Sigma (Foveon imager) underwater?

So, at the risk of opening up a huge Bayer vs. Foveon technical "mine is longer" debate, and will full knowledge that the SD1 is not even released yet and Sigma's got a pretty short track record with DSLR bodies in general (much of it an apparent learning curve; the SD14 got somewhat panned for white balance issues and color casts among other things)....is there anyone else besides me who is curious about Sigma's new X3 imager underwater?

The supposed advantages of getting all colors at all photosites (regardless of how you count them and what constitutes a 'pixel' vs. a Bayer array) is that there's no AA filter on the sensor and the color has less interpolation spatially. So presumably stronger adjacent pixel color differences can be represented. The new sensor also has about 15MP 'sites' which is getting up to par with other crop DSLRs. Given much underwater photography doesn't rely on super high ISO, assuming strobe use, any total sensitivity or noise disadvantage vs. a Bayer sensor array DSLR might somewhat muted as well.

I've been reading the slow rollup to the SD1 release for some time and am seriously thinking of jumping on one when it releases (already made the commitment to leave Olympus 4:3rds, been unloading my more expensive Zuiko lenses on eBay and holding back just the E-30 and main 14-54 lens so I'm not naked without a camera while I wait). I know either Canon or Nikon would be a 'smarter' choice in terms of sheer market share and housing availability, but I always was a bit of a contrarian. BS-Kinetics has already indicated on their website that an SD1 housing will likely be coming, adapted from their current "Lut" carbon fiber housing made originally for the SD14.

Am I alone in being intrigued by what a different sensor technique might 'see' given how detailed and colorful even the smallest things can be underwater? The lack of an AA is probably not providing much 'advantage' as you're going to just be limited by the optics of the port or dome system and the clarity of the water, unlike on land...but its the lack of any requirement to interpolate color at all (spatially, I mean - of course the Foveon sensors have to 'interpolate' true color at each photosite from the results received from the 3 detectors at different color penetration 'depths' in the silicon) seems like it might have some advantages in terms of detail resolution, perhaps? The one thing that worries me the most is that the 'red' part of the X3 sensor is measuring the deepest penetration layer into the silicon, so presumably the 'gain' of the red channel might be a little reduced against the blue and green...and reds are what we tend to lose the most underwater to begin with. But there's a similar loss in terms of smaller actual red filtered site density on a Bayer, which is kind of contrary to underwater sensitivity desires in the same way.

Any other impressions? Want to talk me off this interest in a weird dead-end niche sensor approach cliff for my next DSLR purchase?

Wow, 105 reads and not a single reply....kind of implies I am indeed alone in my curiosity

I'll bite (with a nice controversial, conversation starting post):

I think the sensor would give excellent images - but my concern is whether anyone would put the time and effort into the housing etc to make it a usable system underwater.

A month or so ago, I was sent some underwater images by the Hasselblad UK rep - which have amazing image quality. Fabulous. But the images themselves are very, very poor - because the system just doesn't have the ergonomics/lens options to let you actually get a great shot a real subject in the underwater world. That system would be brilliant in the pool/underwater studio - when models can be directed and conditions controlled, but useless for wildlife/diving photography. All the image quality in the world is no benefit if the system is not good enough to actually get you the shot.

Unless someone put some serious effort into a housing for the Sigma - it might suffer a similar fate. Which is always the concern when you make a slightly leftfield choice in UW camera gear.

Alex

I agree with Alex; I think we forget the importance of ergonomics / ease of use, and availability of lenses and accessories etc is. Popular brands like Nikon and Canon will provide overall the best combination of image quality, lens availability, housings, ports, etc, etc. One really has to look at the overall system. This last weekend I got to use the Phase 645 back with Mamiya, and Nauticam housing. All fantastic pieces of equipment, and capable of amazing images, but when you put it all together probably much more limited in terms of variation of shooting setups. For example, the 28mm lens does not use the leaf shutter, so only 1/125 max flash synch is possible, a problem when trying to have a black backgound. Shooting the longer lenses no problem, so that leads me to a slightly more restricted range of shooting situations than the Nikon 1/320 with all lenses.

Shooting portraits in the pool, or to extrapolate - mid range fish portraits etc, the Phase etc would be unbeatable. But start to push the wide angle or macro extremes, you will probably run into limitations due to your equipment when compared to Nikon or Canon.

So you hve to look at the whole setup, unless you are looking to very specific shooting requirements in which case one system may outshine the other.

Am I alone in being intrigued by what a different sensor technique might 'see' given how detailed and colorful even the smallest things can be underwater? The lack of an AA is probably not providing much 'advantage' as you're going to just be limited by the optics of the port or dome system and the clarity of the water, unlike on land...but its the lack of any requirement to interpolate color at all (spatially, I mean - of course the Foveon sensors have to 'interpolate' true color at each photosite from the results received from the 3 detectors at different color penetration 'depths' in the silicon) seems like it might have some advantages in terms of detail resolution, perhaps? The one thing that worries me the most is that the 'red' part of the X3 sensor is measuring the deepest penetration layer into the silicon, so presumably the 'gain' of the red channel might be a little reduced against the blue and green...and reds are what we tend to lose the most underwater to begin with. But there's a similar loss in terms of smaller actual red filtered site density on a Bayer, which is kind of contrary to underwater sensitivity desires in the same way.

 

Any other impressions? Want to talk me off this interest in a weird dead-end niche sensor approach cliff for my next DSLR purchase?

Well I read your original post and didn't bother to reply then but here goes.

Alternative systems have to offer more than an incremental advantage. If they don't then the incremental advantage is just that and this simply doesn't make it worth pursuing as a niche within other niches. So producing an underwater housing for a camera which uses a different sensor that is unlikely to have a significant benefit (if any) isn't a logical thing to do.

I use a Leica M8 on land - no AA filter and it produces lovely images especially with the Leica M lenses. But so do my Canons and to be brutally blunt about it, only someone who was image literate and knew where the subtle differences were to be seen would be able to split the Leica/Canon images apart. I use it for two very simple reasons - because of its size, and perhaps most importantly, because I enjoy using a Leica rangefinder. (As an aside, I think that someone actually produces a housing for the M8 but I have absolutely no idea why - its an utterly pointless thing to do). The Leica M8 is a niche product which fits in fine for those of us who want to use it but despite being a very different camera it can't produce spectacularly different images - even though it has no AA filter, and microlenses fitted in front of a Kodak sensor

If you want to use a Foveon sensor underwater and think that the challenge of housing and using it would be enjoyable then please do so. Don't expect to produce world shatteringly different images with it because of its technical differences though because its highly unlikely that you will. As has been pointed out, there are many other reasons why existing products work well already.

Thanks for the replies. Alex, yours isn't controversial at all. I think I'm in "lust" so I'm looking for a slap of cold water to snap me back to reality. But as with most idiot youths (um, in the "photographer lifetime" sense, I'm getting to be an old fart in general) it may not sink in too quickly.

I think the ergonomics and whole system issues are very important, too, but since BS-Kinetics has already promised a housing (and their housings look to be similar to at least Ike's in overall size/shape/ergo, if not somewhere between Ike and the aluminum housings) I'm not as worried about that aspect in this case. I am still rather worried about price of course (supply - one housing, demand - so far, me tentatively and one pro of their own ). I know I'd for sure be buying into a niche re: the camera and housing, possibly even flash issues (probably manual control only, no TTL), but the lenses are Sigma lenses with their mount and would be expected to optically perform about the same as they do on Canon or Nikon mounts (which is to say, sometimes on par with their counterparts, sometimes a bit worse, depending on what you're willing to pay).

Regarding port optics and the like, I also assume the BS-Kinetics ports to be on par with the other offerings in the DLSR world. This is only APS-C size sensor and lenses after all, so some of the Hassleblad or MF comparisons don't really 'scale' in that regard. No extra humongous domes or extensions required with all the resulting extra optical precision necessitated. Overall system bulk wouldn't be any bigger than any other APS-C DSLR housing.

I am still curious if there's much impression in the pro community that (if you accept the marketing mantra of Foveon, which is itself a question not a certainty) having better pixel-to-pixel color capture vs. positionally interpolated might have advantages for the typical underwater subject. We all know DR (not ISO sensitivity per say - but range within the image) will make a big difference, for example, capturing that beautiful sun-dappling effect on shallow shots (shark or dolphin, etc.). That's one thing that I think doesn't get done justice compared to what our eyes see, and you really can see a big difference between older digitals and newer in that regard. Could there be a similar advantage to be had here for colorful underwater subjects? Or do you think the optical quality of water itself (even really clear water has a much shorter viewing range than air) is enough to negate that?

As for expecting to create world shattering images because of a technical difference, I know better. I won't ever create world shattering images because of personal inexperience and incompetence, period. Same likely goes for my above water shooting - I'll never be more than an amateur, and I recognize that. But I am interested in the Sigma above water, like Paul is with the Leica, so I'm still trading off the idea of splitting into two different systems for above and below water. Doubling up in that regard might not cost much more in the long run, especially if I do a smarter job of buying the body/housing/lens(es) to my skill level and frequency of U/W use, vs. buying a lower demand thus premium housing and port setup. Of course swapping between what I use a lot above for my infrequent but much anticipated trips below carries a burden of decreasing my familiarity and might impact the same as poorer ergonomics.

Anyway, the entire subject is a bit of technical mental wanking, I suppose. Was just curious if I was the only one, um, curious about the sensor at all.