu want decent image quality.

I too prerf to limit my lenses and choice, although I have (too) many I tend to use 24/1.4 and 100/2.8 macro underwater for 99% of images and fixed 24/35/50/85 and 135Ls above water (and I do use ALL of these).

I've gone of puffed rice....

[Alex_Mustard]

I always use dioptres with rectilinear wide angles lens to improve corner sharpness. I find it makes a noticeable improvement to image quality.

Although I did say I would no longer debate the issue on Wetpixel. It is an impossible argument to win online because the differences are small (both work) - people have different standards of what is acceptable to them, and it is very hard to get decent test shots because there are so many variables.

But I do believe that they make a difference because they help to curve the plane of focus of the lens to match that of the dome port.

My big rant is against what I believe was damaging mis-information about the lack of need for dioptres for domes that has been circulating the community for the last couple of years. In my opinion it all started when one manufacturer said our new dome is so good you don't need a dioptre. This was a good marketing statement. And the other manufacturers/retailers quickly responded saying the same thing - knowing that the lenses all worked behind their domes without dioptres. Nobody wanted their domes to look inferior, so they started saying there was no need for dioptres too. Soon everyone was claiming you didn't need a dioptre.

And to some extent they are right. The lenses all work without dioptres. But my lenses work better a dioptre. And so should all lenses in theory. Read an underwater optics book and the therory of a dome port (or dome porthole as they are often traditionally known in such texts) includes two optical elements - the dome and the dioptre.

The argument about new domes should have been that bigger domes reduce the strength of dioptre required. And therefore lessen the need for a dioptre - but don't remove it. Paul showed above that theoretically even the awesome Seacam superdome theoretically needs a +2. Which ties up with much of what Stephen Frink found in the great tests he published on the front page of Wetpixel a few months ago (I can't find the link). Big domes produce great image quality - but they do require a dioptre (in theory and practice).

All that said, the difference between using a dioptre and not using a dioptre is small. And there are plenty of great real world photos out their taken both ways. For me I always use one. But since I can't really prove it, I don't try and persuade people otherwise. Except in this post!

Anyway, that is my opinionate slant on modern dome port theory with rectilinear lenses. I never use a dioptre with my fisheyes.

Alex

[StephenFrink]

the discussion gets too technical for me to understand. i think i need some time to digest all the theories behind

hopefully Stephen will come up with an optimal setting for this lens soon, so that I can just apply it without any hassle.

Just a quick schedule update. My Canon rep only sent the new 16-35 yesterday, which means I'll get it on Monday (no Saturday Fed Ex in the Florida Keys). I'm out of town next week so my testing will be delayed a week now, until the second week of May. Sorry. It will be a high priority once I return.

[bmyates]

I always use dioptres with rectilinear wide angles lens to improve corner sharpness. I find it makes a noticeable improvement to image quality...

Alex (and/or Paul, or others),

So you use a diopter (dioptre on that side of the pond) with all WA lenses in dome ports (except FE's), whether primes or zooms?

What (if anything) is the downside of adding a diopter, i.e., do you sacrifice anything by using one?

I haven't used a diopter with a zoom lens since my old Minolta SLR days, when I had an Ike housing and they said I needed one (+4 as I recall) in order for the lens to focus close enough to focus on the virtual image created by the dome uw. But if there's no real downside, I'm not sure why everyone wouldn't get a diopter for every WA/zoom lens they own.

[craig]

Like any other element in the optical path, a diopter can contribute to image degradation. Diopters introduce some CA though on wide angle it isn't very significant. Diopters interacted with my Nikon 12-24 to produce odd lens distortions and, of course, there's the issue with doing over-unders. Overall though, I agree with Alex that diopters shouldn't be feared when using dome ports and it doesn't make sense to me to waste the majority of a lens's focus range by not using one. When housing a new lens it seems reasonable to do a few dives in a pool to verify that the system works properly. Most people would check a new lens out above water after all.

[Walt Stearns]

Getting perfect, or at least good corner sharpness with extreme rectilinear wides (especially the zooms) in the full frame digital area is like trying to catch a unicorn.

Here is a link from someone who has performed some tests on Canon’s new 16-35 Series II L http://www.16-9.net/lens_tests/. The results are interesting but not compelling enough to make me go out and try one. Seeing my Subal FE 2 dome port stuck out there on the end of a 60mm extension ring for 17-40 is enough to make me cringe as it is. Pushing it another 10mm further out for the newer 16-35 II would start making my housing look more like a blunderbuss (http://hegewisch.net...s/blunbuss.html).

Diopters, been there, done that, even back in the day of film. Found then, still see it now, as short fix through compromise.

To change tact, we all know that for extreme wideangle coverage underwater, a fisheye will beat super wide rectilinear in the middle to low aperture range day in, day out. And its not just a full frame issue either. Even lenses like 12-24 at 12mm on crop sensors get (maybe not as apparent) soft corners too.

Maybe the zoom lens we’re looking for is a semi-fisheye 16-28. It will likely have a little barrel distortion, but the corner detail would hypothetically be on par with your fisheye. Unfortunately, we are too small a market for a lens manufacture to even consider doing something for our cause.

Hey, it my opinion and I am sticking to it.

Until then I’ll keep on using what I got till something really does come along that will do it better.

Canon 5D, 17-40, 15 fisheye with 1.4 teleconverter, 70mm Macro.

[Paul Kay]

Just to clear one thing up. Soft corners are not a function of sensor size/format. They are to do with the lens's field of view and become noticeable when this exceeds about 90 degrees (~20mm lens on full frame).

Although it has been said elsewhere on wetpixel before this affects rectilinear lenses but NOT fisheye lenses as rectilinear are designed to focus on a flat field whilst fisheyes focus on a curved field which is what a dome port produces. But not everyone (myself included) like fisheye views and de-fishing is as unsatisfactory a solution as soft corners!

The 16~35/2.8L (I or II) will suffer from some degree of corner softening from about 16~20mm as will any other wide-angles (17~40/4L or other fixed focals or zooms) in this focal length range. The amount will depend on the subject distance (most noticable softness occurs in subjects with a spread of subject distances including close foreground detail), aperture being used (the smaller the aperture, the greater the depth of field and the more of the curved virtual image which is sharp), focal length set (the longer the better), dome port used (its size - larger is good up to a point - I suspect that Seacam's Superdome is probably about the optimal largest size - and its position relative to the lens) and the diopter being used (type, power, etc) or not. That's a lot of variables! Shooting appropriately set-up and with carefully selected subjects and the result should be high quality images (even in the corners), but not sorting everything out precisely will lead to unpleasant corner blur, as ever with ultra-wides.

There is of course the other issue of how the MkI and MkII lenses project onto the sensor corners which might have an effect on land but will almost certainly not have as much effect underwater as the issues I mention above. This is simply because the image is degraded before it is projected onto the sensor.

Lastly, I'm not sure that I'd class diopters as a 'compromise'. They are a practical solution to realigning focus and can produce improved image quality in some circumstances, but like any other optical component can also influence images in other, less desirable ways.